CD10 delineates a subset of human IL-4 producing follicular helper T cells involved in the survival of follicular lymphoma B cells

The follicular lymphoma (FL) microenvironment is characterized by a strong infiltration of helper T cells displaying a complex phenotype, including an overexpression of both activation and exhaustion markers, and a specific gene expression profile (GEP), underlying altered T-cell activation, motility, and polarization.^1-5  Recently, we demonstrated more precisely that genes related to follicular helper T cells (T_FH), the specialized CD4^pos T cells involved in normal germinal center (GC) B-cell survival and differentiation,⁶  represent a significant part of FL-specific microenvironment signature and revealed their unique capacity to support malignant B-cell growth.^7,8  FL-T_FH are regarded as a promising therapeutic target in this still incurable disease.⁹  FL-T_FH are characterized by a specific cytokine profile, combining overexpression of interleukin (IL)-4, interferon (IFN)-γ, and tumor necrosis factor (TNF) -α, and decreased expression of helper T 17–related genes.⁸  However, specific markers associated with FL-T_FH heterogeneity and identifying precisely the tumor-supportive FL-T_FH subset are lacking.

In reactive lymphoid tissues, CD57 has been initially proposed as a marker of B-cell supportive GC-T_FH,^10,11  but further GEP and functional studies revealed that CD57^pos and CD57^neg T_FH are rather similar.¹²  Neuropilin 1 (Nrp-1) was also detected on a subset of T_FH, but no specific function could be attributed to Nrp-1^pos T_FH.¹³  Interestingly, CD10, a marker of immature T and B cells and GC B cells virtually absent on circulating mature T cells,¹⁴  has been reported on a subset of poorly characterized CD5^pos T cells within reactive lymphoid hyperplasia (RLH), FL, and marginal zone lymphoma,¹⁵  as well as on malignant T_FH in angioimmunoblastic T-cell lymphoma.^16,17  Such results raise the possibility that CD10 expression highlights a subset of T_FH within normal and malignant lymph nodes (LNs).

Combining GEP, histology, phenotype, and functional approaches, we demonstrate that CD10 expression is restricted to a unique subset of GC-T_FH, specifically amplified in the FL context. Moreover, CD10^pos FL-T_FH exhibit a peculiar IL-4^hiIFN-γ^lo TNF-α^hi cytokine profile associated with a strong capacity to sustain directly and indirectly malignant B-cell survival.

Details are provided in the supplemental Materials and Methods (available on the Blood Web site).

Subjects were recruited under institutional review board approval and the informed consent process according to the Declaration of Helsinki. Samples comprised LNs obtained from patients with FL, nodular lymphocyte predominant Hodgkin lymphoma (NLPHL), and mantle cell lymphomas (MCLs); tonsils collected from children undergoing routine tonsillectomy; and reactive LNs with follicular hyperplasia. CD3^posCD4^posCXCR5^hiICOS^hiCD25^neg T_FH, CD10^pos T_FH, and CD10^neg T_FH were sorted using a FACSAria (BD Biosciences) (purity >98%). Tonsil and FL B cells were purified using the human B-cell isolation kit II (Miltenyi Biotec).

Membrane and intracellular staining were performed using standard flow cytometry techniques. Data were acquired on a CyAn ADP flow cytometer and analyzed using Kaluza software (Beckman Coulter). Tissue sections were used for single immunohistochemical (Programmed cell death 1 [PD-1]), double immunohistochemical (CD10/PAX5), and double immunohistofluorescence stainings (CD10/CD3, CD10/inducible T-cell costimulator [ICOS], CD10/C-X-C motif chemokine ligand [CXCL] 13).

GEP of 7 FL-T_FH and 7 tonsil-T_FH was analyzed using GeneChip HG-U133 Plus 2.0 microarrays (Affymetrix) and normalized using Partek software. Microarray data are registered to the Gene Expression Omnibus under accession number GSE66384.

Purified FL malignant B cells were cultured alone, with an activation cocktail (CD40 ligand, IL-2, IL-4), or in the presence of purified CD10^pos or CD10^neg T_FH (ratio 1:1). After 48 hours, B-cell apoptosis was assessed on CD20^posCD2^neg B cells using active caspase-3 phycoerythrin apoptosis kit (BD Biosciences).

Statistical analyses were performed with the GraphPad Prism software using nonparametric Wilcoxon test for matched pairs, or Mann-Whitney U tests.

Scattered PAX5^negCD10^hi cells could be identified within neoplastic follicles in 16/19 FL samples and were characterized as CD3^posCD4^pos T cells with a mature GC-T_FH phenotype, that is, expressing high levels of C-X-C motif chemokine receptor (CXCR) 5 and PD-1, together with ICOS and CXCL13 (Figure 1A-C). FL-infiltrating PD-1^negCXCR5^neg non-T_FH and PD-1^intCXCR5^int pre-T_FH, as well as blood CD4^posCD45RA^negCXCR5^pos T cells representing circulating memory FL-T_FH did not express CD10 (Figure 1B and data not shown). To evaluate if CD10 expression on T_FH was FL specific, we tested other B-cell lymphomas with a follicular growth pattern. Whereas cells expressing PD-1 were rarely detected in MCL samples, in agreement with a lack of PAX5^negCD10^pos T cells, we confirmed the presence of numerous PD-1^pos cells forming rosettes around neoplastic cells in NLPHL, but these cells were essentially CD10^neg (data not shown). The lack of MCL-infiltrating T_FH and the absence of CD10 expression on PD-1^pos T cells in the NLPHL microenvironment were confirmed by flow cytometry (data not shown). By contrast, we identified PAX5^negCD10^pos T cells at a highly variable frequency within the GC of all RLHs. These cells tended to concentrate at the periphery of the GC and coexpressed CD3, ICOS, and CXCL13 (supplemental Figure 1A-B). To evaluate how CD10 expression could impact T_FH function, we first sorted paired CD10^pos and CD10^neg T_FH from tonsils and demonstrated similar expression of canonical T_FH genes (supplemental Figure 1C). Moreover, CD10^pos and CD10^neg T_FH displayed the same capacity to sustain immunoglobulin production by autologous purified B cells (supplemental Figure 1D). We next checked whether CD10^pos tonsil-T_FH could be enriched for CD57- or Nrp-1-expressing cells (supplemental Figure 1E). As reported for Nrp-1,¹³  CD10 was expressed by a higher proportion of CD57^pos than CD57^neg T_FH (22.2% [6.1% to 51.4%] vs 15.6% [3.4% to 35.2%], P < .001, n = 20). However, the expression of CD10 and Nrp-1 was mutually exclusive. Finally, CD10 was not expressed on CXCR5^hiICOS^hiFoxp3^posCD25^pos tonsil follicular regulatory T cells. Thus, CD10 identifies a distinct subset of fully functional GC-T_FH in human RLH. Interestingly, conversely to NRP1 and CD57, CD10/MME was significantly upregulated in FL-T_FH compared with tonsil-T_FH, as revealed by GEP analysis. Accordingly, the percentage of CD10^pos T_FH was increased in FL samples (6.17% [2.7% to 22.9%]) compared with tonsils (1.56% [0.2% to 10.1%], P < .001) (Figure 1D-E), raising the hypothesis that CD10^pos T_FH expansion could play a specific role in FL pathogenesis.

CD10^pos FL-T_FH were essentially Ki-67^neg but expressed high levels of HLA-DR indicating a nonproliferating but activated status (supplemental Figure 2). In addition, neither CD10^pos nor CD10^neg FL-T_FH expressed T cell immunoglobulin mucin-3 (TIM-3) confirming that they represent fully activated and not exhausted T cells (data not shown). We next looked for the capacity of T_FH to produce cytokines after in vitro restimulation. Interestingly, CD10^pos tonsil-T_FH produced similar levels of IL-21 and IFN-γ but significantly less IL-17A and more IL-4 than their CD10^neg counterpart, indicating that CD10 highlights a unique subset of T_FH with specific functional properties (Figure 2A). Compared with tonsil-T_FH, FL-T_FH displayed an IL-21^hiIL-17^lo phenotype independently of their CD10 expression. However, CD10^pos FL-T_FH exhibited a higher capacity to secrete IL-4 and a reduced capacity to secrete IFN-γ compared with CD10^neg FL-T_FH. In agreement, IL-4-producing FL-T_FH and IFN-γ-producing FL-T_FH belonged to nonoverlapping cell subsets (Figure 2B). These results prompted us to evaluate the direct role of CD10^pos vs CD10^neg FL-T_FH on malignant FL B cells. CD10^pos FL-T_FH were more efficient than their CD10^neg counterpart to support autologous malignant FL B-cell survival in vitro (Figure 2C). Finally, the previously reported overexpression of TNFA by FL-T_FH⁸  appeared uniformly supported by CD10^pos and CD10^neg FL-T_FH subsets, containing both ∼35% of TNF-α-producing cells. Altogether, our data suggest that these 2 FL-T_FH subsets may have different roles within the FL cell niche. CD10^pos FL-T_FH produce high amounts of IL-4 that could trigger B-cell activation, survival, and production of the Treg-recruiting CC chemokine ligands 17 and 22.^8,18  IL-4 also contributes to the polarization of tumor-associated macrophages¹⁹  that favor FL B-cell growth.^20,21  In addition, CD10^pos FL-T_FH exhibit a TNF^hiIFNγ^lo phenotype that could favor the induction and maintenance of a B-cell supportive lymphoid stroma network.²²  Conversely, overexpression of IFN-γ by CD10^neg T_FH could promote activation of cytotoxic CD8^pos T cells displaying efficient antitumor activity²³  but also expression of stroma-derived indoleamine-2,3 dioxygenase that could inhibit not only T-cell but also malignant B-cell proliferation.²⁴ 

In conclusion, our study supports the current understanding of the FL cell niche as an intricate network of cell interactions in which each cell subset should be exquisitely characterized individually and in relationship with the other partners before being proposed as a biomarker or therapeutic target.

The authors thank the Labex Immunotherapy Graft Oncology, the Centre de Ressources Biologiques–Santé (BB-0033-00056; http://www.crbsante-rennes.com) of Rennes Hospital for its support in the processing of biological samples, the Unité Mixte de Service Centre National de la Recherche Scientifique (CNRS) 3480/US INSERM 018 Biologie Santé Innovation Technologique (BIOSIT) for cell-sorting core facility, Christophe Ruaux for providing tonsil samples, and Christiane Copie and Maryse Baia for slide-scanning facility.

This work was supported by research funding from the Institut National du Cancer (INCA AAP PLBIO-13-085 and INCA_6530), the Ligue Nationale Contre le Cancer (Equipe Labellisée and “Carte d'identité des Tumeurs” program), and the Fondation “Association pour la Recherche contre le Cancer.” J.M. has been funded by FP7-PEOPLE-2011-Initial Training Network Stroma Cell-Immune Cell Interactions in Health and Disease (STROMA).

Contribution: P.A.-T. designed and performed research, analyzed data, and wrote the manuscript; S.H. designed and performed research, analyzed data, and corrected the manuscript; C.A., J.M., M.S.B., R.J., J.L.P., C.M., and N.M. performed research; and P.G. and K.T. supervised and designed research, analyzed data, and wrote the manuscript.

Conflict-of-interest disclosure: S.H. was supported by the Novartis Foundation (Basel, Switzerland). The remaining authors declare no competing financial interests.

Correspondence: Philippe Gaulard, INSERM U955, Centre Hospitalier Universitaire Henri Mondor, Faculté de Médecine de Créteil 8 rue du Général Sarrail, F-94010 Créteil, France; e-mail: philippe.gaulard@hmn.aphp.fr; and Karin Tarte, INSERM U917, Faculté de Médecine, 2 Ave du Pr Léon Bernard, F-35043 Rennes, France; e-mail: karin.tarte@univ-rennes1.fr.