1012 - High-dimensional Single-cell Characterisation of the Chronic Lymphocytic Leukaemia Tumour Microenvironment using Imaging Mass Cytometry

Introduction:

In chronic lymphocytic leukaemia (CLL), the lymphoid tumour microenvironment (TME) can drive disease progression. Cellular and molecular interplay skews the TME towards a pro-survival and proliferative milieu, culminating in progressive accumulation of CD5+ B-cell clones. A systematic evaluation of the CLL TME may enable improved diagnostic, prognostic and therapeutic patient stratification. Imaging mass cytometry (IMC) permits the simultaneous multiplexed assessment of over 40 antigens across multiple cells on a single tissue slide, superseding the limitations of conventional immunohistochemistry. Applying IMC, we aimed to evaluate the CLL TME landscape with single-cell resolution to facilitate deep phenotyping and characterisation of its cellular composition, with the intention of exploring how spatially resolved features associate with clinical outcome.

Methods:

Archival formalin-fixed paraffin-embedded lymph node tissue slides were obtained from patients with CLL (n = 7). Non-malignant lymph node specimens (n = 2) were available for comparison. Tissue slides were stained with a previously optimised panel of primary metal isotope-conjugated antibodies designed to identify lineage-specific cell subsets, functional status and stromal elements. Images were acquired using the Fluidigm Hyperion Imaging System. Following extraction of raw data, automated, deep-learning based cell segmentation was employed to identify single-cells. A mixed strategy of marker-based cell characterisation (ASTIR) and consensus clustering was employed to categorise cell populations. For clustered populations, cell identities were assigned manually based on canonical marker expression. Computational and statistical analyses were performed using R or Python.

Results:

After segmenting a total of 163,212 cells from IMC images, we profiled the landscape of the CLL TME. We identified a diversity of cell populations including B-cells, T-cell subsets, monocytes, macrophages, dendritic cells, NK-like-cells, and other stromal components. Compared with non-malignant lymph node data we observed that CLL considerably alters the TME. Indeed, across CLL tissues, we observed substantial heterogeneity regarding the presence of cellular populations. In CLL, the median relative frequency of malignant cells was 66.4% (13.4% to 85.1%). CLL B-cells expressing Ki67, indicative of a proliferative state, comprised 0.2% to 8.4% (median, 2.1%) of the tumour bulk. During preliminary statistical analyses, we found that the frequencies of CLL B-cells per sample were negatively associated with macrophage proportions (Spearman; R = -0.84, p = 0.024), although up-scaling of our sample size and careful statistical analysis to eliminate technical artefacts are needed. Through subclustering key parent lineages, we uncovered heterogeneous cellular subpopulations including previously-described CD4+ CD45RO+ T-cells as well as, interestingly, CD45RAhi and CD57+ CLL B-cells which to our knowledge have not been commonly characterised. Taken together, these findings highlight the diversity of the CLL TME and will guide further phenotypic and spatial analyses using IMC.

Conclusions:

IMC has previously been utilised to spatially characterise the anatomy of solid tumours and haematologic neoplasms such as diffuse large B-cell lymphoma at a cellular level, but to our knowledge has not yet been applied in CLL. Our work thus far has confirmed the feasibility of highly multiplexed analysis of the CLL TME composition using IMC. Furthermore, we have demonstrated a diversity of cellular populations, of which preliminary analysis suggests a substantially altered composition compared with normal lymph node. These data are consistent with the premise that CLL can shape the TME. We have additionally uncovered inter-patient variability which mirrors the characteristically heterogeneous clinical course and genomic paradigm of the disease. In future, we will perform an in-depth characterisation of the spatial context of cellular neighbourhoods to gain insights into interactions between proliferating tumour cells and their microenvironment. Therefore, we aim to derive meaningful clinical correlates associated to outcome that can be ultimately translated into practice.