Single-cell whole-genome sequencing reveals the functional landscape of somatic mutations in B lymphocytes across the human lifespan

Abstract

Accumulation of mutations in somatic cells has been implicated as a cause of aging since the 1950s. However, attempts to establish a causal relationship between somatic mutations and aging have been constrained by the lack of methods to directly identify mutational events in primary human tissues. Here we provide genome-wide mutation frequencies and spectra of human B lymphocytes from healthy individuals across the entire human lifespan using a highly accurate single-cell whole-genome sequencing method. We found that the number of somatic mutations increases from <500 per cell in newborns to >3,000 per cell in centenarians. We discovered mutational hotspot regions, some of which, as expected, were located at Ig genes associated with somatic hypermutation (SHM). B cell-specific mutation signatures associated with development, aging, or SHM were found. The SHM signature strongly correlated with the signature found in human B cell tumors, indicating that potential cancer-causing events are already present even in B cells of healthy individuals. We also identified multiple mutations in sequence features relevant to cellular function (i.e., transcribed genes and gene regulatory regions). Such mutations increased significantly during aging, but only at approximately one-half the rate of the genome average, indicating selection against mutations that impact B cell function. This full characterization of the landscape of somatic mutations in human B lymphocytes indicates that spontaneous somatic mutations accumulating with age can be deleterious and may contribute to both the increased risk for leukemia and the functional decline of B lymphocytes in the elderly.

Keywords: B lymphocyte; aging; functional genome; single-cell whole-genome sequencing; somatic DNA mutation.

Fig. 1.

Somatic mutations accumulate with age in human B lymphocytes. (A) The number of somatic mutations per genome as a function of age. Red asterisks indicate individual cells. Each box corresponds to summary statistics of four cells from one donor. Regression was performed on the median numbers of mutations of the four cells from each donor, using exponential model nonlinear least squares regression. (B) Rainfall plot illustrating distances of neighboring mutations. (Top) Density of mutations (56 cells pooled) in kilobase bins. (Bottom) Distances of each mutation to its closest other mutation. (C) Mutation hotspots in Ig H chain regions visualized using the UCSC Genome Browser. In the mutation panels, each bar represents one SNV. In the ATAC sequencing (ATAC-seq) panel, obtained from pooled B cells from one young individual and one old individual, each bar represents one open chromatin peak. Other panels were obtained from annotations of the UCSC Genome Browser. The three vertical boxes highlight three hotspots identified in this region. (D) The SHM and CSR status of the 48 adult B cells; all eight B cells from cord blood are SHM⁻CSR⁻. As expected, the majority of adult B cells are classified into two categories: SHM⁻CSR⁻ (primarily naïve B cells) and SHM⁺CSR⁺ (memory B cells). A few cases of SHM⁺CSR⁻ and SHM⁻CSR⁺ cells are known as nonswitched memory B cells and GC-independent memory B cells, respectively.

Fig. 2.

Mutation signatures. (A) Contributions of signatures A, B, and D to all mutations in B lymphocytes in the different age groups. Signature C is from spontaneous mutations previously obtained for human primary fibroblasts. (B) Mutation signatures in the context of their flanking base pairs (in alphabetical order, e.g., in the first column C > A category from left to right, ApCpA > ApApA to TpCpT > TpApT). (C) Comparison between signature B and COSMIC1. The fractions of mutations are presented in alphabetical order (e.g., in the first column C > A category from top down, ApCpA > ApApA to TpCpT > TpApT). (D) Comparison between signature D and COSMIC 9.

Fig. 3.

Accumulation of mutations in the functional genome and genome overall during aging. Each data point represents the ratio of the number of mutations per cell to the average of the mutation number per cell in newborn B lymphocytes (functional genome and whole genome calculated separately). The ratios of the functional genome are in red, and those in the genome overall are in black. P values are two-tailed and were estimated using the Wilcoxon signed-rank test.