Ascorbic Acid Promotes KIR Demethylation during Early NK Cell Differentiation

Abstract

Variegated expression of killer Ig-like receptors (KIR) in human NK cells is a stochastic process exclusive to subsets of mature NK cells and CD8⁺ T cells. Allele-specific KIR expression is maintained by DNA methylation within the proximal promoter regions. Because KIR genes are densely methylated in NK cell progenitors, there is an implied stage of human NK cell development in which DNA demethylation takes place to allow for active transcription. When and how this process occurs is unknown. In this study, we show that KIR proximal promoters are densely methylated in less mature CD56^bright NK cells and are progressively demethylated in CD56^dim NK cells as they mature and acquire KIR. We hypothesized that ten-eleven translocation (TET) enzymes, which oxidize 5mC on DNA could mediate KIR promoter demethylation. The catalytic efficiency of TET enzymes is known to be enhanced by ascorbic acid. We found that the addition of ascorbic acid to ex vivo culture of sorted CD56^bright NK cells increased the frequency of KIR expression in a dose-dependent manner and facilitated demethylation of proximal promoters. A marked enrichment of the transcription factor Runx3 as well as TET2 and TET3 was observed within proximal KIR promoters in CD56^bright NK cells cultured with ascorbic acid. Additionally, overexpression of TET3 and Runx3 promoted KIR expression in CD56^bright NK cells and NK-92 cells. Our results show that KIR promoter demethylation can be induced in CD56^bright, and this process is facilitated by ascorbic acid.

FIGURE 1.

The majority of KIR promoter demethylation occurs between the CD56^bright and CD56^dimCD94^high stages of NK cell development. (A) Flow cytometry plots of KIR2DL1 and KIR3DL1 expression on gated CD56^bright NK cells, CD56^dimCD94^high NK cells and CD56^dimCD94^low NK cells freshly isolated from peripheral blood of two healthy donors in 2 independent experiments. The indicated NK cell subsets were sorted, and methylation of CpG sites within the proximal promoter regions of KIR2DL1 and KIR3DL1 were analyzed by bisulfite sequencing. (B) Schematics of the proximal promoter regions for KIR2DL1 and KIR3DL1 that were analyzed by bisulfite sequencing. Vertical lines represent the locations of CpG sites relative to the transcriptional start sites (indicated by arrows). (C) Bisulfite sequencing results. Each row represents a single sequencing reaction. Open circles represent unmethylated cytosines, and filled circles represent methylated cytosines. Missing circles represent ambiguous nucleotide calls from sequencing reactions.

FIGURE 2.

CD56^bright NK cells cultured with ascorbic acid exhibit elevated frequencies of KIR expression. Sorted (A) CD56^bright NK cells, (B) CD56^dimCD94^high and (C) CD56^dimCD94^low NK cells (n = 6) were cultured for 7 days on EL08–1D2 feeder cells with 10 ng/ml IL-15 and the indicated concentrations of ascorbic acid. Shown are the frequencies of KIR2DL1, KIR2DL2/3 and KIR3DL1 expression as determined by flow cytometry. Results are from 3 independent experiments. (D) qRT-PCR analysis of the relative fold expression for the indicated KIR transcripts in CD56^bright NK cells (n = 4) cultured for 7 days on EL08–1D2 feeder cells with 10 ng/ml IL-15 and either DMSO or 50 ng/ml ascorbic acid. Results are from 2 independent experiments. All data was normalized against ACTB. Data are shown as mean ± SEM. One-way ANOVA with multiple comparisons was used to determine statistical significance. *p < 0.05, **p < 0.01

FIGURE 3.

Ascorbic acid facilitates demethylation of KIR proximal promoter regions in cultured CD56^bright NK cells. CD56^bright NK cells were sorted from the peripheral blood of two healthy donors and cultured for 7 days on EL08–1D2 feeder cells with 10 ng/ml IL-15 and either DMSO or 50 ng/ml ascorbic acid in 2 independent experiments. Cells were then harvested, and methylation of CpG sites within the proximal promoter regions of KIR2DL1 and KIR3DL1 were analyzed by bisulfite sequencing. Each row represents a single sequencing reaction. Open circles represent unmethylated cytosines, and filled circles represent methylated cytosines. Missing circles represent ambiguous nucleotide calls from sequencing reactions.

FIGURE 4.

Runx3, TET2 and TET3 are enriched within the proximal promoters of KIR genes during culture with ascorbic acid. CD56^bright NK cells were cultured for 7 days on EL08–1D2 feeders with 10 ng/ml IL-15 and either DMSO or 50 ng/ml ascorbic acid. Binding of Runx1, Runx2, Runx3, TET2 and TET3 to the proximal promoters of KIR2DL1, KIR2DL2 and KIR3DL1 was analyzed by ChIP followed by qRT-PCR (n = 4). Fold enrichment was calculated relative to input. Results are from 3 independent experiments. Data are shown as mean ± SEM. One-way ANOVA with multiple comparisons was used to determine statistical significance. *p < 0.05

FIGURE 5.

TET3 and Runx3 promote KIR expression. (A) Sorted CD56^bright NK cells (n = 4) were retrovirally transduced with a control GFP vector or a TET3 overexpression vector and cultured for 5 days on EL08–1D2 feeder cells with 10 ng/ml IL-15 and either DMSO or 50 ng/ml ascorbic acid. Shown are FACS plots of KIR (combination of KIR2DL1, KIR2DL2/3 and KIR3DL1) expression on GFP⁺ NK cells from each condition from a representative donor (left). Cumulative data from all 4 donors is also shown (right). Results are from 2 independent experiments. (B) NK-92 cells were transduced with a control mCherry vector or a TET3 overexpression vector. Transduced cells were sorted, and a population of cells transduced with the TET3 overexpression vector was transduced again with a Runx3 overexpression vector. NK-92 cells were then cultured for 5 days with 10 ng/ml IL-15 and either DMSO or 50 ng/ml ascorbic acid. Shown are FACS plots of KIR3DL1, KIR2DL2/3 and KIR3DL1 expression on transduced cells from each condition from a representative experiment. Cumulative data from 3 independent experiments is also shown. Data are shown as mean ± SEM. One-way ANOVA with multiple comparisons was used to determine statistical significance. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

FIGURE 6.

No evidence of genome-wide transcriptional or epigenetic alterations in CD56^bright NK cell cultured with ascorbic acid. CD56^bright NK cells were sorted from the peripheral blood of two healthy donors and cultured for 7 days on EL08–1D2 feeder cells with 10 ng/ml IL-15 and either DMSO or 50 ng/ml ascorbic acid (n = 4) in 2 independent experiments. Cells were then harvested and analyzed by RNA-seq and whole genome DNA methylation arrays. (A) Principle components analysis and (B) hierarchical clustering of RNA-seq data from CD56^bright NK cells from each donor cultured with and without ascorbic acid. (C) Scatter plots of average beta methylation intensity values for CD56^bright NK cells from each donor cultured with DMSO compared to ascorbic acid. Red lines represent the line of best fit and cut offs for probes with greater then 2-fold difference between culture conditions. (D) Hierarchical clustering of whole genome DNA methylation array data from CD56^bright NK cells from each donor cultured with and without ascorbic acid.