Somatic Variants in the Human Lens Epithelium: A Preliminary Assessment

Abstract

Purpose: We hypothesize that somatic mutations accumulate in cells of the human lens and may contribute to the development of cortical or posterior sub-capsular cataracts. Here, we used a Next-generation sequencing (NGS) strategy to screen for low-allelic frequency variants in DNA extracted from human lens epithelial samples.

Methods: Next-Generation sequencing of 151 cancer-related genes (WUCaMP2 panel) was performed on DNA extracted from post-mortem or surgical specimens obtained from 24 individuals. Usually, pairwise comparisons were made between two or more ocular samples from the same individual, allowing putative somatic variants detected in lens samples to be differentiated from germline variants.

Results: Use of a targeted hybridization approach enabled high sequence coverage (>1000-fold) of the WUCaMP2 genes. In addition to high-frequency variants (corresponding to homozygous or heterozygous SNPs and Indels), somatic variants with allelic frequencies of 1-4% were detected in the lens epithelial samples. The presence of one such variant, a T > C point substitution at position 32907082 in BRCA2, was verified subsequently using droplet digital PCR.

Conclusions: Low-allelic fraction variants are present in the human lens epithelium, at frequencies consistent with the presence of millimeter-sized clones.

Figure 1

Dissection of the central and peripheral regions of the human lens epithelium. (A) The base of a 35-mm Petri dish (dark blue) is covered with four layers of Parafilm and imprinted with a 6-mm-diameter circle (light blue). The lens is positioned in the Petri dish with the anterior pole (AP) facing down. (B) A 7-mm-diameter circle is removed from the posterior pole (PP) and radial cuts are made in the remaining portion. (C) The lens is centered over the 6-mm circle, and the capsule is pinned to the base of the Petri dish bottom. The fiber cell mass is removed. (D) The central capsule (with its adherent layer of epithelial cells) is pinned to secure the tissue during dissection, and the central portion of the epithelium/capsule is removed using the 6-mm circle as a guide. The remaining part of the lens epithelium/capsule is collected separately.

Figure 2

Work flow used to identify somatic variants in genomic DNA extracted from human lens epithelial cells. Sequencing libraries were enriched for genes of the WUCaMP2 panel (Table 2) by targeted hybridization capture. The resulting library was sequenced on an Illumina HiSeq 2500 platform to obtain paired-end reads. Sequencing results were filtered at several levels. The reads were then aligned to a human reference genome (hg19), and PCR duplicates were removed. Variants were called using SAMtools and VarScan2 software. Finally, selected variants were inspected manually using the IGV visualization tool.

Figure 3

Variant allele frequency distribution in the peripheral lens epithelial sample from donor N146 (see Table 1). A total of 628 variants (blue diamonds) are detected with frequencies ranging from 1% to 100%. For each variant, the corresponding read depth is indicated (green triangle). Variants with frequencies of 50% or 100% represent heterozygous or homozygous SNPs, respectively. Seventy-eight low-allelic fraction (<10%) variants are detected. Many of the low-allelic fraction variants were not called as true somatic variants because they were associated with read depths of less than 300 (shaded area) or because they failed other quality control criteria (see Materials and Methods). Four variants (red diamonds) passed all filters (including sufficient read depth [red triangles]). The variant circled in red (the associated read depth value is also encircled in red) corresponds to a T > C substitution in the BRCA2 gene (see Table 5).

Figure 4

Somatic variants in matched ocular specimens. Tissue comparisons consisted of epithelia from right versus left lenses (A), central versus peripheral epithelium from individual lenses (B), or central versus peripheral lens epithelium versus cornea or retina (C). The number of somatic variants detected in each region is indicated. Additional information about the individual samples or the sequencing results is available in Tables 1, 4, and 5. P, peripheral epithelium; C, central epithelium; YO, years old.

Figure 5

The presence of a BRCA2 somatic variant detected in sample N146 (see Table 5; Fig. 3) is validated by ddPCR analysis. A TaqMan probe labeled with HEX fluorophore was designed to detect the wild-type (WT) BRCA2 allele, and a probe labeled with FAM fluorophore was designed to detect the variant BRCA2 sequence. The threshold lines (pink) were determined by the positive control (A), to separate clusters with WT or variant DNA. Green dots represent droplets with WT DNA, blue dots represent droplets with variant DNA, brown dots represent droplets with both WT and variant DNA, and black dots represent droplets with no target DNA. The variant was called as present in droplets with fluorescence units in FAM > 4000 and with fluorescence units in HEX < 3000. The DNA from the peripheral lens region (C) shows the presence of the variant, whereas samples from the central lens region (D) and from the retina (E) show the presence of WT DNA only. In the negative control (B), no HEX or FAM droplets were identified.