Non-canonical functions of DNMT3A in hematopoietic stem cells regulate telomerase activity and genome integrity

Abstract

DNMT3A is a critical regulator of hematopoietic stem cell (HSC) fate decisions and the most recurrently mutated gene in human clonal hematopoiesis (CH). DNMT3A is described as a DNA methyltransferase enzyme, but cells with DNMT3A loss of function show minor changes in DNA methylation that do not correlate with altered gene expression. To explore the possibility that Dnmt3a has DNA-methylation-independent functions in HSCs, we created an allelic series of mice with varying levels of DNA-methylation-impaired Dnmt3a. Clonal expansion of Dnmt3a-deficient HSCs was rescued by Dnmt3a proteins lacking DNA methylation capacity, suggesting that Dnmt3a has important non-canonical functions in HSCs. Dnmt3a-null HSCs can be transplanted indefinitely, implying the ability to circumvent mechanisms that limit the replicative lifespan of HSCs, such as telomere shortening. Dnmt3a-null HSCs show increased telomerase activity and sustain telomere length over serial transplantation, revealing a previously unidentified role for DNMT3A mutations in regulating HSC longevity that is unrelated to DNA methylation function.

Keywords: DNA methylation; DNMT3A; hematopoietic stem cell; telomerase; telomere.

Figure 1:. Evidence for DNA methylation-independent functions of Dnmt3a in hematopoiesis

A) Homing assay showing quantification of GFP+ cells in BM of recipient mice 24-hours post-transplant (n = 5 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. B) Quantification of apoptotic cells by AnnexinV staining of wild-type (WT) HSPCs transduced with empty vector (EV) control lentivirus, or lentivirus expressing Dnmt3a or Dnmt3L isolated from the BM of recipient mice two-weeks post-transplant (n = 4–10 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01. C) Peripheral blood engraftment (%GFP+ cells) of WT HSPCs transduced with indicated lentiviral vectors (n = 5–10 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, ****p<0.0001. D) Hematopoietic lineage distribution spectrum of WT HSPCs transduced with indicated lentiviral vectors in blood of recipient mice 16-weeks post-transplant (n = 5–10 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. E) Peripheral blood engraftment (%GFP+ cells) of Dnmt3a^KO HSPCs transduced with indicated lentiviral vectors (n = 5–10 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. ****p<0.0001. F) Quantification of GFP+ HSCs in the BM of recipient mice 18-weeks post-transplant resulting from transduction of WT or Dnmt3a^KO HSPCs with indicated lentivirus (n = 5–10 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01. G) Average DNA methylation levels of CpGs within DMRs amongst comparator genotypes for WT HSPCs transduced with empty vector (EV) control lentivirus, or lentivirus over-expressing Dnmt3a or Dnmt3L. H) Heatmap showing average DNA methylation level of DMRs comparing WT HSPCs transduced with empty vector (EV) control lentivirus, or lentivirus over-expressing Dnmt3a or Dnmt3L. See also Figure S1.

Figure 2:. Altered function of Dnmt3a-deficient HSCs is rescued by Dnmt3a proteins lacking DNA methylation capacity

A) Number of colony-forming units (CFU) from third round of serial replating of wild-type (WT) or Dnmt3a^KO (KO) HSCs transduced with indicated lentiviral vectors. EV = empty vector, Dnmt3a = wild-type Dnmt3a, V712G = Dnmt3a^V712G, E752A = Dnmt3a^E752A, R832A = Dnmt3a^R832A (n = 5–12 per group, 3 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. ****p<0.0001.. B) Average DNA methylation levels of CpGs within DMRs amongst comparator genotypes for wild-type (WT) or Dnmt3a^KO (KO) HSPCs transduced with empty vector (EV) control lentivirus, or lentivirus over-expressing indicated Dnmt3a variants. C) Donor-derived 16-week peripheral blood engraftment for indicated genotypes across non-competitive serial transplantation (n = 3–8 per group for each transplant round, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05. D) Frequency of donor-derived HSCs in BM of recipient mice 18-weeks post-transplant over serial transplantation (n = 3–8 per group for each transplant round, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. E) Quantification of number of donor-derived HSCs per mouse (per two hindlimbs) for indicated genotypes in quaternary recipient mice (n = 5–8 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. **p<0.01, ****p<0.0001. See also Figure S1 and S2, and Table S1.

Figure 3:. Clonal expansion of Dnmt3a-null HSCs in serial transplant is rescued by DNA methylation-deficient Dnmt3a

A) 16-week peripheral blood engraftment and 18-week BM HSC abundance in secondary recipients of HSCs from indicated CAGG-Cre^ERT2 genotypes (n = 13–22 per group, 3 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, ***p<0.001. B) 16-week peripheral blood engraftment and 18-week BM HSC abundance in secondary recipients of HSCs from indicated Vav-Cre genotypes (n = 16–23 per group, 3 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. C) Quantification of donor-derived HSCs from indicated genotypes across serial WBM competitive transplant stages (n = 7–10 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. D) Percentage of HSCs in G₀ from Ki67/DAPI cell cycle analysis of donor-derived HSCs from indicated genotypes post-secondary transplant (n = 4 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. E) Cleaved caspase 3/7 activity in donor-derived HSCs from indicated genotypes post-tertiary transplant following 16-hour overnight culture (n = 3–4 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. F) Quantification of donor-derived HSPC and myeloid progenitor cell populations in BM of mice transplanted with indicated HSC genotypes 18-weeks post-tertiary transplant (n = 7–10 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. See also Figure S3 and S4.

Figure 4:. HSC DNA methylation profile does not correlate with gene expression and functional output

A) Principle component analysis based on global DNA methylation patterns in donor-derived bone marrow cells from secondary recipients of indicated CAGG-Cre^ERT2 HSC genotypes. B) DNA methylation level of CpGs within DMRs (defined by comparison of control versus Dnmt3a^KO cells) across indicated CAGG-Cre^ERT2 genotypes. C) DNA methylation levels within DMRs (defined by comparison of control versus Dnmt3a^KO cells) across indicated CAGG-Cre^ERT2 genotypes. D) Principle component analysis based on global DNA methylation patterns in donor-derived bone marrow cells from secondary recipients of indicated Vav-Cre HSC genotypes. E) DNA methylation level of CpGs within DMRs (defined by comparison of control versus Dnmt3a^KO cells) across indicated Vav-Cre genotypes. F) DNA methylation levels within DMRs (defined by comparison of control versus Dnmt3a^KO cells) across indicated Vav-Cre HSPC genotypes. G) Principle component analysis based on global transcriptome in donor-derived HSCs from secondary recipients of indicated Vav-Cre HSC genotypes. The eclipse represents the region that contains 95% of all samples that can be grouped from the underlying Gaussian distribution per genotype. H) Gene expression heatmap showing expression levels of DEGs (defined by comparison of control versus Dnmt3a^KO HSCs) across indicated Vav-Cre HSC genotypes. I) Gene expression in donor-derived HSCs of indicated genotypes from quaternary recipients assessed by qPCR (n = 3–4 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05. See also Figure S5 and S6, and Table S2, S3, S4, S5, S6, S7.

Figure 5:. DNMT3A loss-of-function increases TERT expression and telomerase activity

A) TRF gel showing telomere length in BM cells from control and Dnmt3a^KO mice across indicated transplant stages. B) Representative microscopy images of TelC immunostaining in HSPCs. C) Quantification of TelC staining intensity in control and Dnmt3a^KO HSPCs (n = 14–34 per group). Data are reported as means ± SEMs, analyzed by unpaired two-tailed t-test. *p<0.05. D) Quantification of Tert mRNA expression in control and Dnmt3a^KO HSPCs (Lineage− Sca-1+ c-Kit+) by qPCR (n = 16 per group). Data are reported as means ± SEMs, analyzed by unpaired two-tailed t-test. ***p<0.001. E) Telomerase activity quantified by TRAP assay showing higher telomerase activity (red box) in Dnmt3a^KO BM cells. Positive control is human embryonic stem cells (hESCs). LD = loading control. F) DNA methylation analysis of subtelomeric regions by bisulfite sequencing from HSPCs of indicated genotypes. G) Expression levels of Terc and Tert in HSCs (Lineage− Sca-1+ c-Kit+ CD48− CD150+) of indicated genotypes by qPCR (n = 5–8 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. H) Telomerase activity quantified by TRAP assay (red box) in BM cells of indicated genotypes. Positive control is human embryonic stem cells (hESCs). LD = loading control. I) TRF gel showing telomere length in BM cells of indicated genotypes post-secondary transplant. J) TRF gel showing telomere length in BM cells of indicated genotypes post-quaternary transplant. * = insufficient control cells could be obtained to generate necessary DNA. See also Figure S7.

Figure 6:. Dnmt3a loss-of-function partially rescues some functional defects of telomerase-deficient HSCs

A) Donor-derived 16-week peripheral blood engraftment for indicated genotypes across non-competitive serial transplantation (n = 3–8 per group for each transplant round, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01, ****p<0.0001. B) Donor-derived 18-week BM engraftment for indicated genotypes across non-competitive serial transplantation (n = 3–8 per group for each transplant round, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. ***p<0.001, ****p<0.0001. C) Donor-derived HSC frequency of indicated genotypes in BM 18-weeks post-transplant across non-competitive serial transplantation (n = 3–8 per group for each transplant round, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. ****p<0.0001. D) Number of donor-derived HSCs per mouse (per two hindlimbs) for indicated genotypes in BM of quaternary recipient mice (n = 5–8 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. ****p<0.0001. E) TRF gel showing telomere length of BM cells from indicated genotypes at conclusion of secondary and quaternary transplantation. F) TRAP assay showing telomerase activity from BM cells of indicated genotypes at conclusion of tertiary transplant. Red box highlights products of interest. Positive control is human embryonic stem cells (hESCs). LD = loading control. G) Expression of Terc in HSPCs of indicated genotypes by qPCR from cell post-primary and post-tertiary transplant (n = 3 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, ***p<0.001. H) Expression of Tert in HSPCs of indicated genotypes by qPCR from cell post-primary and post-tertiary transplant (n = 3 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01. I) c-circle assay to quantify alternative lengthening of telomeres (ALT) activity in BM cells of indicated genotypes post-secondary and post-quaternary transplantation. Positive control = U2OS cells. J) Representative images for co-immunofluorescence staining for PML and TRF2 to identify APBs in mouse HSCs. K) Quantification of APBs in HSCs of indicated genotypes isolated from quaternary transplant recipient mice (n = 10 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01.

Figure 7:. Loss of Dnmt3a suppresses the DNA damage response at dysfunctional telomeres

A) Western blot showing levels of the DNA damage response marker γH2AX in donor-derived BM cells of indicated genotypes post-tertiary transplant. B) Flow cytometric quantification of γH2AX+ HSCs in quaternary recipients (n = 4–9 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, **p<0.01, ****p<0.0001. C) Representative immunostaining for 53BP1 and TRF1 to identify telomere dysfunction-induced foci (TIF) in donor-derived HSPCs of indicated genotypes post-tertiary transplant. D) Quantification of TIFs (53BP1/TRF1 foci) per cell in in donor-derived BM cells of indicated genotypes post-tertiary transplant (n = 33–58 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, ****p<0.0001. E) Quantification of apoptotic HSCs across serial transplantation (n = 3–8 per group for each transplant round, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. **p<0.01, ***p<0.001, ****p<0.0001. F) Cleaved caspase 3/7 activity in donor-derived HSCs from indicated genotypes post-quaternary transplant following 16-hour overnight culture (n = 1–11 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. G) Representative comet assay microscopy images of donor-derived c-Kit+ BM cells of indicated genotypes from quaternary recipient mice. H) Quantification of DNA damage by comet assay olive moment in c-Kit+ BM cells of indicated genotypes from quaternary recipient mice (n = 145 per group). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. ****p<0.0001. I) Quantification of DNA damage by comet assay olive moment in c-Kit+ BM cells of indicated genotypes following two-hour camptothecin treatment (n = 170 per group, 2 independent experiments). Data are reported as means ± SEMs, analyzed by one-way ANOVA with Tukey correction. *p<0.05, ****p<0.0001. J) Western blot showing Dnmt3a and SPT16 levels in 32D cells expressing indicated HA-tagged Dnmt3a variant proteins. K) Co-immunoprecipitation and western blot showing interaction between Dnmt3a variant proteins and SPT16. Bottom panel shows interaction after treatment with daunorubicin.