Changes in human lymphocyte subpopulations in tonsils and regional lymph nodes of human head and neck squamous carcinoma compared to control lymph nodes

Efficient interactions between T, B and antigen-presenting cells in T-dependent immune responses take place at the secondary lymphoid organs [1‐3]. T cells are located mainly in the paracortical zone, which includes the interfollicular regions. B cells are placed in small primary follicles in the cortex, which become secondary follicles or germinal centres (GCs) after antigenic stimulation [4].

Most recent studies of the GC reaction focused either on B cells, centroblasts and centrocytes, or on follicular dentritic cells. It is known that T cells have a crucial role in the development of the GC reaction, mediated by both cellular contacts and humoral factors (interleukins). GC T cells express CD40L (CD154) [5,6], a molecule which allows the interactions with CD40⁺ B-lymphocytes. High affinity B cells, selected by antigen retained in the surface of follicular dendritic cells (FDCs) [1,7], become antibody producing plasma cells or memory B cells. This distinction is determined by the signals of the CD40/CD40L interaction and by the type of interleukins secreted by T cells [5,6,8,9]. Non-selected B cells, however, die by apoptosis [2,8,10,11,12]. The activation and generation of memory T cells in the secondary follicles of lymphoid tissues remain unclear, although much is known about these processes in B cells. Some authors have shown that, in mice, T cells migrating to follicles are also able to proliferate the developing GC [13,14]. During the GC reaction, T cells become concentrated both, at and near the junction of the follicular mantle with the light zone. Some cells remain there after the end of the GC reaction [13].

The origin, migration and role of intra-GC T cells in human follicles is not accurately known, as it is not possible to study the kinetic of the GC reaction in humans. This study compares control lymph nodes with human tonsils and tumour reactive lymph nodes from patients with head and neck's carcinomas.

The identification and distribution of cells in these nodes has been achieved through the study of several markers and other membrane antigens. The markers used were CD69, which is a very early activation antigen, CD45RA, a marker mostly associated to virgin cells and CD45RO, a marker associated to memory cells. In addition to the differences between control and reactive lymph nodes, an interesting distribution of the B and T cells in several layers was found, when tissue sections were performed.

These results suggest a speculative model of the cellular traffic into the GC, giving a crucial role to the T cells in the regulation of the GC reaction.

The quantitative and histological changes, at the cellular level, that take place in human secondary lymphoid tissues are described in this paper. These result from either chronic stimulation by pathogens in tonsils, or by tumour stimulation in the draining lymph nodes of human head and neck squamous carcinomas. An increase in the percentage of B-lymphocytes in these organs compared to control lymph nodes, due to the generation of GCs after antigen stimulation, was observed by flow cytometry. In tissue sections, the phenotype and morphology of the primary follicle B cells, in control lymph nodes, were seen to be similar to the secondary follicle mantle B cells in tumor reactive lymph nodes or in tonsils. Several authors consider that after the stimulation, during the immune response, the human B cells placed outside the follicles, in the T cell areas, migrate to the region of B cells and FDCs that form the primary follicle [1,4,21,22]. The stimulated B cells proliferate giving rise to centroblasts, which push the naive B cells to the periphery of the GC, thus generating the follicular mantle [2,4,22].

Regarding the T cell population, the most significant data is the decrease of naive T-lymphocytes (CD45RA⁺) in reactive lymph nodes. Previous studies in other species have demonstrated that naive circulating lymphocytes (L-selectin⁺) penetrate in the lymph node by the blood vessels and go out by efferent lymphatics [15,16,23]. This decrease in naive T cells could be due to a change of phenotype (from CD45RA⁺ naive T-lymphocytes to CD45RO⁺memory cells) or to the dilution of this population due to the higher number of reactive lymph node cells. In addition, it has been described that memory cells are retained more often in the sites were antigen is present [15,24]. In fact, our studies in tissue sections show an increase of CD45RO⁺ cells compared to control lymph nodes. These CD45RO⁺ cells are located mainly in two zones: 1) between the follicular mantle and the light zone of the GC and 2) surrounding the secondary follicle.

Surprisingly, in the three types of tissues studied, CD69 antigen was only highly expressed in a T-lymphocyte fraction, while B-lymphocytes showed either a negative or a low expression. CD69 has been reported to be an early activation marker [19,20], therefore its presence was expected only in reactive lymph nodes and tonsils but not in controls. The study by confocal microscopy confirmed the flow cytometry results. CD69^dull B cell population is constituted by cells that form the primary follicles in control lymph nodes and also the GC follicular mantle in reactive lymph nodes and tonsils. It has been described that neither primary follicle B cells nor follicular mantle B cells proliferate during the immune response [6]. This could imply that both have the same origin. The small increase in the CD69 expression found in tonsils, by flow cytometry analysis, is probably due to the greater number of GCs in this tissue, which obviously provokes an increase in the number of mantle cells.

CD69^bright T-lymphocytes surround B-lymphocytes in primary follicles and are in close contact with the internal part of the CD19/CD69^dull mantle B cells in the secondary follicles. This suggests an unknown functional association between these CD69 positive T-lymphocytes and mantle B cells. As the CD69 ligand is still unknown, it is possible to speculate that this molecule could be involved in interactions between these two lymphocyte subpopulations.

By confocal microscopy, GCs in tonsils and in tumour reactive lymph nodes can be seen with a distribution of cells in several layers: a CD19^bright B cell zone, a CD69^bright/CD45RO⁺ intra-GC T cell layer, a follicular mantle B cell layer and an external T cell zone. A group of T-lymphocytes CD3⁺/CD69^bright/CD45RO⁺ is organized between the light zone of the GC and the follicular mantle, as a crown of cells. Their phenotype supports findings reported by other authors in mice [5]. Moreover, it is similar to the narrow T cell layer that surrounds the primary follicles in human control lymph nodes (Figure 4). It seems possible that these T cells (CD3⁺/CD69^bright/CD45RO⁺), already observed in the primary follicle, could give place to the GC T cells when an immune response is initiated.

This distribution in a four layer-structure of different lymphoid populations (B / T / B / T cells), indicates a very organized structure. From our results and with data from several authors, a speculative model of migration and activation of lymphocytes in the GCs is suggested (Figure 6). Some authors [4,25] support the hypothesis that T-B interaction takes place in the area that surrounds the primary follicle and that, later, stimulated B and T cells migrate to the follicle, where they proliferate [13], forming the GCs. In the inner part of the GC, these CD69^bright/CD45RO⁺ T-lymphocytes could be generated by T-B cellular interaction, migrating to the marginal zone of GC, in close contact with the mantle (Figures 4 and 5). In this model, it is proposed that proliferating memory T cells could continuously be exported to the CD45RO⁺ layer that surrounds the secondary follicle and that follicular mantle B-lymphocytes could participate in this migration. The CD69 molecule might participate in this process. Moreover, scattered memory T-lymphocytes were found among mantle B cells. The expression of CD69 varies in these T-lymphocytes, it diminishes moving out from the internal part of the GC and it is lacking in cells located near to the outside of the follicle. It is likely that these GC T-lymphocytes could interact with mantle B cells, and also could express all the necessary molecules to be able to interact with B-lymphocytes [6]. Interactions with mantle B cells may provoke an exit of T cells to the mantle outer zone. This may avoid continuous GC T-lymphocyte generation and expansion, which would lead to an unregulated or exacerbated immune response. T cells would join the physiological barrier created by mantle B cells and stop the immune response, keeping a resting memory T-lymphocyte store outside the follicle.

To confirm some of these hypotheses, experiments are currently being carried out in our laboratory. These include the study of whether the memory T lymphocytes from the outer layer are progeny of the lymphocytes placed in the inner layers.

Human control lymph nodes have a higher number of T lymphocytes than tonsils or tumor reactive lymph nodes. These are mostly T cells with naive phenotype (CD45RA⁺). Histologically, control lymph nodes only show primary lymph follicles. Tonsils and tumor reactive lymph nodes display an increase in the number of B cells and a decrease in the number of CD45RA⁺ T lymphocytes. Studies in tissue sections show the formation of secondary lymphoid follicles with a very organized structure constituted by alternative layers of subpopulations of B and T lymphocytes: CD19^bright B cell zone, CD69^bright CD45RO⁺ intra- GC T cell layer, a follicular mantle B cell layer and an external T cell zone. The characteristics of mantle B cells are nearly identical to the primary follicle B cells. We propose a hypothetical model of lymphocyte migration in the GC.