CCND1-CDK4-mediated cell cycle progression provides a competitive advantage for human hematopoietic stem cells in vivo

Abstract

Maintenance of stem cell properties is associated with reduced proliferation. However, in mouse hematopoietic stem cells (HSCs), loss of quiescence results in a wide range of phenotypes, ranging from functional failure to extensive self-renewal. It remains unknown whether the function of human HSCs is controlled by the kinetics of cell cycle progression. Using human HSCs and human progenitor cells (HSPCs), we report here that elevated levels of CCND1-CDK4 complexes promoted the transit from G0 to G1 and shortened the G1 cell cycle phase, resulting in protection from differentiation-inducing signals in vitro and increasing human leukocyte engraftment in vivo. Further, CCND1-CDK4 overexpression conferred a competitive advantage without impacting HSPC numbers. In contrast, accelerated cell cycle progression mediated by elevated levels of CCNE1-CDK2 led to the loss of functional HSPCs in vivo. Collectively, these data suggest that the transition kinetics through the early cell cycle phases are key regulators of human HSPC function and important for lifelong hematopoiesis.

Figure 1.

Elevated CCND1–CDK4 (4D) expression induces a G0-to-G1 transition and increased signaling in human HSPCs. (A) Scheme of the construct: CCND1 and CDK4 are linked by T2A, followed by an IRES-GFP-reporter sequence (4D). The control vector lacks CCND1 and CDK4 sequences (mock). (B) Expression analysis was performed on 4D- or mock-transduced human CD34-enriched cord blood (CB) cells by deep sequencing, and the plot shows the difference of the mean normalized read counts per condition (mock, 4D) of genes related to cell cycle (cyclins, CDKs and CKIs). WThe following criteria were applied to determine differential expression of genes (marked in red): false discovery rate (FDR) of 5% and a log₂ fold change of >1, respectively ≤1. Read counts <10 were considered as background levels (dashed line). Due to usage of codon-optimized sequences CDK4 read counts were annotated to a different gene but read coverage of the codon-optimized (CO) CDK4 sequence revealed its expression in 4D- but not in mock-transduced samples (Fig. S1). Data were pooled from two independent biological replicates using two and three pooled CB samples. Endo, endogenous. (C) Western blot of CDK4 and CCND1 overexpressed in mouse NIH3T3 cells. The antibody recognizing CDK4 is specific for human protein, the antibody recognizing CCND1 is mouse–human cross-reactive. Data shown is representative of three independent experiments. (D) Cell cycle profiles of 4D- and mock-transduced human CD34⁺ CB HSPCs 3 d after transductions. 1 CB sample was used for the shown experiment. (E) Plots show the frequency (left) and MFI (middle left) of phosphorylated Rb at S807/811. Frequencies of cells in S–G2–M phase (middle right), and frequencies of Ki-67 negative cells (right) are shown. Graph bars depict independent values from 6–10 experiments. In two experiments, triplicates were measured to determine S–G2–M phase and Ki-67 expression, and the mean is depicted. Due to paucity of cells in 6 (pRb), 5 (Ki67), and 8 (S–G2–M) experiments, single measurements were performed and shown. Viability of GFP⁺ CD34⁺ HSPCs was determined before each assay using DAPI or Sytox negativity and found >97% in all cases. Two to five pooled CB samples were used for each experiment. (F) Frequencies of cells undergoing DNA fragmentation (sub-G1) are shown (left). Sample preparation as described in E. (right) The frequency of apoptotic cells after transduction with 4D relative to control-treated HSPCs 3 d after transduction. Data are pooled from three independent experiments. Single cord blood sample were used for each experiment. (G) Plots show frequencies of pRb(S807/811)⁺ cells and cells in the S–G2–M phase among GFP⁺ and GFP⁻ cells of the same probe. Sample preparation as described in E. *, P = 0.05–0.01; **, P = 0.01–0.001; ***, P < 0.001.

Figure 2.

4D overexpression shortens G1 cell cycle phase. (A) Phase length calculations in human HSPCs based on cumulative BrdU incorporation 4 d after transduction. Phase lengths were calculated by determining the cell cycle transition rates of equation systems 1 and 2 (see Materials and methods). (top) Simultaneously, the equations of system 1 (lines) were fitted to time-course data of cells gated on BrdU and DNA content (dots), showing (top left) BrdU-positive, (top middle) G1 BrdU-negative, and (top right) G2/M BrdU-negative fractions, as well as (bottom) equations of system 2 (lines) to the same time-course data gated on DNA content only (dots), showing the steady-state distribution of (bottom left) S-phase, (bottom middle) G1-phase, and (bottom right) G2–M-phase fractions. Shaded regions indicated one standard deviation of the estimated measurement noise. (B) Plots show cell cycle phase lengths in human HSPCs after transduction with 4D calculated by the fitted values described in A. 95% confidence intervals are shown. Significant differences (**) between cell cycle parameters were evaluated at a confidence level of 99% by profile likelihood. Data shown is summarized from two independent experiments and two pooled CB samples were used each time. (C) GFP⁺ CD45⁺ (left) and GFP⁺ CD45⁺ CD34⁺ (right) cell numbers after cultivation of 4D- or mock-transduced HSPCs in medium supplemented with hSCF, hTPO, and mFlt3l. Before each plating, viability was measured and found comparable between 4D- and mock-transduced cells (mock: assay start day 4, >98 ± 0.8%; day 8, 95 ± 1.1%; day 13, 94 ± 0.5%; day 16, 97 ± 1.5%; 4D: assay start day 4, 99 ± 0.5%; day 8, 95 ± 1%; day 13, 95 ± 0.8%; day 16, 88 ± 7%). Plot shown is representative of two independent experiments using two or three pooled CB samples each. (D) Serial replating assay performed 3–5 d after transduction (left). Colonies were counted after 14 d and 5 × 10⁴ CD45⁺ cells were replated. Colony types were determined after the first plating round (right). Each data point represents the mean of duplicates or triplicates (n = 7–8) and colony numbers were calculated to an input of 50,000 cells per well. Viability was comparable between mock- and 4D-transduced cells at each time point of analysis. Data are pooled from 7 independent experiments using two or three pooled CB samples. **, P = 0.01–0.001.

Figure 3.

4D-mediated G1 transit improves engraftment in vivo. (A) 2–4 × 10⁵ 4D- or mock-transduced HSPCs were transplanted into unconditioned NSG-Kit^Wv/+ recipient mice (Cosgun et al., 2014) and blood of recipients was periodically analyzed for the frequency of human CD45⁺ leukocytes for up to 28 wk. Data are pooled from three independent experiments; mock, n = 11; 4D, n = 10. Two to three pooled CB samples were used for each experiment. (B) Plot shows GFP expression in human peripheral blood (PB) leukocytes in samples described in A. (C) Dot plots show GFP expression in 4D- or mock-transduced hCD34⁺ donor CB HSPCs (left) and in hCD45⁺ BM cells (right) 28 wk after transplantation. Data are representative for 3 independent experiments using at least 2 pooled CB samples each. (D) Bar graphs show GFP⁺ cells in CD34⁺ donor CB cells (input, left) and in hCD45⁺ BM cells in recipient mice 22–28 wk after transplantation (1° recipient; middle) as outlined in A. Fold-change of GFP contribution to both populations (frequency in 1° recipient relative to CD34⁺ cell input’, right). (E) Dot plots show GFP expression in CD34⁺ CB donor cells before (Input) and 28 wk after transplantation in HSCs (hCD45⁺ CD34⁺ CD38⁻ CD90⁺ CD45RA⁻), HSPCs (hCD45⁺ CD34⁺), mature lymphoid (hCD45⁺ CD3⁺ CD19⁺), and myeloid (hCD45⁺ CD33⁺ and/or hCD45⁺ CD16⁺) cells in the BM of recipient mice. (F) Plot shows changes of GFP⁺ cells in the indicated populations relative to GFP⁺ cells in donor CD34⁺ CB HSPCs. Absolute cell numbers were used for normalization. Samples as described in C. (G and H) Composition of mature donor-derived cells in the BM (G) and spleen (H). GFP⁻ and GFP⁺ fractions from mock- and 4D-transduced donor cells are shown. Samples as described in C. (I) Plot shows changes of GFP⁺ cells in the indicated BM populations relative to GFP⁺ cells in donor CD34⁺ CB HSPCs. MPPs, CD34⁺ CD38⁻ CD90⁻ CD45RA⁻; MLPs, CD34⁺ CD38⁻ CD90^-/lo CD45RA⁺; B/NK/GMPs, CD34⁺ CD38⁺ CD45RA⁺; and CMPs, CD34⁺ CD38⁺ CD45RA⁻ (van Galen et al., 2014). Samples as described in C. (J) BFU-E (left), CFU-G + GM (middle), and CFU-M (right) colony formation of 4D- or mock-transduced sorted GFP⁺ hCD45⁺ BM leukocytes in the presence of hSCF, hGM-CSF, hG-CSF, hIL-3, and hEPO. Colonies were scored after 14 d. Nontransduced GFP⁻ human donor cells were used as control (w/o). Data are summarized from three independent experiments. In two experiments (∆ and ◇), BM cells from recipients of the same donor cells were pooled. Samples as described in C. (K) Frequency of Ki67⁻ cells in mock- or 4D-transduced donor-derived hCD34⁺ GFP⁺ HSPCs 22–28 wk after transplantation (left). Frequency of hCD34⁺ GFP⁺ HSPCs in S–G2–M in the same samples (right). Data are summarized from two independent experiments, mock n = 8, 4D n = 7. (L) Plot shows the expression of the 4D construct in donor-derived hCD34⁺ GFP⁺ HSPCs 22 wk after transplantation in comparison to hCD34⁺ GFP⁻ HSPC competitors. Data are summarized from 3 recipient mice using 3 pooled CB samples. *, P = 0.05–0.01; ***, P < 0.001.

Figure 4.

4D provides a competitive advantage for HSPC engraftment. (A) Plot shows the frequency of successful engrafted secondary recipient mice that had received 4–6 × 10⁶ hCD45⁺ enriched BM cells from primary recipient mice 23–24 wk earlier. Frequencies of hCD45⁺ cells <1% were considered as nonengrafted and excluded from subsequent analysis (dashed line). Data are summarized from two independent experiments using 2 or 3 donor and 3 recipient mice each. (B) GFP expression in hCD45⁺ BM leukocytes in primary (1°) and secondary (2°) recipient mice (dot plots). Bar graphs show the fold-change of GFP frequencies in hCD45⁺ BM cells relative to hCD45⁺ BM cells in the primary recipients (left) and to CD34⁺ CB donor HSPCs (right). Samples as described in A. (C) Dot plots show frequencies of CD34⁺ cells within human lineage negative leukocytes in the BM of secondary recipient mice. Data from one representative experiment are shown. (D) Plot shows changes of GFP⁺ cells in the indicated populations relative to GFP⁺ cells in donor hCD45⁺ CD34⁺ BM cells from primary recipients. Absolute cell numbers were used for normalization. Samples as described in A. (E) Composition of mature donor-derived GFP⁺ and GFP⁻ cells in the BM of secondary recipients. (F) Composition of mature donor-derived GFP⁺ and GFP⁻ cells in the spleen of secondary recipients. (G) Plot shows fold-change difference of human CD34⁺ HSPCs 6 d after cultivation with LPS relative to samples without LPS. Data are pooled from 5 experiments using 2–5 pooled and 1 experiment using 3 individual CB samples. Statistic was calculated based on mean values of each separate CB preparation. Different dot shapes indicate technical replicates from independent experiments. (H) Plot shows the expression of 4D transcript relative to mock-transduced cells before and after cultivation with LPS determined by qPCR. Data are pooled from two independent experiments using two pooled CB samples (triangles) and 3 individual CB samples (circles). (I and J) Plots show maintenance of human CD34⁺ HSPCs after 8 d of cultivation with hGM-CSF (I) or hIL-3 (J). Fold change in CD34⁺ GFP⁺ cells relative to mock control is shown. Data are pooled from three independent experiments using 5 (IL-3) and 7 (GM-CSF) single CB samples.

Figure 5.

Elevated levels of CCNE1–CDK2 (2E) confer a competitive disadvantage to human HSC function in vivo. (A) Plot shows expression levels of the 2E construct relative to mock controls in GFP⁺ CD34⁺ CB samples 3 d after transduction. Data are summarized from two independent experiments using three single CB samples each. (B) Cell cycle profile 3 d after overexpression of mock, 4D, or 2E in human HSPCs gated on GFP⁺ CD34⁺ cells. Data are representative for 15 independent experiments. Single (12 expts) or 2–3 pooled (3 expts) CB samples were used. Data are quantified in (C and D). (C) Frequencies of GFP⁺ CD34⁺ HSPCs containing fragmented DNA (sub-G1, left) and apoptotic cells (right). (D) Frequencies (left) and MFI (second left) of phosphorylated pRb at S807/811 in mock- and 4D-tranduced HSPCs. Frequencies of GFP⁺ CD34⁺ HSPCs in S–G2–M phase (second right), and of Ki-67 negative cells (right). Experiments were done 3–5 d after transduction, and data are pooled from 3 (pRb), 2 (Ki67), and 15 (S–G2–M) independent experiments. Two to three pooled CB samples were used for each experiment. (E) Plots show cell cycle phase lengths in human HSPCs 3 d after overexpression of 2E, as outlined in Fig. 2 B. Data were pooled from two independent experiments and single CB samples were used each time. (F) Donor leukocyte numbers in the BM of NSG-Kit^Wv/+ recipient mice 22–28 wk after transplantation of 2–4 × 10⁵ HSPCs containing 2E- or mock-transduced cells. Data are summarized from two independent experiments using two pooled CB samples for the transplantation into three recipient mice each. (G) Dot plots show 2E-transduced (GFP⁺) human HSPCs in CD34⁺ cells of the transplantation mix before (Input) and in human CD45⁺ BM cells 28 wk after transplantation (1° recipient). Plots are representative for samples as described in F. Bar graphs summarize GFP frequencies in donor CD34⁺ HSPCs (left) and engrafted hCD45⁺ BM cells in primary recipients (middle). n = 6. The fold-change of GFP expression relative to the GFP frequencies in CD34⁺ CB donor cells is shown (right). n = 6. (H) Absolute numbers of GFP⁻ and GFP⁺ engrafted hCD45⁺ CD34⁺ BM HSCPs in the BM of recipient mice. n = 6. (I) Bar graphs show the fold-change in GFP expression in human HSPCs in secondary recipient mice relative to that in hCD45⁺ BM cells in primary recipients (left) and relative to that of CD34⁺ donor CB HSPCs (middle) after transplantation of 4–6 × 10⁶ hCD45⁺ enriched BM cells from primary recipient mice into secondary NSG-Kit^Wv/+ recipients. Cell numbers of human GFP⁺ and GFP⁻ CD34⁺ HSPCs in the BM of secondary recipient mice are shown (right). Data are pooled from two independent experiments with three mice each.