Analysis Of T Cell Receptor Repertoire Reveals Evidence For Antigen-Specific Response In CLL Lymph Nodes

It is now widely accepted that in chronic lymphocytic leukaemia (CLL), proliferation of the malignant B cell clone takes place in pseudofollicles positioned within lymphoid tissue. This complex environment is believed to provide CLL cells with signals necessary for survival and expansion. In recent years the role for T cells in these processes has started to emerge. By analysing formalin fixed paraffin embedded (FFPE) CLL lymph node (LN) sections by multi-parameter confocal microscopy, our group previously demonstrated that dividing CLL cells stay in close contact with CD4+ CD25+ Foxp3- T cells. In addition, in an in vitro assay, activated T cells were found to be capable of inducing CLL cell proliferation. Our subsequent flow cytometry analyses looked at the phenotype of the T cells isolated from the CLL lymph nodes by fine needle aspiration (FNA). LN T cells were found to express effector memory cell markers more commonly than their PB counterparts and also had much higher levels of the activation/exhaustion marker PD1.

Since PD1 is known to be a marker of T follicular helper cells (Tfh), in our current work we asked if Tfh cells are a significant component of the T cell infiltrate of CLL lymph nodes. We used confocal immunofluorescent microscopy on FFPE CLL LN sections and looked for coexpression of CD4, PD1 and ICOS – also used to identify Tfh cells. PD1 was expressed on 25% of CD4+ cells (95% CI 21-30) but only 4% of CD4+ cells expressed both PD1 and ICOS (95% CI 2.5-5.6) (n=6). The number of Tfh cells infiltrating CLL LN was shown to be low when compared with the numbers found in normal reactive LN germinal centres (33.3% ±3.0 CD4+ cells co-express PD1 and ICOS) and more closely resembled the interfollicular areas of the normal reactive LN, where 0.71% ±0.23 CD4+ cells co-express PD1 and ICOS (n=6).

As PD1 is expressed on chronically activated T cells, we next sought evidence for CLL T cell antigen-specificity. LN-FNA and matching peripheral blood (PB) CD4+ T cells from 6 CLL patients were sorted into PD1hi and PD1lo subsets and subject to spectratyping of their T cell receptor Vβ (TCRVβ) repertoire. Diversity was then assessed using an arbitrary scale in which a higher value indicated a loss of TCR diversity. There was a significant reduction in TCR diversity observed in the PD1hi subset in both LN and PB compartments compared to PD1lo T-cells (PB PD1hi vs PD1lo; 19.83 +/- 1.62 vs 14.67 +/- 2.42; p=0.0049; LN PD1hi vs PD1lo - 21.33+/-0.56 vs 16.00+/- 1.77; p=0.022). Importantly, CLL PB PD1hi cells were also shown to have a significantly reduced TCR diversity compared to PB PD1hi cells from normal age-matched controls (19.83+/-1.62 vs 12.67+/-2.54; p=0.039; n=6). TCR oligoclonality observed in the CLL LN and PB PD1hi CD4 T cells strongly supports the role of antigen in maintaining these two populations. Since PD1hi cells were far more frequent in the CLL LN than blood, these antigen driven interactions likely occur within the LN.

In order to look for further evidence of CLL LN T cell antigen-specificity we then performed high throughput TCRVβ CDR3 sequencing (Illumina Genome Analyser, analysed by Immunoseq; Adaptive Biotechnologies) of CLL LN-FNA and PB CD4+ cells. Analysis of TCR sequences obtained from two separate LN and matching PB samples sampled concurrently from the same individual (n=4 patients) revealed that there is a significantly higher commonality in TCR CDR3 clonotypes between two matching LN samples then between any LN and its matching PB sample (p=0.016). These results support the theory that the CLL LN is a site of specific antigen stimulation of CD4+ T cells.

Further support for this hypothesis came from the analysis of TCRVβ sequencing results from two sets of two CLL LN and PB samples collected from the same individual 1 month apart. Six persistent TCR CDR3 clonotypes were detected, which were present in all four LN but not in the PB. These clones made up around 1% of the total number of TCR sequences found in each of the LN.

In conclusion, our data provides strong new evidence that, just like the neoplastic B cells, CLL LN CD4+ T cells are antigen experienced and that their accumulation in CLL lymph nodes may be antigen-driven.