Evolution of multiple cell clones over a 29-year period of a CLL patient

Abstract

Chronic lymphocytic leukaemia (CLL) is a frequent B-cell malignancy, characterized by recurrent somatic chromosome alterations and a low level of point mutations. Here we present single-nucleotide polymorphism microarray analyses of a single CLL patient over 29 years of observation and treatment, and transcriptome and whole-genome sequencing at selected time points. We identify chromosome alterations 13q14-, 6q- and 12q+ in early cell clones, elimination of clonal populations following therapy, and subsequent appearance of a clone containing trisomy 12 and chromosome 10 copy-neutral loss of heterogeneity that marks a major population dominant at death. Serial single-cell RNA sequencing reveals an expression pattern with high FOS, JUN and KLF4 at disease acceleration, which resolves following therapy, but reoccurs following relapse and death. Transcriptome evolution indicates complex changes in expression occur over time. In conclusion, CLL can evolve gradually during indolent phases, and undergo rapid changes following therapy.

Figure 1. Sample and clinical information.

(a) Schema of the sample isolation and the sequencing strategy. FACS, fluorescence-activated cell sorting. WGS, whole-genome sequencing. RNA-seq, RNA sequencing. (b) Clinical information and corresponding samples analysed. Top, white blood cell counts (WBC) from diagnosis to year 29 following diagnosis. The dashed grey line indicates the normal upper level of WBC. Additional clinical information is shown in Supplementary Fig. 1. Bottom, the matched sequencing samples.

Figure 2. Representative CNV profiles detected by SNP microarray.

(a) The CNV profiles of 16 time points during the years 8–28 from diagnosis are shown. (b) The SNP array plots of 13q− in years 10 and 20 are shown.

Figure 3. CNV profiles detected by WGS and single-cell WGS.

(a) The timeline of CNVs detected by combined SNP microarray and WGS. CNLOH, copy-neutral loss of heterogeneity. (b) The CNV profiles of 6q, 12 and 13q14 by single-cell WGS. Two cells were removed (Supplementary Figs 11 and 12). (c) The CNV profiles of the 13q14 region. The cells with a normal karyotype in this region were removed. The CNV profiles of all cells were showed in Supplementary Fig. 10b.

Figure 4. Dynamic changes revealed by single-cell RNA-seq analysis.

(a) Hierarchical clustering of single-cell RNA-seq data from 300 single cells of five time points. Each column represents a single cell, and each row represents a gene. (b) The Z-scores of cells from different time points in different clusters. (c,d) Cell expression profiles in pseudo-temporal ordering. Points represent single cells. Lines connecting points represent the edges of the minimum spanning tree by Monocle program. The thick lines represent the main path of the pseudo-temporal ordering.

Figure 5. The inferred tumour evolution path of the CLL patient.

The upper panel represents the timeline of the main treatments. The middle panel represents the inferred CNVs evolution model and the lower panel represents the percentage of cells from different clusters at expression level.