Characterization of intratumoral follicular helper T cells in follicular lymphoma: role in the survival of malignant B cells

Abstract

Accumulating evidences indicate that the cellular and molecular microenvironment of follicular lymphoma (FL) has a key role in both lymphomagenesis and patient outcome. Malignant FL B cells are found admixed to specific stromal and immune cell subsets, in particular CD4(pos) T cells displaying phenotypic features of follicular helper T cells (T(FH)). The goal of our study was to functionally characterize intratumoral CD4(pos) T cells. We showed that CXCR5(hi)ICOS(hi)CD4(pos) T cells sorted from FL biopsies comprise at least two separate cell populations with distinct genetic and functional features: (i) CD25(pos) follicular regulatory T cells (T(FR)), and (ii) CD25(neg) T(FH) displaying a FL-B cell supportive activity without regulatory functions. Furthermore, despite their strong similarities with tonsil-derived T(FH), purified FL-derived T(FH) displayed a specific gene expression profile including an overexpression of several genes potentially involved directly or indirectly in lymphomagenesis, in particular TNF, LTA, IL4 or CD40LG. Interestingly, we further demonstrated that these two last signals efficiently rescued malignant B cells from spontaneous and rituximab-induced apoptosis. Altogether, our study demonstrates that tumor-infiltrating CD4(pos) T cells are more heterogeneous than previously presumed, and underlines for the first time the crucial role of T(FH) in the complex set of cellular interactions within FL microenvironment.

Figure 1. Phenotypic characterization of CD4^pos T cells in malignant and reactive SLO

Frequency of CXCR5^hiICOS^hi (A, B), CXCR5^hiPD-1^hi (B), and FOXP3^posCD25^pos (C) among CD4^pos T cells from tonsils (Tons), reactive LN (rLN), FL LN, and DLBCL LN samples. (A) Open squares represent rLN samples with a strong follicular hyperplasia. Bars: median. *p<0.05; **p<0.001; ***p<0.0001.

Figure 2. Phenotypic characterization of FL T_FR

(A) Left: Representative plots of CD25 and FOXP3 expression among CXCR5^hiICOS^hiCD4^pos T cells in FL. Right: Frequency of CD25^posFOXP3^pos subset among CD4^posCXCR5^hiICOS^hi T cells from tonsils (Tons) and FL LN. (B) CXCR5 expression of CD25^posFOXP3^pos cells among CD4^pos T cells in Tons and FL LN. (C) Correlation between the percentage of T_FR, and the percentages of Treg or T_FH in FL LN.

Figure 3. FOXP3^posICOS^pos cells in FL LN and reactive SLO

Immunohistochemistry in biopsies of (A) two cases of FL (1,2; respectively x100 and X400; 3,4; respectively X100 and X400), and (B) a reactive tonsil (1,2; respectively x100 and X250) and a LN with follicular hyperplasia (3,4; respectively X100 and X250). Double staining for FOXP3 (nuclear, blue) and ICOS (membrane, brown) show only very few cells expressing both markers in reactive tonsils and LN, mainly surrounding reactive GC (arrows) whereas a significant proportion of cells – either scattered or in small clusters - in the neoplastic follicles of FL are double stained for FOXP3 and ICOS.

Figure 4. Gene expression of FL T_FH and T_FR

(A) PCA of data resulting from the gene expression analyses of Th1, Th2, Th17, Treg, tonsil-derived T_FH (Tons), FL-derived T_FR and T_FH. (B) Hierarchical clustering of Treg, FL T_FH and T_FR, and Tons T_FH. (C) TBX21, GATA3, RORC, FOXP3, BCL6, PRDM1, and PDCD1 gene expression in FL T_FH and T_FR, and Tons T_FH were compared to that in Th1, Th2, Th17 and Treg (n=4 for Tons T_FH, FL T_FH and T_FR, n=3 for Th1, Th2, Th17, and Treg). The arbitrary value of 1 was assigned to blood naive CD4^pos T cells. Bars: mean+/−SD.

Figure 5. Functional characterization of FL T_FH and T_FR

(A,B) Purified FL B cells were cultured alone or with autologous T_FH, nonT_FH, or T_FR. (A) Left: representative plots of CD86 (line) or isotype-matched (grey) staining on gated B cells. Right: CD86 expression fold change after coculture with T_FH or nonT_FH. The RMFI for each coculture condition was compared to the RMFI of B cells cultured alone, allowing the calculation of the ratio of RMFI (RRMFI). Bars: mean+/−SD (n=3). (B) Left: representative plots of active caspase-3 staining on gated B cells. Right: Percentages of inhibition of B-cell apoptosis in coculture with T subsets. Bars: mean+/-SD (n=3). (C) Activated (A) or non-activated (NA) CFSE-labelled effector T cells (Teff) were cultured alone or in presence of Teff, Treg, or T_FR, T_FH, and nonT_FH isolated from FL LN. Left: representative plots of CFSE staining. Right: Bars: percentages of CFSE^pos Teff displaying less (<G2, black) or more (● G2, white) than 2 cell divisions (n=3).

Figure 6. CD40L and IL-4 involvement in FL B-cell survival

(A) IL4 and CD40LG expression of Tons and FL T_FH. The arbitrary value of 1 was assigned to blood naive CD4^pos T cells. Bars: mean+/−SD (n=6). *p<0.05, **p<0.01. (B,C) Left: representative plots of IFN-γ (B) or IL-4 (C) expression of activated (A) or non-activated (NA) PD1^pos FL T_FH. Right: expression of IFN-γ (•B) or IL-4 (C) by T_FH or CD3^pos T cells from Tons or FL LN. Bars: median. *p<0.05, **p<0.01. (D) Relative number of viable FL B cells after culture in presence or not of Rituximab and CD40L+/−IL-4. Number of viable FL B cells obtained without CD40L+IL-4 and without Rituximab was assigned to 100. Bars: median (n=9). **p<0.01.