The Polycomb group gene Ezh2 prevents hematopoietic stem cell exhaustion

Abstract

The molecular mechanism responsible for a decline of stem cell functioning after replicative stress remains unknown. We used mouse embryonic fibroblasts (MEFs) and hematopoietic stem cells (HSCs) to identify genes involved in the process of cellular aging. In proliferating and senescent MEFs one of the most differentially expressed transcripts was Enhancer of zeste homolog 2 (Ezh2), a Polycomb group protein (PcG) involved in histone methylation and deacetylation. Retroviral overexpression of Ezh2 in MEFs resulted in bypassing of the senescence program. More importantly, whereas normal HSCs were rapidly exhausted after serial transplantations, overexpression of Ezh2 completely conserved long-term repopulating potential. Animals that were reconstituted with 3 times serially transplanted control bone marrow cells all died due to hematopoietic failure. In contrast, similarly transplanted Ezh2-overexpressing stem cells restored stem cell quality to normal levels. In a "genetic genomics" screen, we identified novel putative Ezh2 target or partner stem cell genes that are associated with chromatin modification. Our data suggest that stabilization of the chromatin structure preserves HSC potential after replicative stress.

Figure 1.

Expression of Ezh2 in MEFs. (A) At first MEFs showed rapid proliferation, but senesced after 8 population doublings. (B) RT-PCR analysis of young (p1) and aged (p5) MEFs. (C) Protein levels of Ezh2 in untreated MEFs at passage 1 and 9. (D) Schematic representation of retroviral vectors used to overexpress Ezh2. (E) Detection of Ezh2 protein in MEFs transduced with control and Ezh2 vector at different passages after transduction. (F) Relative expression of Ezh2 mRNA in MEFs transduced with the control vector (▪) or with the vector containing Ezh2 () at different passages after transduction. Expression levels of Ezh2 were determined by quantitative PCR and calculated relative to MEFs transduced with control at passage 1 after transduction. (G) Growth of MEFs after retroviral transduction (control, ▪; Ezh2, .

Figure 2.

Expression of Ezh2 in hematopoietic cells. (A) To isolate the different lineage-negative (Lin^-) populations from BM, the 5% most Lin^- cells were selected. (B) Cells in the Lin^- population were sorted based on Sca-1 and c-kit surface markers. (C) Expression of Ezh2 as measured by qPCR in the different Sca-1 and c-kit populations relative to Sca-1^-c-kit^- BM cells. (D) Growth of Lin^-Sca-1⁺c-kit⁺ cells cultured in the presence of GM-CSF and SCF. (E) Relative expression of Ezh2 was monitored by qPCR at different time points after initiation of differentiation with GM-CSF and SCF. Day 0 was set at 1. Insert shows gel analysis of amplified RT-PCR products.

Figure 3.

Overexpression of Ezh2 in HSCs. (A) WBC counts of mice given transplants with BM cells transduced with control (▪) or Ezh2 () vectors (n = 11 recipients per group, 2 independent experiments). Mean values plus or minus standard error of the mean (SEM) are shown. (B) Chimerism at different time points after transplantation. The percentage of donor-derived transduced CD45.1⁺GFP⁺ WBCs is shown. Average values ± 1 SEM of 2 independent experiments are shown. (C) Ezh2 protein expression in the spleen about 120 days after primary transplantation of control or Ezh2 CD45.1 BM cells. (D) CAFC frequencies of sorted LSK GFP⁺ cells 120 days after primary transplantation (control, ▪; Ezh2, ). (E) Chimerism levels of secondary recipients, competitively transplanted with various ratios (2:1, ⋄; 1:1, ○; 1:2, □; 1:5, ▵) of transduced/nontransduced and freshly isolated BM cells, analyzed 3 months after transplantation. Values show data from individual recipients in 2 independent experiments (n = 35/group). (F) CRI calculated for CD45.1⁺GFP⁺ (transduced) versus CD45.1⁺GFP^- (nontransduced) cells (see insert). Values (+ 1 SEM) are averages 3 months after secondary transplantation, based on 11 and 24 individual mice in the first and second experiment, respectively (control, ▪; Ezh2, ). (G) CRI calculated for transduced cells (CD45.1⁺GFP⁺) compared to freshly isolated BM cells (CD45.2⁺; see insert). Averages values (+ 1 SEM) of 11 and 24 individual mice of the first and second experiment, respectively, are shown (control, ▪; Ezh2, ). (H) LTRA frequencies in CD45.1⁺GFP⁺ (transduced) cell fractions calculated from limiting dilution analyses 3 months after secondary transplantation in 2 independent experiments (control, ▪; Ezh2, ).

Figure 4.

Effects of overexpression of Ezh2 in HSCs in tertiary recipients. (A) Representative fluorescence-activated cell sorting (FACS) plot showing donor contribution in a recipient transplanted with 2.5 × 10⁶ twice serially transplanted Ezh2-overexpressing cells in competition with 5 × 10⁵ freshly isolated CD45.2⁺ BM cells. Left panel shows contribution when gated on myeloid cells; right panel shows donor contribution for lymphoid cells. (B) The percentage of CD45.1⁺GFP⁺ cells in the peripheral blood in all recipients (n = 28) 3 months after tertiary transplantation. Cells were cotransplanted in different ratios (5:1, ▵; 2:1, ⋄; 1:1, ○) with freshly isolated BM cells. (C) Survival curve of tertiary recipients that received transplantations of serially transplanted BM cells without cotransplantation of freshly isolated BM cells (control, ▪; Ezh2, ). (D) CRI comparing transduced CD45.1⁺GFP⁺ stem cells versus nontransduced CD45.1⁺GFP^- stem cells in primary, secondary, and tertiary recipients (control, ▪; Ezh2, ). (E) CRI comparing transduced CD45.1⁺GFP⁺ stem cells with freshly isolated BM cells after 1, 2, and 3 serial transplantations (control, ▪; Ezh2, ).

Figure 5.

Variation in CAFC day 35 frequency and Ezh2 expression levels in BXD strains. (A) Variation in HSC frequency (CAFC d35/10⁵ BM cells) in BXD recombinant inbred mice is associated with a QTL mapping to chromosome 18. Data can be retrieved from www.genenetwork.org., (B) Ezh2 transcript levels were measured in LSK cells isolated from the BM of 30 BXD strains, using Affymetrix gene chips. (C) Variation in Ezh2 expression in LSK cells isolated from BXD mice is regulated by a QTL mapping to chromosome 18. (D). Average Ezh2 expression levels (+ 1 SEM) in BXD recombinant inbred mice segregated according to the presence of the B6 (▪) or D2 () allele for marker D18Mit83.

Figure 6.

Putative new stem cell targets or partners of Ezh2. (A) The top 100 stem cell genes whose expression correlated with Ezh2 expression in BXD mice were classified according to function by WebGestalt (http://genereg.ornl.gov/webgestalt/). □, the expected number of genes with a certain function to be found if 100 transcripts were randomly selected; ▪, the actually retrieved numbers. Detailed information on genes present in the 3 significantly (P < .01) enriched clusters (arrows) is provided in Table 1. (B) Correlation of relative expression of Ezh2 and Eed in the 30 BXD strains. Each data point in the figure refers to an individual BXD strain. (C) Correlation of relative expression of Ezh2 and Geminin (Gmnn) in the 30 BXD strains. Each data point in the figure refers to an individual BXD strain.