Promoter DNA methylation regulates progranulin expression and is altered in FTLD

Abstract

Background: Frontotemporal lobar degeneration (FTLD) is a heterogeneous group of neurodegenerative diseases associated with personality changes and progressive dementia. Loss-of-function mutations in the growth factor progranulin (GRN) cause autosomal dominant FTLD, but so far the pathomechanism of sporadic FTLD is unclear.

Results: We analyzed whether DNA methylation in the GRN core promoter restricts GRN expression and, thus, might promote FTLD in the absence of GRN mutations. GRN expression in human lymphoblast cell lines is negatively correlated with methylation at several CpG units within the GRN promoter. Chronic treatment with the DNA methyltransferase inhibitor 5-aza-2'-deoxycytidine (DAC) strongly induces GRN mRNA and protein levels. In a reporter assay, CpG methylation blocks transcriptional activity of the GRN core promoter. In brains of FTLD patients several CpG units in the GRN promoter are significantly hypermethylated compared to age-matched healthy controls, Alzheimer and Parkinson patients. These CpG motifs are critical for GRN promoter activity in reporter assays. Furthermore, DNA methyltransferase 3a (DNMT3a) is upregulated in FTLD patients and overexpression of DNMT3a reduces GRN promoter activity and expression.

Conclusion: These data suggest that altered DNA methylation is a novel pathomechanism for FTLD that is potentially amenable to targeted pharmacotherapy.

Figure 1

GRN expression in human lymphoblast cell lines is inversely correlated to its promoter methylation. (a) GRN net secretion was measured by ELISA in LCLs derived from neurologically healthy individuals (LCL #1-13), unaffected relatives of FTLD patients (LCL #14, 15, highlighted in blue) and FTLD-patients (LCL#16, 17, highlighted in red). n = 3, mean ± SEM. (b) Scheme of GRN promoter region. Red bars depict PCR-amplicons analyzed for DNA methylation levels by MassARRAY. Blue bars indicate full length and short promoter region that was cloned into the pCpGL vector for luciferase assays (compare Figure 3). White circles display CpG units in amplicons A-1 to A-5 and A-DAC quantified by MassARRAY. CpG units that were not analyzed are not shown. Asterisks indicate significant correlation between GRN mRNA expression or GRN secretion and GRN methylation at respective CpG unit (*p < 0.05, linear regression analysis, Benjamini Hochberg multiple testing and FDR correction, compare Additional file 1: Table S3). (c) Average DNA methylation levels in amplicons A-1 to A-5 for individual LCLs are plotted. Mean ± SD. Color code as in (a). (d) Correlation between GRN mRNA expression and average DNA methylation at CpG units 1, 2, 6, 8 and 11 is shown. GRN mRNA expression was quantified by qPCR and normalized to PGK1 and GAPDH. Relative mRNA expression levels were plotted against average DNA methylation levels. Correlation between parameters was quantified by linear regression analysis, r² and p-values are given. Color code as in (a). (e) Correlation between GRN secretion and average DNA methylation at CpG units 1, 2, 6, 8 and 11. GRN secretion was determined by ELISA and relative units (R.U.) were plotted against average DNA methylation levels. Correlation between parameters was quantified by linear regression analysis, r² and p-values are given. Color code as in (a).

Figure 2

The DNMT-inhibitor 5-aza-2′-deoxycytidine (DAC) reduces DNA methylation in the GRN promoter and increases GRN expression in LCLs. LCL #3 and #14 were treated daily with 0.5 μM DAC for 9 days. (a) Quantitative DNA methylation analysis of amplicon A-DAC (compare Figure 1b) shows progressive demethylation in the GRN promoter region. Average DNA methylation across all 15 CpG units in the A-DAC amplicon is shown. Mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, Student’s t-test. (b) GRN mRNA expression levels were analyzed by qPCR and normalized to PGK1. GRN mRNA expression strongly increased, while the expression of GAPDH did not change. Mean ± SD, n = 3, **p < 0.01, ***p < 0.001, ANOVA with Tukey’s Multiple Comparison test. (c) GRN secretion was quantified by ELISA and normalized to absolute cell number. Chronic treatment of LCL #3 and LCL #14 with 0.5 μM DAC for 9 days significantly increased GRN protein secretion into the cell culture supernatant, while the secretion in mock treated control cells remained low. Mean ± SD, n = 3, **p < 0.01, ***p < 0.001, ANOVA with Tukey’s Multiple Comparison test.

Figure 3

DNA methylation inhibits GRN promoter activity at distinct CpG units.In vitro methylated and unmethylated pCpGL plasmids containing the GRN core promoter region driving expression of firefly luciferase were transiently cotransfected into (a) HEK 293FT cells and (b) primary rat cortical neurons together with a Renilla luciferase expressing plasmid. The full length GRN promoter pCpGL plasmid, a GRN promoter construct with site specific mutations of the significant CpG units in amplicons A-1 and A-2, and a short GRN promoter construct lacking amplicons A-1 and A-2 were transiently cotransfected into (c) HEK 293FT cells and (d) primary rat cortical neurons together with a Renilla luciferase expressing plasmid. Luciferase reporter activity was measured 48 h (a + c, HEK 293FT) or 72 h (b + d, neurons) after transfection. Relative luciferase activity was determined by normalizing firefly luciferase against Renilla luciferase activity. The empty vector was used as negative control. Mean ± SEM, n ≥ 3. ***p < 0.001, Student’s t-test, sign. significant.

Figure 4

GRN promoter DNA methylation is altered in FTLD-TDP patients. Genomic DNA from human frontal cortex samples was subjected to bisulfite conversion to analyze DNA methylation in the GRN promoter by MassARRAY (See Figure 1b). (a) Box plots show average percentage of DNA methylation in amplicons A-1 to A-5 comparing healthy controls (white), FTLD patients (red), AD patients (blue) and PD patients (yellow). Whiskers depict 5–95 percentile, **p < 0.01, Kruskal Wallis test with Dunn’s Multiple Comparison Test. (b) Negative correlation of average DNA methylation across amplicon A-1 to GRN mRNA expression. Correlation between parameters was quantified by linear regression analysis, r² and p-values are given. Color code as in (a). (c) Relative GRN mRNA expression levels in brains from controls, FTLD, AD and PD patients quantified by qPCR and normalized to the housekeeping genes PGK1 and GAPDH. Mean ± SD.

Figure 5

In FTLD-TDP patients, DNMT3a mRNA expression levels are altered. Relative mRNA expression levels of DNMT1 (a), DNMT3a (b) and DNMT3b (c) in brain samples quantified by qPCR and normalized to the housekeeping genes PGK1 and GAPDH. Mean ± SD, *p < 0.05, Kruskal Wallis test with Dunn’s Multiple Comparison Test.

Figure 6

Overexpression of DNMT3a alters GRN promoter activity in primary cortical neurons and reduces GRN mRNA expression in LCLs. (a) pCpGL plasmid containing the GRN core promoter and a DNMT3a overexpression construct were transiently transfected in HEK 293FT cells (left panel) and in rat primary cortical neurons (right panel). Relative luciferase activity was determined by normalizing firefly luciferase against Renilla luciferase activity. Empty vectors were used as negative control. Firefly luciferase expression was significantly reduced upon DNMT3a overexpression. Mean ± SEM, n ≥ 3. ***p < 0.001, ANOVA with Tukey’s Multiple Comparison test. (b) Lentiviral expression of DNMT3a in LCLs #3 and #14. Overexpression was verified by qPCR five days after viral transduction. n = 5, mean ± SEM, *p < 0.05, Student’s t-test. (c) GRN mRNA expression levels were significantly reduced in DNMT3a overexpressing LCLs as quantified by qPCR and normalized to PGK1 expression levels. n = 5, mean ± SEM, *p < 0.05, Student’s t-test.