Expansion of PD1-positive T Cells in Nodal Marginal Zone Lymphoma

Abstract

The diagnosis of nodal marginal zone lymphoma (NMZL) can be challenging, with the differential diagnosis including other low-grade B-cell lymphomas, reactive hyperplasia, and even some cases of peripheral T-cell lymphoma (PTCL). PTCL may have a perifollicular growth pattern mimicking NMZL. We and others have noted an atypical distribution of T-follicular helper (T_FH) cells in some cases of NMZL. This study was prompted by the diagnosis of NMZL in several cases in which a marked increase of T_FH cells, as determined by staining for programmed death-1 (PD1), had prompted suspicion for a diagnosis of PTCL. We analyzed PD1 staining in 48 cases of NMZL to characterize the extent and pattern of the PD1-positive infiltrate. Three main patterns of PD1 staining were identified: follicular pattern (peripheral, n= 16; central, n = 9; mixed, n = 3), diffuse pattern (n = 4), and a reduced or normal staining pattern in residual follicles (n = 16). A comprehensive analysis of other T_FH markers was undertaken in 14 cases with a high content of PD1-positive cells that were confirmed as B-cell lymphoma by clonality analysis. We describe in detail 5 of these cases in which PTCL was an initial consideration. This study illuminates the diverse immunohistochemical patterns encountered in NMZL and highlights a diagnostic pitfall important for diagnostic accuracy.

Marginal zone lymphoma (MZL) is a heterogenous disease with 3 different subtypes recognized in the World Health Organization (WHO) classification: nodal marginal zone lymphoma (NMZL), extranodal MZL, and splenic MZL.¹ In particular, NMZL remains a challenging diagnosis, requiring exclusion of nodal involvement by the other types of MZL and other, mainly B cell, lymphomas for diagnosis.^1,2 However, cases of MZL with increased numbers of T cells have also been described, raising further diagnostic issues. In 1999, Campo et al³ noted an increase in intrafollicular T cells in the context of “splenic-type” NMZL. Subsequently, a study by Bob et al⁴ reported an extrafollicular expansion of pro- grammed death-1 (PD1) (CD279)-positive T cells in a subset of NMZL, and, more recently, a case of florid T-follicular helper (T_FH) cell hyperplasia in extranodal MZL has been published.⁵ PD1, a marker of germinal center–associated T_FH cells, is strongly expressed in the T_FH of reactive follicles.⁶ However, a prominent PD1-positive T_FH cell infiltrate, in either an intrafollicular or extrafollicular location, although not specific, is a feature associated with peripheral T-cell lymphomas (PTCLs) of T_FH cell derivation. Notably, NMZL and PTCL exhibit overlapping histologic patterns, with both tumor types typically showing perifollicular localizations.^7–9 Thus, an increase in PD1-positive cells in NMZL could represent a potential diagnostic pitfall.^6,10,11

This study was prompted by several cases of NMZL submitted in consultation to some of the authors (E.S.J., C.L.), in which the initial diagnosis under consideration was PTCL of T_FH cell derivation. Further immunophenotypic and molecular studies led to a final diagnosis of NMZL. These challenging cases led us to conduct a comprehensive study of 48 cases of NMZL, focusing on the distribution and extent of T_FH cells in relation to the neoplastic B cells. We describe the patterns of staining noted and review in greater detail 5 cases of NMZL in which a T-cell lymphoma of T_FH cell derivation was initially proposed.

MATERIALS AND METHODS

Histology and Immunohistochemistry

Cases of NMZL were identified from the files at the National Cancer Institute and the Cancer University Institute of Toulouse Oncopole. Additional cases were contributed by coauthors as members of the International Lymphoma Study Group, reviewed by an expert panel (E. S.J., E.C., S.P., S.H.S.), with a consensus diagnosis of NMZL reached. The study was approved by the Institutional Review Board of the National Cancer Institute.

Hematoxylin and eosin (H&E)-stained sections from each case were evaluated. Histology reports, available flow cytometry results, and immunohistochemistry slides were reviewed. Additional immunohistochemical studies were performed on available material with CD20, CD3, CD79a, CD10, BCL6, MUM1 (IRF4), CD4, CD8, PD1, ICOS, CXCL13, FOXP3, PDL1, CD21, IgD, kappa, and lambda. Immunohistochemistry was performed on a Ventana Benchmark automated immunostainer using UltraView detection. The panel of antibodies, clones, and sources are presented in Table 1.

TABLE 1.

Antibodies Used in the Immunophenotypic Analysis of NMZL

Antigen	Clone	Dilution	Source
CD20	L26	Predilute	Roche
CD3	2GV6	Predilute	Roche
CD10	SP67	Predilute	Roche
BCL6	G/191E/A8	Predilute	Roche
PD1	NAT 105	Predilute	Roche
ICOS	SP98	1:200	Abcam
CXCL13	BLC/BCA-1	1:500	R&D Systems
FOXP3	236A/E7	1:100	Abcam
CD4	SP35	Predilute	Roche
CD8	SP57	Predilute	Roche
IgD	Rabbit	1:500	Dako
Kappa	Rabbit	Predilute	Roche
Lambda	Rabbit	Predilute	Roche
CD21	1F8	1:30	Dako
CD79a	SP18	Predilute	Roche
MUM1	MRQ-43	Predilute	Cell Marque
PDL1	SP142	1:3	Abcam

The cytologic composition of the cases was evaluated and the presence of the following cells types recorded: centrocyte-like cells, monocytoid cells, and plasmacytoid cells/plasma cells. The morphology of these cell types has been described elsewhere.² Plasmacytoid differentiation required the presence of light-chain restricted plasmacytoid cells/plasma cells by immunohistochemistry.

PD1/PAX5 Double Staining Immunohistochemistry

Immunohistochemical PD1/PAX5 double staining was performed on formalin-fixed paraffin-embedded (FFPE) tissue sections. Sections were deparaffinized and rehydrated in graded alcohol, then placed in a low pH antigen retrieval solution (Dako S1699), and steamed for 30 minutes. Sections were blocked with TRIS-goat (5%) solution and incubated for 60 minutes at room temperature with PD1 monoclonal antibody diluted to 1:200 (Abcam; ab52587). The slides were rinsed in a TRIS buffer solution followed by a 30-minute incubation of anti-mouse secondary antibody. The slides were rinsed again, and Dako Liquid DAB+ Substrate Chromogen System (K3468) was used as the detection chromogen. Slides were placed in TRIS buffer solution and transferred to a Roche-Ventana BenchMark machine for automated PAX5 staining using the standard manufacturer protocol. Antigen retrieval was conducted in Cell Conditioning 1 (CC1, 950–124) buffer for 60 minutes. The slides were incubated with Ventana prediluted anti-PAX5 (SP34) Rabbit Monoclonal Primary Antibody for 60 minutes. The signal was detected using ultraView Universal Alkaline Phosphatase Red Detection Kit (760–501). Hematoxylin was used to counterstain.

Assessment of PD1 Staining Patterns

The pattern of PD1 staining in normal lymphoid tissue has been previously described in detail.^6,10 PD1-positive cells are present at the periphery of a normal reactive germinal center and are concentrated in the light zone.¹¹ Small numbers of PD1-positive cells may be associated with monocytoid B-cell clusters in reactive lymph nodes,¹¹ and a few PD1-positive cells are present in interfollicular areas.⁶ We considered cases to have a “normal” PD1 staining pattern if strong PD1 reactivity was confined to intrafollicular cells in a polarized fashion, concentrated in the light zone, with minimal PD1 reactivity outside the follicles (Fig. 1D). Extrafollicular PD1 staining is encountered in reactive conditions such as viral lymphadenitis or Castleman disease, as well as in lymphoma, although the staining intensity and number of positive cells are variable.^4,10,11 We, therefore, interpreted the pattern of PD1 staining, taking into account both the predominant location of the PD1-positive cells (follicular or extrafollicular) and the strength of the staining in comparison with that of T cells in residual follicles.

FIGURE 1.

Patterns of PD1-positive cells. A, Follicular pattern with peripheral PD1 staining. Four cases showing the spectrum of PD1 expression. B, Follicular pattern with a central accumulation of intrafollicular PD1-positive cells. C, Diffuse pattern showing moderate extrafollicular PD1 staining. D, Normal. Strongly PD1-positive cells are confined to and show polarization within reactive follicles.

Immunoglobulin Gene and T-Cell Receptor Gene Rearrangement Studies

DNA was extracted from FFPE tissue sections using QIAGEN QIAamp DNA FFPE Tissue Kit on a QIAcube robotic system according to the manufacturer’s protocol (QIAGEN, Germantown, MD).

Polymerase chain reaction (PCR) for immunoglobulin gene (IGH and IGκ loci) and T-cell receptor (TRG locus) rearrangements was performed. The IGH locus was assessed using either laboratory-designed consensus primers directed against V_H framework II, V_H framework III, and the joining region (J_H), as described previously,^12–14 or commercially available BIOMED-2 primer sets to all 3 V_H regions (InvivoScribe Technologies, San Diego, CA). IGκ analysis was performed using BIOMED-2 primer sets (InvivoScribe Technologies).¹⁵ The TRG locus was assessed using a single multiplexed PCR reaction, as described previously,¹⁶ with primers directed at all known Vγ family members and the Jγ1/2, JP1/2, and JP joining segments. PCR products were separated by capillary electrophoresis on an ABI 3130xl Genetic Analyzer (Applied Biosystems, Foster City, CA), and electropherograms were analyzed using the Gene-Mapper software, version 4.0. TRG PCR is capable of detecting a clonal population comprising a minimum of 2% to 5% of the total T-cell population, and, as performed in our laboratory, it can identify > 90% to 95% of all TRG gene rearrangements occurring in a clonal T-cell proliferation.

RESULTS

Demographic, Histologic, and Immunophenotypic Features

The study cohort included 48 patients: 18 male individuals and 30 female individuals with a median age of 65 years (range: 30 to 89 y). The most common sites of biopsy included cervical (n = 15), axillary (n = 9), and inguinal (n = 7) lymph nodes. Other sites included abdominal, submandibular, and supraclavicular lymph nodes (each n = 2) and subclavicular lymph nodes (n = 1). The site of the biopsy was not specified in 10 cases.

The cytologic composition of the cases was heterogenous, and most cases had a mixture of cell types. A monocytoid B-cell component was present in most cases (n = 36). Plasmacytoid differentiation was present in 23 cases. Light-chain restriction was demonstrable in 31 cases, by immunohistochemistry (n = 22/48) alone, flow cytometry alone (n = 2/16), or both flow cytometry and immunohistochemistry (n = 7/16). Most light-chain restricted cases showed kappa restriction (n = 23/31). Immunoglobulin (Ig) D was positive in 13/46 cases. One case was positive for CD5, but negative for CD23 and Cyclin D1. Seven of 25 (28%) of cases with an available Epstein-Barr virus–encoded small nuclear RNA (EBER) status were positive for Epstein-Barr virus in <1% of cells.

Analysis of PD1 Staining Patterns

Three main patterns of PD1 staining were identified and are detailed below: increased follicular staining (n = 28), a diffuse increase with extrafollicular PD1-positive cells (n = 4), and a normal or reduced pattern of PD1 staining (n = 16) (Fig. 1). All cases with increased PD1-positive T cells had a concomitant increase in total T-cell numbers, such that the infiltrate was T-cell rich. In contrast, cases with normal or reduced PD1 staining were B-cell rich, and more readily appreciated as B-cell lymphoma. Cases with increased PD1 staining did not show other differences, as compared with the entire cohort, with respect to clonal plasma cells, eosinophils, or morphology of the neoplastic B cells.

Pattern 1: Follicular Pattern (n = 28, 58%)

In the follicular pattern, staining was either predominantly peripheral or central within the follicles. The peripheral pattern (n = 16) was characterized by cords of PD1-positive cells concentrated at the periphery of expanded follicles (Fig. 1A). A subset (n = 4) included in the peripheral group had overlapping features with the normal pattern but were associated with a greater extent of PD1-positive cells than normally seen in reactive germinal centers.

The central pattern (n = 9) had a well-demarcated, uniform accumulation of intrafollicular PD1-positive cells, without the typical “polarization” usually present in reactive follicles (Fig. 1B). In some of these cases, the accumulation of T cells was extensive, such that the B-cell population was reduced and peripheralized within the nodules. Three cases had mixed peripheral/central patterns.

Pattern 2: Diffuse Pattern (n = 4; 8%)

Four cases had an increase in extrafollicular PD1-positive cells, with moderate to strong intensity. However, remnants of residual follicles were identifiable by focally strong staining (Fig. 1C).

Pattern 3: Normal/Reactive Pattern or Reduced Staining (n = 16; 33%)

Twelve cases had PD1-positive cells confined to the follicles in a normal pattern (Fig. 1D). The residual germinal centers or regressed follicles were usually identifiable morphologically. Four cases had a reduction in PD1-positive cells and were characterized by marked follicular colonization by B cells, and absent mantle zones by IgD staining.

Immunophenotype of PD1-positive Cells

To further characterize the PD1-positive cells, we performed additional immunohistochemical stains for ICOS, CXCL13, and FOXP3 in 14 cases with prominent infiltrates (11 cases with a follicular pattern of PD1 staining and 3 cases with diffuse PD1 expression, Table 2). Within this cohort of 14, > 25% of the CD4-positive T cells were positive for PD1, and, in most cases, it was > 50%. By double staining (11 cases), PD1/PAX5 coexpression was seen in only rare cells, confirming that the PD1-positive cells were of T-cell lineage.

TABLE 2.

Immunophenotypic Features and Clonality Analysis of 14 Cases of NMZL With Prominent PD1-positive T Cells

		T-Cell Marker Expression Within the PD1-positive T-Cell Infiltrate							Neoplastic B Cells		Molecular Studies
Case No.	PD1 Pattern	ICOS	CXCL13	CD10	BCL6	CD4	CD8	FOXP3	IgD	Kappa/Lambda	EBER	IG	TRG
1	Follicular	+	+	Rare cells only	Weak, focal +	+	−	Rare +	−	Kappa	ND	Clonal	Polyclonal
2	Follicular	ND	ND	Rare cells only	Weak +	+	−	ND	Weak+	Lambda	< 1% (< 1/HPF)	Clonal	Polyclonal
3	Follicular	Unsatisfactory	+ subset	+ subset	−/unsatisfactory	+	−	Rare +	−	NC	ND	Clonal	Polyclonal
4	Follicular	+	+ subset	+ subset	Weak, focal +	+	−	ND	−	Lambda	ND	Clonal	Polyclonal
5	Follicular	+ weak subset	+ subset	+ subset	+ in follicles	+	−/+	Scattered +	−	Lambda	−	Clonal	Polyclonal
6	Follicular	+	+ subset	+ subset	+ in follicles	+	−	Scattered +	ND	Kappa	< 1% (< 3/HPF)	Clonal	Polyclonal
7	Follicular	+	ND	Rare cells only	NC	+	−	ND	NC	Kappa	−	Clonal	Polyclonal
8	Follicular	+	+ subset	+ subset	+	+	−	Rare +	−	NC	−	Clonal	Polyclonal
9	Follicular	+	ND	+	+	+	−	ND	+	Kappa	ND	Clonal	Polyclonal
10	Follicular	Unsatisfactory	+ subset	Rare cells only	Weak, focal +	+	−	Rare +	NC	NC	ND	Clonal	Polyclonal
11	Follicular	+ weak	+ subset	+ subset	Weak, focal +	+	−	Rare +	+	NC	ND	Clonal	Polyclonal
12	Diffuse	Weak, focal + interfollicular	Scattered + in follicles	Scattered + in follicles	+ in follicles	+	−/+	Scattered +	−	Lambda	−	Clonal	Polyclonal
13	Diffuse	Unsatisfactory	ND	Scattered + in follicles	+ in follicles; weak, focal + interfollicular	+	−/+	ND	−	NC	−	Clonal	Polyclonal
14	Diffuse	+ in extrafollicular component	ND	+ in follicles	+ in follicles	+	−	ND	ND	Lambda	< 1% (< 5/HPF)	Clonal	Polyclonal

NC indicates noncontributory; ND, not done.

In cases with a follicular pattern of PD1 expression, the intrafollicular PD1-positive cells were also CD4 positive. ICOS was usually positive in a similar distribution to the strongly PD1-positive cells; however, CD10 and CXCL13 positivity was usually present only in a subset of the cells. Assessment of BCL6 was challenging, as BCL6 is positive in normal germinal center B cells, and, in our experience, nonspecific staining may be present outside follicles. Nevertheless, BCL6 positivity interpretable as in excess of B cells and in a similar pattern to T-cell rich areas was occasionally seen in cases with a follicular pattern of PD1 staining (illustrated in Figs. 3, 4).

FIGURE 3.

Follicular pattern mimicking PTCL. Case 2. A, The lymph node is effaced by a diffuse, vaguely nodular proliferation (H&E) composed of B cells (CD20) (B) surrounding nodules of T cells (CD3) (C) that express PD1 (D). E, Weak BCL6 expression is present in a similar distribution to the PD1-positive areas, which are also predominantly CD4 positive (F) and contain fewer CD8-positive T cells (G).

FIGURE 4.

Follicular pattern mimicking PTCL. Case 8. A, A nodular proliferation effaced the lymph node (H&E). B, CD20 stains B cells at the periphery of the nodules. C, The center of the nodules shows a marked accumulation of T cells (CD3) that express strong PD1 (D), ICOS (E), and BCL6 (F). A subset of cells expresses CXCL13 (G) and CD10 (H).

The extrafollicular PD1-positive component was usually negative or only weakly positive for additional T_FH markers; however, 1 case (case 14) had strong expression of ICOS in interfollicular areas. CD4-positive cells were greater in number than CD8-positive cells outside the follicles, suggesting that the PD1-positive cells also expressed CD4, although, in most cases, there was a substantial CD8-positive T-cell infiltrate. Weak, focal interfollicular BCL6 was present in 1 case. FOXP3, where performed, stained only scattered cells in a perifollicular or interfollicular distribution.

We performed programmed death-ligand 1 (PDL1) immunohistochemistry on 6 cases (5 follicular and 1 diffuse), which showed that PDL1 expression was limited to admixed histiocytes/dendritic cells, but the tumor cells were negative.

MZL Mimicking PTCL of T_FH Cell Phenotype

The extent and magnitude of the PD1-positive infiltrate evident by immunohistochemistry was variable and occurred along a spectrum. In 9 of the 14 cases, the extent of the T-cell infiltrate and the presence of an apparently abnormal PD1 staining pattern (either follicular or diffuse) was such that the initial diagnostic impression was that of PTCL of T_FH cell phenotype. We discuss in detail 5 cases that posed such a diagnostic challenge.

Histopathologic and Immunophenotypic Findings

Follicular Pattern of PD1 Staining (Cases 1, 2, and 8)

In case 1, there was a dense perifollicular infiltrate composed of PD1, ICOS, and CXCL13-positive T cells that surrounded “naked” reactive germinal centers, raising the possibility of a pattern 1 angioimmunoblastic T-cell lymphoma (AITL). However, other features of AITL, such as dilated peripheral cortical sinuses and expansion of follicular dendritic cell (FDC) meshworks beyond B-cell follicles were not identified (Fig. 2). The neoplastic B cells were mainly interfollicular and perifollicular.

FIGURE 2.

Follicular pattern mimicking PTCL. Case 1. A, The neoplastic cells have a perifollicular growth pattern around reactive follicles (H&E). B, CD20 stains reactive germinal centers and perifollicular atypical B cells. C, Most of the perifollicular cells are T cells (CD3) that express PD1(D), ICOS (E), and CXCL13 (F). G, PAX5 (red) and PD1 (brown) coexpression is not present, confirming that PD1 is not expressed by B cells.

In case 2, PD1-positive cells formed large nodular aggregates, which enveloped and surrounded regressed FDC meshworks. The neoplastic B cells were mainly interfollicular. BCL6 showed dispersed positive cells within the large nodules, which were otherwise mainly composed of T cells. A subpopulation of the BCL6-positive cells was strongly positive, consistent with normal centroblasts (Fig. 3).

In case 8, there was a central, uniform accumulation of intrafollicular PD1-positive T cells, which were surrounded by neoplastic B cells. The intrafollicular T cells were strongly positive for CD3, CD4, BCL6, PD1, and ICOS. A subset of intrafollicular cells was positive for CD10 and CXCL13, although they were much fewer than the PD1-positive, ICOS-positive cells (Fig. 4).

Two of the 3 cases showed light-chain restriction, with kappa excess in case 1 and lambda restriction in case 2. In addition, EBER was positive in rare small cells (< 1%) in case 2.

Diffuse Pattern of PD1 Staining (Cases 12 and 13)

In case 12, PD1-positive cells were markedly increased, with a predominantly diffuse pattern. However, there were focal nodular aggregates resembling residual follicles. ICOS, CXCL13, and CD10 stained a subset of cells within the residual follicles. Weak positivity with ICOS was also focally present in the interfollicular areas. The neoplastic B cells were mainly concentrated in nodules, but with focal larger cells in the interfollicular region. A lambda predominant plasma cell population was found to be present by immunohistochemistry (Figs. 5A–E).

FIGURE 5.

Diffuse pattern mimicking PTCL. Case 12. A, There is a nodular and interfollicular infiltrate in the lymph node (H&E) composed of serpiginous nodules of B cells (CD20) (B) and a heavy T-cell infiltrate (CD3) (C) with diffuse expression of PD1 in the extrafollicular areas (D). E, PAX5 (red) positive cells do not express PD1 (brown) consistent with the expression of PD1 on T cells. Case 13. F, There is diffuse architectural effacement (H&E) with significant numbers of T cells (CD3) (G). H, BCL6 is strongly positive in residual follicles, with focal weak interfollicular staining. I, PD1 is strongly expressed in follicles, but also shows moderate extrafollicular staining.

In case 13, PD1-positive cells were diffusely increased in the extrafollicular areas. A dense T-cell infiltrate was present throughout the node, admixed with the neoplastic B cells, positive for CD79a. Remnants of residual follicles were positive with BCL6, and there was focal, weak interfollicular positivity. CD10 stained scattered intrafollicular cells only (Figs. 5F–I).

Immunoglobulin Gene and T-Cell Receptor Gene Rearrangement Analysis

Clonal IG gene rearrangements and polyclonal TRG gene rearrangements were present in all 5 cases.

DISCUSSION

AITL and nodal PTCLs of T_FH cell origin constitute a new group in the revised fourth edition of the WHO Classification that includes follicular T-cell lymphoma and other nodal PTCLs previously categorized as variants of PTCL, not otherwise specified.¹ Many previous reports have described T-cell lymphomas with prominent perifollicular zones surrounding hyperplastic follicles, a morphology that is reminiscent of MZL and that presents a diagnostic pitfall.^7–9 This marginal zone–like pattern shares features with the so-called “early” phase of AITL, also referred to as pattern 1 AITL.^17,18

PD1 has been identified as a useful marker of germinal center–associated T cells and therefore of the group of T-cell lymphomas derived from T_FH cells.⁶ However, PD1 staining is not specific for T_FH cells or even T cells, and extrafollicular PD1-positivity is also present in reactive conditions.^10,11,19 T-cell infiltrates have been described in NMZL as nodular infiltrates within residual germinal centers in “splenic-type” MZL cases,³ and strong expression of PD1 in extrafollicular T cells⁴ has also been observed.

Our study examined the location of PD1-positive T cells in 48 cases of nodal involvement by MZL. We identified 3 patterns of infiltration. Although many cases (n = 16, 33%) showed either a reduction in PD1 staining or a normal pattern of PD1 staining in residual, reactive, regressed, or colonized follicles, the majority (n = 32; 66.6%) showed a pattern of PD1 staining that departed from that traditionally classed as “normal.” Cases with an abnormal PD1 distribution showed an infiltrate of PD1-positive cells in follicular (n = 28) or diffuse (n = 4) patterns. The intensity of the PD1 expression varied between MZL cases and between lymph node compartments, being strongest within the residual follicles. Thus, the presence of strongly PD1-positive cells was useful in identifying residual follicles.

We characterized 14 cases in greater detail, in which the number and intensity of PD1-positive T cells were markedly increased. In 9 of the 14 cases, PTCL was an initial diagnostic consideration. The cases were morphologically diverse; however, many resemble a “pattern 1” AITL, with PD1-positive cells surrounding naked germinal centers or follicular T-cell lymphoma, composed of T-cell nodules with a T_FH cell phenotype.^17,18,20,21 In all cases, a review of immunohistochemical stains showed an atypical B-cell population with minimal atypia in the T-cell infiltrate. Peripheralization of B-cell areas, a feature of AITL, was not present, and, while the FDC meshworks showed expansion and colonization in some cases, they remained associated with B cells. Although immunohistochemistry revealed light-chain excess in most cases, the presence of an atypical B-cell infiltrate and demonstrable light-chain restriction by immunohistochemistry does not preclude consideration of PTCL. Monotypic Epstein-Barr virus–negative B-cell proliferations are a well-described phenomenon in the context of PTCL, in particular AITL, and may predominate, obscuring the underlying T-cell process.^22,23 These proliferations are also characterized by varying degrees of plasma cell differentiation, further complicating their separation from MZL, in which plasmacytoid differentiation is a common feature. Rare EBER-positive cells may also be observed, also a common finding in AITL. In all 14 cases, molecular analysis showed clonal IG gene rearrangements and polyclonal TRG gene rearrangements, confirming the diagnosis of MZL with a marked infiltrate of PD1-positive T cells.

The significance of this prominent PD1-positive T-cell population in a subset of MZL is uncertain. Although often used in the diagnostic setting as a marker of T_FH cells, PD1 expression may be present on other cells and T-cell subsets depending on the context of the immune response. PD1 is expressed on CD4⁺ and CD8⁺ T cells during T-cell activation, where it binds to PDL1 or PDL2 and plays a crucial role in limiting the immune response and restoring immune homeostasis in the peripheral tissues.¹⁹ The expression is also found on regulatory T cells, T-follicular regulatory cells, memory T cells, B cells, and natural killer cells.¹⁹ Sustained expression of PD1 is present in CD4⁺ and CD8⁺ T-cell subsets in the setting of chronic antigenic stimulation and T-cell exhaustion, and tumors may exploit this immune checkpoint pathway as a mechanism to evade the host immune system.²⁴ The analysis of 14 cases with T_FH cell markers (ICOS, CXCL13, BCL6, CD10) and a T-regulatory cell marker (FOXP3) showed that, particularly in the follicular pattern, the PD1-positive cells may also express other T_FH markers, thus contributing to the diagnostic consideration of a T_FH cell–derived lymphoma and potential misinterpretation.

In conclusion, our study shows that the PD1-positive T-cell component of the tumor microenvironment in MZL is variable; however, the functional implication of the infiltrate and the differences between cases with increased follicular localization of PD1 positive cells and those with an extrafollicular increase have yet to be determined. Of crucial importance to the diagnostic pathologist, however, is that these infiltrates may be substantial and present a pitfall in the distinction with T-cell lymphoma. Careful assessment of histologic features and molecular analysis is essential for correct diagnosis.