PRDM8 reveals aberrant DNA methylation in aging syndromes and is relevant for hematopoietic and neuronal differentiation

Abstract

Background: Dyskeratosis congenita (DKC) and idiopathic aplastic anemia (AA) are bone marrow failure syndromes that share characteristics of premature aging with severe telomere attrition. Aging is also reflected by DNA methylation changes, which can be utilized to predict donor age. There is evidence that such epigenetic age predictions are accelerated in premature aging syndromes, but it is yet unclear how this is related to telomere length. DNA methylation analysis may support diagnosis of DKC and AA, which still remains a challenge for these rare diseases.

Results: In this study, we analyzed blood samples of 70 AA and 18 DKC patients to demonstrate that their epigenetic age predictions are overall increased, albeit not directly correlated with telomere length. Aberrant DNA methylation was observed in the gene PRDM8 in DKC and AA as well as in other diseases with premature aging phenotype, such as Down syndrome and Hutchinson-Gilford-Progeria syndrome. Aberrant DNA methylation patterns were particularly found within subsets of cell populations in DKC and AA samples as measured with barcoded bisulfite amplicon sequencing (BBA-seq). To gain insight into the functional relevance of PRDM8, we used CRISPR/Cas9 technology to generate induced pluripotent stem cells (iPSCs) with heterozygous and homozygous knockout. Loss of PRDM8 impaired hematopoietic and neuronal differentiation of iPSCs, even in the heterozygous knockout clone, but it did not impact on epigenetic age.

Conclusion: Taken together, our results demonstrate that epigenetic aging is accelerated in DKC and AA, independent from telomere attrition. Furthermore, aberrant DNA methylation in PRDM8 provides another biomarker for bone marrow failure syndromes and modulation of this gene in cellular subsets may be related to the hematopoietic and neuronal phenotypes observed in premature aging syndromes.

Keywords: Aging; Aplastic anemia; DNA methylation; Dyskeratosis congenita; Epigenetic clock; Hematopoietic differentiation; Neuronal differentiation; PRDM8; Telomere; iPSC.

Fig. 1

Telomere age and epigenetic age are increased in dyskeratosis congenita and aplastic anemia. a Telomere lengths of 105 healthy donors [17], 65 aplastic anemia (AA), and 17 dyskeratosis congenita (DKC) patients were measured in granulocytes and correlated to chronological age. DKC and AA patients show a reduced telomere length compared to healthy controls. b Offset of predicted telomere age (delta age) was higher for DKC and AA patients than for healthy controls [17]. t test: *** P < 0.001, **** P < 0.0001. c Epigenetic age predictions of 70 AA and 18 DKC samples revealed much lower correlation to chronological age than 243 healthy controls [17, 25]. d The difference between predicted epigenetic age and chronological age (delta age) was higher for DKC and AA than for healthy controls, as described for other samples before [17]. t test: ** P < 0.01, **** P < 0.0001. e, f Telomere age and epigenetic age (e), as well as delta telomere age and delta epigenetic age (f), do not correlate in 62 AA and 13 DKC samples

Fig. 2

Hypermethylation in PRDM8 in dyskeratosis congenita and aplastic anemia. a DNA methylation was measured by MassARRAY at the CpG site cg27242132 in blood samples of 62 new aplastic anemia (AA) and 12 new dyskeratosis congenita (DKC) patients, as compared to 10 previously described healthy controls [18]. t test: ** P < 0.01. b, c DNA methylation in cg27242132 does not correlate to epigenetic age (b) or telomere age (c). d, e Barcoded bisulfite amplicon sequencing (BBA-seq) was used to determined DNA methylation around the two relevant CpGs (cg27242132 and cg19409579) in independent samples: 12 controls, 8 AA, and 5 DKC validated higher methylation in DKC and AA compared to controls across the amplicons with 13 (assay 1) and 17 (assay 2) neighboring CpGs. f–k The frequencies of DNA methylation patterns within individual BBA-seq reads are exemplarily depicted for both PRDM8 amplicons in a healthy donor (f, g), a DKC patient (h, i) and a AA patient (j, k)

Fig. 3

DNA methylation in PRDM8 in premature aging syndromes and differential gene expression. a To evaluate aberrant DNA methylation in PRDM8 (at cg27242132) in other premature aging syndromes, we used available DNA methylation profiles of 103 healthy controls (GSE36054, GSE32148, GSE49064), 4 DKC samples (GSE75310), 29 Down syndrome samples (GSE52588), 4 Werner syndrome samples (GSE42865), and 3 Hutchinson-Gilford-Progeria syndrome (HGPS) samples (GSE42865). DNA methylation levels (β values) revealed hypermethylation in all premature aging syndromes. t test: * P < 0.05, ** P < 0.01, *** P < 0.001. b–d Expression of all PRDM8 transcripts (b), the long PRDM8 transcript (NM020026.3; c), and the short transcript (NM001099403.2; d) was analyzed by qRT-PCR in 10 controls, 27 AA, and 14 DKC patients. t test: * P < 0.05; ns, not significant

Fig. 4

Impaired hematopoietic differentiation of PRDM8^−/− iPSCs. a Scheme of two main transcripts of PRDM8 (NM_020226.3 and NM_001099403.2) and sites of genomic editing. Two pairs of guide RNAs (gRNAs) were designed targeting the intron/exon boundary at the start codon of both transcripts. b Genome editing was confirmed by gene expression analysis after 14 days of embryoid body assay (normalized to GAPDH and PRDM8 expression in the undifferentiated control cells). c Phase contrast pictures of EBs after 16 days of hematopoietic differentiation. The control EB produces hematopoietic progenitor cells, whereas this is not the case for the PRDM8^+/− clone and the PRDM8^−/− clone, which consistently revealed enhanced growth. Scale bar, 500 μM. d Knockout of PRDM8 resulted in a significantly lower number of hematopoietic progenitor cells. t test: * P < 0.05; ns, not significant. e Cytospins supported impaired hematopoiesis after PRDM8 knockout. Scale bar, 500 μM. f The colony forming unit (CFU) potential is lost in PRDM8 knockout clones. g Flow cytometry substantiates hematopoietic differentiation of the control iPSCs (read line: autofluorescence; blue line: with antibodies)

Fig. 5

PRDM8 knock out leads to impaired neuronal differentiation. a Representative phase contrast pictures of neuronal differentiations on day 13. Control cells reveal typical aggregates of immature sensory neurons, which are absent in the PRDM8 knockout cells. Scale bar, 200 μM. b Quantitative RT-PCR analysis after 10 days of neuronal differentiation shows upregulation of neuronal markers in control cells, but not in PRDM8 knockouts (color code depicts mean fold change versus GAPDH). c Gene expression changes in RNA-Seq after 27 days of peripheral neuron differentiation (RPM, reads per million; genes with a log2 fold change above 2 or below − 2 are depicted). d CpG sites that are either 20% hypermethylated (red, 1511 CpGs) or hypomethylated (blue, 1738 CpGs) in the PRDM8^−/− clone compared to the isogenic control. e Epigenetic age of control and knockout iPSCs after neuronal differentiation was close to 0 year.