The unfolded protein response governs integrity of the haematopoietic stem-cell pool during stress

Journal name:
Nature
Volume:
510,
Pages:
268–272
Date published:
DOI:
doi:10.1038/nature13228
Received
Accepted
Published online
Corrected online

The blood system is sustained by a pool of haematopoietic stem cells (HSCs) that are long-lived due to their capacity for self-renewal. A consequence of longevity is exposure to stress stimuli including reactive oxygen species (ROS), nutrient fluctuation and DNA damage1, 2. Damage that occurs within stressed HSCs must be tightly controlled to prevent either loss of function or the clonal persistence of oncogenic mutations that increase the risk of leukaemogenesis3, 4. Despite the importance of maintaining cell integrity throughout life, how the HSC pool achieves this and how individual HSCs respond to stress remain poorly understood. Many sources of stress cause misfolded protein accumulation in the endoplasmic reticulum (ER), and subsequent activation of the unfolded protein response (UPR) enables the cell to either resolve stress or initiate apoptosis5, 6. Here we show that human HSCs are predisposed to apoptosis through strong activation of the PERK branch of the UPR after ER stress, whereas closely related progenitors exhibit an adaptive response leading to their survival. Enhanced ER protein folding by overexpression of the co-chaperone ERDJ4 (also called DNAJB9) increases HSC repopulation capacity in xenograft assays, linking the UPR to HSC function. Because the UPR is a focal point where different sources of stress converge, our study provides a framework for understanding how stress signalling is coordinated within tissue hierarchies and integrated with stemness. Broadly, these findings reveal that the HSC pool maintains clonal integrity by clearance of individual HSCs after stress to prevent propagation of damaged stem cells.

At a glance

Figures

  1. Elevated expression of PERK branch genes of the UPR in HSCs compared to progenitors and further amplification after tunicamycin-induced stress.
    Figure 1: Elevated expression of PERK branch genes of the UPR in HSCs compared to progenitors and further amplification after tunicamycin-induced stress.

    a, Forty UPR-related genes from the nodes in Extended Data Fig. 1a showed differential expression between HSCs and progenitors (false discovery rate (FDR) <0.05). CMP, common myeloid progenitor; GMP, granulocyte macrophage progenitor; MEP, megakaryocyte erythrocyte progenitor; MPP, multipotent progenitor; MLP, multilymphoid progenitor. b, Expression of key UPR genes in HSPC and progenitor fractions was measured by qPCR. Results are shown as mean±s.e.m. of n = 6 cord blood samples. c, d, UPR branch activation depends on cell type and stressor. Sorted HSPCs or progenitors were plated in the presence of thapsigargin (c) or tunicamycin (d). RNA was isolated at different time points to measure gene expression by qPCR. DMSO controls were the same between c and d. Data are shown as mean±s.e.m. of n = 3 cord blood samples; P value was calculated based on treated/control cells and indicates differential response between HSPCs and progenitors. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

  2. HSCs are predisposed to apoptosis compared to progenitors after treatment with the ER stress agent tunicamycin.
    Figure 2: HSCs are predisposed to apoptosis compared to progenitors after treatment with the ER stress agent tunicamycin.

    a, b, Lower survival of cord-blood- and bone-marrow-derived HSPCs compared to progenitors in the presence of tunicamycin. Sorted HSCs/HSPCs and progenitors were plated with 0.6 μgml−1 tunicamycin (a) or 3µgml−1 tunicamycin (b). Symbols represent viable cell counts of individual samples where populations are connected by a black line; the blue line indicates mean±s.e.m. of n = 16 cord blood samples and n = 5 bone marrow samples (a) or n = 7 cord blood samples and n = 5 bone marrow samples (b). c, Reduced clonogenic potential of HSCs compared to progenitors following tunicamycin treatment. HSCs or progenitors were sorted into methylcellulose containing DMSO or tunicamycin. Data are shown as mean±s.e.m. of n = 4 cord blood samples; NS, not significant. d, Tunicamycin treatment causes higher apoptosis in HSCs compared to progenitors. Cord blood cells were plated with tunicamycin and stained for primitive surface markers and Annexin-V/Sytox. Quantification of viable cells is shown as mean±s.e.m. of n = 5 cord blood samples. P values indicate different viability between HSCs and progenitors. *P<0.05, **P<0.01, ****P<0.0001.

  3. HSCs are predisposed to UPR-induced apoptosis through PERK-eIF2[agr]-ATF4-CHOP-GADD34 signalling.
    Figure 3: HSCs are predisposed to UPR-induced apoptosis through PERK–eIF2α–ATF4–CHOP–GADD34 signalling.

    a, Bidirectional lentiviral reporter vector. All transduced cells are marked by TagBFP; GFP brightness measures the ATF4 mRNA translation rate, which is regulated by upstream open reading frames (uORFs) and depends on peIF2α (ref. 13). b, c, Higher ATF4 reporter activity in HSPCs compared to progenitors. HSPCs and progenitors were sorted, transduced with the ATF4 reporter and treated with tunicamycin. b, Representative flow plots outline calculation of the transgene ratio. MFI, mean fluorescence intensity. c, Results are shown as mean±s.e.m. of n = 4 cord blood samples. d, Constitutively active GADD34OE has a more pronounced effect on HSPCs compared to progenitors. Transduced HSPCs (left) or progenitors (right) were treated with tunicamycin. Symbols represent n = 3 cord blood samples where control (ctrl) and constitutively active GADD34OE (ca-GADD34OE) groups are connected by a black line; P values were calculated using paired t-tests. e, f, Modulating the PERK pathway rescues HSCs from apoptosis. HSPCs and progenitors were plated with tunicamycin and the GADD34 inhibitor salubrinal (e) or the PERK inhibitor GSK2606414 (f). Viability is shown as mean±s.e.m. of n = 5 (e) or n = 4 (f) cord blood samples. *P<0.05, **P<0.01, ***P<0.001.

  4. ERDJ4OE protects from tunicamycin-induced apoptosis and increases HSC output and frequency.
    Figure 4: ERDJ4OE protects from tunicamycin-induced apoptosis and increases HSC output and frequency.

    a, ERDJ4 expression in sorted cord blood populations (Extended Data Table 1). P values were calculated in comparison to CD49f+ HSCs; qPCR results are shown as mean±s.e.m. of n = 3 cord blood samples. B/NK, B and NK cell progenitor. b, c, Validation of lentiviral vectors for ERDJ4OE. b, Transduced cord blood cells were analysed by qPCR. PGK and SFFV refer to lentiviral promoter driving ERDJ4 expression. Results are shown as mean±s.e.m. of n = 2 cord blood samples. c, Transduced K562 cells were analysed by western blot. ERK2 is shown as a loading control. d, ERDJ4OE protects from tunicamycin-induced apoptosis. Transduced HSPCs were treated with tunicamycin. Symbols represent n = 11 cord blood samples where control and ERDJ4OE groups are connected by a grey line; black line indicates mean ± s.e.m.; P value was calculated using a paired t-test. e, ERDJ4OE confers a competitive advantage in vivo. Engraftment of transduced cord blood cells was analysed 20weeks after injection. Data are shown as mean±s.e.m. of n = 3 cord blood samples with 5 mice per group. f, g, ERDJ4OE increases the number of engrafting HSCs. f, Engraftment of transduced cord blood cells was analysed 10weeks after injection of different cell doses. P value was calculated using the Mann–Whitney U-test. Every symbol represents one mouse; data represent n = 3 cord blood samples with 4 mice per group; line shows median. g, HSC frequency was calculated on the basis of the number of engrafted mice. h, Progenitors retain normal engraftment capacity with ERDJ4OE. Engraftment of transduced progenitors was assessed 2 and 4weeks after injection. Data represents n = 3 cord blood samples with 3 mice per group; line shows median. i, ERDJ4OE moderates a surge in CHOP and GADD34 expression after transplantation. Transduced cord blood cells were expanded in vitro and transplanted into mice. CHOP and GADD34 expression in sorted GFP+ cells was analysed by qPCR. Data are shown as mean±s.e.m. of n = 3 cord blood samples. *P<0.05, **P<0.01, ****P<0.0001.

  5. Expression analysis of UPR-related genes.
    Extended Data Fig. 1: Expression analysis of UPR-related genes.

    a, Enrichment of UPR-related genes in human HSCs compared to progenitors. CD49f+ HSC-enriched genes were analysed for GO category overrepresentation. Node size represents the number of genes; white, yellow and orange colour correspond to FDR <0.15, <0.1 and <0.01. b, Simplified scheme illustrating UPR signalling events. Three branches of the UPR are activated upon ER stress: IRE1, PERK and ATF6. IRE1 splices cytosolic XBP1 mRNA to enable translation of the XBP1s transcription factor, which upregulates chaperones and ER-associated degradation (ERAD) machinery to resolve ER stress37, 38. PERK initiates a different branch of the UPR through phosphorylation of eIF2α, which attenuates global protein synthesis, thus permitting time to restore ER homeostasis39. Prolonged ER stress leads to PERK signalling-mediated upregulation of the proapoptotic transcription factor CHOP and its target GADD34. GADD34 dephosphorylates eIF2α leading to restoration of global protein translation. However, if ER stress is not resolved, GADD34 upregulation can lead to further accumulation of misfolded proteins, oxidative stress and apoptosis16. Yellow highlighted arrows indicate transcriptional regulation. c, Amplification curves of qPCR reactions for UPR-related genes. Fluorescence signal during 40 cycles of qPCR reactions on cord-blood-derived cDNA is shown for a representative experiment. Green line indicates threshold that was used to calculate mRNA quantity. d, Dissociation curves were generated to check for the presence of aspecific amplicons or primer dimers, which would be visible as additional peaks. Each line represents the dissociation curve of one qPCR reaction, colours indicate different genes. e, Slopes and R2 values of standard curves are shown for a representative experiment. These values were calculated separately for each experiment, based on a cDNA dilution series. ce were performed using SDS v2.3 software. f, Agarose gel analysis of qPCR amplicons. qPCR reactions were run on a 3% agarose gel to check for reaction specificity: nonspecific amplicons would be visible as additional bands. The expected product size is shown above the gel; the ladder sizes are indicated on the right. g, Adult bone marrow cells were sorted into HSPC and progenitor fractions. mRNA levels for CHOP and ERDJ4 were measured by qPCR. Results are shown as mean±s.e.m. of n = 5 bone marrow samples. ****P<0.0001.

  6. Differential response of HSPCs and progenitors to ER-stress-inducing agents.
    Extended Data Fig. 2: Differential response of HSPCs and progenitors to ER-stress-inducing agents.

    a, b, HSPC and progenitor fractions were sorted and plated in the presence of (a) thapsigargin or (b) tunicamycin. mRNA was isolated after 0.5, 1, 6, 16 and 40h and expression levels of GRP78, ERDJ4, GADD34 and ATF4 were assessed by qPCR. The DMSO-treated controls were the same between a and b. Data are shown as mean±s.e.m. of n = 3 cord blood samples, P value was calculated based on fold change of treated over DMSO control cells and indicates differential response between HSPCs and progenitors. c, Adult bone marrow HSPCs and progenitors were sorted and plated in the presence of tunicamycin. After 16h, mRNA was isolated and expression levels of CHOP, ERDJ4 and GRP94 were assessed by qPCR. Data are shown as mean±s.e.m. of n = 5 bone marrow samples.

  7. Survival of HSCs is lower compared to progenitors after tunicamycin, but not thapsigargin treatment.
    Extended Data Fig. 3: Survival of HSCs is lower compared to progenitors after tunicamycin, but not thapsigargin treatment.

    a, Thapsigargin has similar toxicity for sorted HSC and progenitor fractions. Sorted HSCs and progenitors were plated in TSGF6 culture conditions in the presence of thapsigargin or DMSO control. Symbols represent viable cell counts of individual samples where fractions are connected by a black line; the blue line indicates mean±s.e.m. of n = 7 cord blood samples. b, c, Reduced clonogenic capacity of sorted HSCs compared to progenitors after tunicamycin treatment. Total colony number is shown in Fig. 2c. Here, data are separated into colony types based on morphological appearance. Data are shown as mean±s.e.m. of n = 4 cord blood samples. G, granulocyte; M, macrophage; GM, granulocyte/macrophage; BFU, erythroid burst forming unit; mix, multilineage. d, HSCs have lower survival compared to progenitors after tunicamycin treatment, even after cell cycle induction. Sorted HSC and progenitor fractions were plated in TSGF6 culture conditions with double cytokine concentrations for 72–96h to induce G0 exit of the HSC fraction33. Then, cells were plated in the presence of tunicamycin. Viable cell counts as a percentage of DMSO controls are shown. Symbols represent individual samples where fractions are connected by a black line; the blue line indicates mean±s.e.m. of n = 5 cord blood samples at 0.6μgml−1 and n = 3 cord blood samples at 3μgml−1 tunicamycin. e, f, Increased apoptosis of HSCs compared to progenitors after tunicamycin treatment. e, Cord blood cells were plated with tunicamycin and stained for primitive surface markers, Annexin-V and Sytox. Representative flow plots are shown. f, Sorted HSCs and progenitors were plated in the presence of tunicamycin. The percentage of viable Annexin-V cells after 40h compared to DMSO controls is shown as mean±s.e.m. of n = 4 cord blood samples. **P<0.01, ***P<0.001, ****P<0.0001.

  8. ATF4 reporter enables visualization of increased ATF4 translation after tunicamycin treatment.
    Extended Data Fig. 4: ATF4 reporter enables visualization of increased ATF4 translation after tunicamycin treatment.

    a, ATF4 reporter validation. Two upstream ORFs (uORFs) that are 5′ of the ATF4 coding sequence in the ATF4 mRNA ensure more efficient translation of ATF4 when eIF2α phosphorylation levels are high13, 14. A bidirectional lentiviral vector was constructed that gives constitutive expression of TagBFP to mark transduced cells. In the other direction, the SFFV promoter drives expression of the 5′ end of the ATF4 mRNA which fuses with a GFP reporter gene 3′ of the termination codon of uORF2. HeLa cells were transduced with pSMALB-ATF4.5rep (referred to as ATF4 reporter) and treated with tunicamycin. After 30h, GFP fluorescence was read out by flow cytometry. Histogram plots show n = 2 technical duplicates (two black lines for DMSO control, two red lines for tunicamycin treatment). b, c, Reporter fluorescence depends on uORFs. HeLa cells were transduced and treated with tunicamycin. As expected, ATF4–GFP translation is (b) repressed in the negative control that has a mutated uORF1 start codon and (c) constitutively high in the positive control with mutated start codons for both uORFs13. Histogram plots show n = 2 technical duplicates. d, ATF4 reporter-transduced cord blood cells were treated with tunicamycin and increasing doses of the PERK inhibitor GSK2606414. The transgene ratio (TGR) is shown as mean±s.e.m. of n = 6 cord blood samples (except at 600nM, n = 3 cord blood samples). *P<0.05, **P<0.01, ****P<0.0001.

  9. Modulation of UPR-associated genes affects haematopoietic stem and progenitor cells in vivo.
    Extended Data Fig. 5: Modulation of UPR-associated genes affects haematopoietic stem and progenitor cells in vivo.

    a, Analysis of haematopoietic stem and progenitor cell frequencies in Chop–/–mice. Flow cytometry was performed on mouse bone marrow (Extended Data Table 2). Bars show the absolute cell production in each population from wild-type or Chop–/– mice. Data are shown as mean±s.d. of n = 5 mice per group. b, Viability analysis of stem and progenitor cell populations in Chop–/– mice. The percentage of viable Annexin-V7-AAD cells within the HSC-enriched LSK and LinSca-1c-Kit+ progenitor fractions was assessed by flow cytometry. Data are shown as mean±s.d. of technical duplicates of n = 5 mice per group. c, ERDJ4OE cells show increased survival after tunicamycin treatment. The haematopoietic TEX cell line28 was transduced with SFFV-Ctrl or SFFV-ERDJ4OE lentiviral vectors and plated in the presence of 0.6μgml−1 tunicamycin (SFFV refers to lentiviral promoter driving transgene expression). After 48h, the number of transduced cells compared to DMSO-treated controls was determined by automated counting of GFP+ cells. Data are shown as mean±s.d. of n = 3 independent experiments, P value was calculated using a paired t-test. d, Tunicamycin-induced apoptosis is reduced by ERDJ4OE. Cells from c were analysed for Annexin-V and cleaved caspase-3 expression by flow cytometry. Data are shown as mean±s.d. of n = 3 independent experiments, P values were calculated using paired t-tests. e, ERDJ4OE endows cord blood cells with a competitive advantage over untransduced cells. Three cord blood pools (Exp. 1–3) were transduced with PGK-Ctrl or PGK-ERDJ4OE lentiviruses and injected into 5 mice each. Dashed line indicates GFP% after transduction (day 0); solid line indicates median GFP% of the human CD45+ graft in the injected femur of xenografted mice (20weeks). Every symbol represents one mouse. f, Similar expansion of PGK-Ctrl and PGK-ERDJ4OE transduced cord blood cells in vitro. Three cord blood pools (Exp. 1–3) were transduced with PGK-Ctrl or PGK-ERDJ4OE lentivirus and expanded for 10days in liquid culture. Total population doublings of transduced GFP+ cells is shown. g, ERDJ4OE increases HSC output. After liquid culture, GFP+ cells from f were sorted and injected at high and low cell doses, indicated below the x axis. Total human CD45+GFP+ engraftment in the injected femur after 10 weeks is shown. P values were calculated using the Mann–Whitney U-test. Every symbol represents one mouse, line shows median. *P<0.05, **P<0.01, ***P<0.001.

  10. Lineage differentiation, progenitor cell frequencies, homing, and serial transplantability are maintained following ERDJ4OE.
    Extended Data Fig. 6: Lineage differentiation, progenitor cell frequencies, homing, and serial transplantability are maintained following ERDJ4OE.

    a, PGK-ERDJ4OE-transduced cord blood maintains multilineage differentiation potential in vivo. Left: gating scheme to assess differentiation of the human graft in mouse bone marrow. Representative flow plots show quantification of CD45+CD19+ B cells, CD45+CD33+ monocytes and granulocytes, and CD45GlyA+ erythroid cells within the GFP+ graft. Right: the differentiation of transduced cord blood cells was assessed in the peripheral blood (PB) at 10 and 20weeks and in the injected (RF) and non-injected (LF) femur at 20weeks after transplantation. Results are shown as mean±s.e.m. of n = 15 mice representing n = 3 cord blood samples. b, ERDJ4OE does not cause aberrant expansion of stem or progenitor cell fractions. To assess the distribution of human stem and progenitor cells, lineage+ and mouse cells were depleted from xenografted mouse bone marrow. The remaining human lineage cells were analysed by flow cytometry. Left: gating scheme to assess differentiation into HSC, MPP, MLP, CMP/MEP and B/NK/GMP fractions (Extended Data Table 1). Right: the frequency of human stem and progenitor cells within the human CD45+GFP+ graft was assessed 20weeks after transplantation of transduced cord blood cells. Results are shown as mean±s.e.m. of n = 3 cord blood samples. c, Homing capacity to the non-injected bone marrow is not altered by ERDJ4OE. Transduced cord blood cells were expanded for 12days in liquid culture conditions and 1–1.6×106 cells were transplanted per mouse. After 19h, mice were euthanized to assess human CD45+GFP+ cell homing to the non-injected femur. Results were normalized to transduction efficiency. Every symbol represents one mouse, results of n = 3 cord blood samples are shown with 2 mice per group each; line shows median. d, Frequency of functional human HSCs in vivo is maintained with ERDJ4OE. Cord blood cells were transduced and injected into primary mice. After 10weeks, mice were killed and transduced GFP+ cells were sorted from their bone marrow. Thirty thousand to one million cells were re-transplanted into secondary mice for serial LDA. After 10weeks, the bone marrow of secondary mice was assessed for human CD45+GFP+ engraftment; mice were scored as positive if the engraftment level was >0.01%. Data from n = 3 cord blood samples was pooled.

Tables

  1. Surface marker phenotypes to separate human stem and progenitor cell subsets
    Extended Data Table 1: Surface marker phenotypes to separate human stem and progenitor cell subsets
  2. Surface marker phenotypes to separate mouse stem and progenitor cell subsets
    Extended Data Table 2: Surface marker phenotypes to separate mouse stem and progenitor cell subsets
  3. Primer sequences used for quantitative RT-PCR
    Extended Data Table 3: Primer sequences used for quantitative RT–PCR

Change history

Corrected online 11 June 2014
Superscript plus and minus symbols were incorrect in the labels for Fig. 3d and have been fixed.

References

  1. Geiger, H., de Haan, G. & Florian, M. C. The ageing haematopoietic stem cell compartment. Nature Rev. Immunol. 13, 376389 (2013)
  2. Yahata, T. et al. Accumulation of oxidative DNA damage restricts the self-renewal capacity of human hematopoietic stem cells. Blood 118, 29412950 (2011)
  3. Shlush, L. I. et al. Identification of pre-leukaemic haematopoietic stem cells in acute leukaemia. Nature 506, 328333 (2014)
  4. Rossi, D. J., Jamieson, C. H. M. & Weissman, I. L. Stems cells and the pathways to aging and cancer. Cell 132, 681696 (2008)
  5. Rutkowski, D. T. & Kaufman, R. J. A trip to the ER: coping with stress. Trends Cell Biol. 14, 2028 (2004)
  6. Walter, P. & Ron, D. The unfolded protein response: from stress pathway to homeostatic regulation. Science 334, 10811086 (2011)
  7. Majeti, R., Park, C. Y. & Weissman, I. L. Identification of a hierarchy of multipotent hematopoietic progenitors in human cord blood. Cell Stem Cell 1, 635645 (2007)
  8. Doulatov, S. et al. Revised map of the human progenitor hierarchy shows the origin of macrophages and dendritic cells in early lymphoid development. Nature Immunol. 11, 585593 (2010)
  9. Notta, F. et al. Isolation of single human hematopoietic stem cells capable of long-term multilineage engraftment. Science 333, 218221 (2011)
  10. Laurenti, E. et al. The transcriptional architecture of early human hematopoiesis identifies multilevel control of lymphoid commitment. Nature Immunol. 14, 756763 (2013)
  11. DuRose, J. B., Tam, A. B. & Niwa, M. Intrinsic capacities of molecular sensors of the unfolded protein response to sense alternate forms of endoplasmic reticulum stress. Mol. Biol. Cell 17, 30953107 (2006)
  12. Tabas, I. & Ron, D. Integrating the mechanisms of apoptosis induced by endoplasmic reticulum stress. Nature Cell Biol. 13, 184190 (2011)
  13. Lu, P. D., Harding, H. P. & Ron, D. Translation reinitiation at alternative open reading frames regulates gene expression in an integrated stress response. J. Cell Biol. 167, 2733 (2004)
  14. Vattem, K. M. & Wek, R. C. Reinitiation involving upstream ORFs regulates ATF4 mRNA translation in mammalian cells. Proc. Natl Acad. Sci. USA 101, 1126911274 (2004)
  15. Palam, L. R., Baird, T. D. & Wek, R. C. Phosphorylation of eIF2 facilitates ribosomal bypass of an inhibitory upstream ORF to enhance CHOP translation. J. Biol. Chem. 286, 1093910949 (2011)
  16. Marciniak, S. J. et al. CHOP induces death by promoting protein synthesis and oxidation in the stressed endoplasmic reticulum. Genes Dev. 18, 30663077 (2004)
  17. Han, J. et al. ER-stress-induced transcriptional regulation increases protein synthesis leading to cell death. Nature Cell Biol. 15, 481490 (2013)
  18. Novoa, I., Zeng, H., Harding, H. P. & Ron, D. Feedback inhibition of the unfolded protein response by GADD34-mediated dephosphorylation of eIF2α. J. Cell Biol. 153, 10111022 (2001)
  19. Boyce, M. et al. A selective inhibitor of eIF2α dephosphorylation protects cells from ER stress. Science 307, 935939 (2005)
  20. Axten, J. M. et al. Discovery of 7-methyl-5-(1-{[3-(trifluoromethyl)phenyl]acetyl}-2,3-dihydro-1H-indol-5-yl)-7H-pyrrolo[2,3-d]pyrimidin-4-amine (GSK2606414), a potent and selective first-in-class inhibitor of protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK). J. Med. Chem. 55, 71937207 (2012)
  21. Shen, Y., Meunier, L. & Hendershot, L. M. Identification and characterization of a novel endoplasmic reticulum (ER) DnaJ homologue, which stimulates ATPase activity of BiP in vitro and is induced by ER stress. J. Biol. Chem. 277, 1594715956 (2002)
  22. Lai, C. W., Otero, J. H., Hendershot, L. M. & Snapp, E. ERdj4 protein is a soluble endoplasmic reticulum (ER) DnaJ family protein that interacts with ER-associated degradation machinery. J. Biol. Chem. 287, 79697978 (2012)
  23. Kurisu, J. et al. MDG1/ERdj4, an ER-resident DnaJ family member, suppresses cell death induced by ER stress. Genes Cells 8, 189202 (2003)
  24. Rouault-Pierre, K. et al. HIF-2α protects human hematopoietic stem/progenitors and acute myeloid leukemic cells from apoptosis induced by endoplasmic reticulum stress. Cell Stem Cell 13, 549563 (2013)
  25. Milyavsky, M. et al. A distinctive DNA damage response in human hematopoietic Stem cells reveals an apoptosis-independent role for p53 in self-renewal. Cell Stem Cell 7, 186197 (2010)
  26. Heijmans, J. et al. ER stress causes rapid loss of intestinal epithelial stemness through activation of the unfolded protein response. Cell Rep. 3, 11281139 (2013)
  27. Wey, S., Luo, B. & Lee, A. S. Acute inducible ablation of GRP78 reveals its role in hematopoietic stem cell survival, lymphogenesis and regulation of stress signaling. PLoS ONE 7, e39047 (2012)
  28. Warner, J. K. et al. Direct evidence for cooperating genetic events in the leukemic transformation of normal human hematopoietic cells. Leukemia 19, 17941805 (2005)
  29. Amendola, M., Venneri, M. A., Biffi, A., Vigna, E. & Naldini, L. Coordinate dual-gene transgenesis by lentiviral vectors carrying synthetic bidirectional promoters. Nature Biotechnol. 23, 108116 (2005)
  30. The MGC Project Team The status, quality, and expansion of the NIH full-length cDNA project: the Mammalian Gene Collection (MGC). Genome Res. 14, 21212127 (2004)
  31. Gentner, B. et al. Stable knockdown of microRNA in vivo by lentiviral vectors. Nature Methods 6, 6366 (2009)
  32. Subach, O. M. et al. Conversion of red fluorescent protein into a bright blue probe. Chem. Biol. 15, 11161124 (2008)
  33. van Galen, P. et al. Reduced lymphoid lineage priming promotes human hematopoietic stem cell expansion. Cell Stem Cell 14, 94106 (2014)
  34. Hu, Y. & Smyth, G. K. ELDA: extreme limiting dilution analysis for comparing depleted and enriched populations in stem cell and other assays. J. Immunol. Methods 347, 7078 (2009)
  35. Maere, S., Heymans, K. & Kuiper, M. BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics 21, 34483449 (2005)
  36. Lopes, C. T. et al. Cytoscape Web: an interactive web-based network browser. Bioinformatics 26, 23472348 (2010)
  37. Yoshida, H., Matsui, T., Yamamoto, A., Okada, T. & Mori, K. XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor. Cell 107, 881891 (2001)
  38. Lin, J. H. et al. IRE1 signaling affects cell fate during the unfolded protein response. Science 318, 944949 (2007)
  39. Harding, H. P., Zhang, Y. & Ron, D. Protein translation and folding are coupled by an endoplasmic-reticulum-resident kinase. Nature 397, 271274 (1999)

Download references

Author information

Affiliations

  1. Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario M5G 2M9, Canada

    • Peter van Galen,
    • Antonija Kreso,
    • Nathan Mbong,
    • Stephanie Xie,
    • Elisa Laurenti,
    • Karin Hermans,
    • Erno Wienholds &
    • John E. Dick
  2. Department of Molecular Genetics, University of Toronto, Toronto, Ontario M5S 1A8, Canada

    • Peter van Galen,
    • Antonija Kreso,
    • Nathan Mbong,
    • Stephanie Xie,
    • Elisa Laurenti,
    • Karin Hermans,
    • Erno Wienholds &
    • John E. Dick
  3. Cambridge Institute for Medical Research, Wellcome Trust/MRC Stem Cell Institute and Department of Haematology, University of Cambridge, Cambridge CB2 0XY, UK

    • David G. Kent &
    • Anthony R. Green
  4. Department of Medicine, School of Clinical Medicine, Addenbrookes Hospital, University of Cambridge, Cambridge CB2 0QQ, UK

    • Timothy Fitzmaurice &
    • Jane C. Goodall
  5. Cambridge Institute for Medical Research, Wellcome Trust/MRC Stem Cell Institute and Department of Medicine, University of Cambridge, Cambridge CB2 0XY, UK

    • Joseph E. Chambers &
    • Stefan J. Marciniak
  6. Department of Pediatrics, McGill University and the Research Institute of the McGill University Health Centre, Westmount, Québec H3Z 2Z3, Canada

    • Kolja Eppert
  7. Departments of Radiation Oncology and Medical Biophysics, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario M5G 2M9, Canada

    • Bradly G. Wouters

Contributions

P.v.G., A.K. and J.E.D. designed the study and analysed and interpreted the data. P.v.G., A.K., N.M., D.G.K., T.F., J.E.C., S.X., K.H. and E.W. performed experiments. E.L. and K.E. performed bioinformatic analyses. S.J.M., J.C.G., A.R.G. and B.G.W. supervised specific experiments. P.v.G. wrote the paper. A.K. and J.E.D. revised the paper. J.E.D. supervised the study.

Competing financial interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to:

Author details

Extended data figures and tables

Extended Data Figures

  1. Extended Data Figure 1: Expression analysis of UPR-related genes. (282 KB)

    a, Enrichment of UPR-related genes in human HSCs compared to progenitors. CD49f+ HSC-enriched genes were analysed for GO category overrepresentation. Node size represents the number of genes; white, yellow and orange colour correspond to FDR <0.15, <0.1 and <0.01. b, Simplified scheme illustrating UPR signalling events. Three branches of the UPR are activated upon ER stress: IRE1, PERK and ATF6. IRE1 splices cytosolic XBP1 mRNA to enable translation of the XBP1s transcription factor, which upregulates chaperones and ER-associated degradation (ERAD) machinery to resolve ER stress37, 38. PERK initiates a different branch of the UPR through phosphorylation of eIF2α, which attenuates global protein synthesis, thus permitting time to restore ER homeostasis39. Prolonged ER stress leads to PERK signalling-mediated upregulation of the proapoptotic transcription factor CHOP and its target GADD34. GADD34 dephosphorylates eIF2α leading to restoration of global protein translation. However, if ER stress is not resolved, GADD34 upregulation can lead to further accumulation of misfolded proteins, oxidative stress and apoptosis16. Yellow highlighted arrows indicate transcriptional regulation. c, Amplification curves of qPCR reactions for UPR-related genes. Fluorescence signal during 40 cycles of qPCR reactions on cord-blood-derived cDNA is shown for a representative experiment. Green line indicates threshold that was used to calculate mRNA quantity. d, Dissociation curves were generated to check for the presence of aspecific amplicons or primer dimers, which would be visible as additional peaks. Each line represents the dissociation curve of one qPCR reaction, colours indicate different genes. e, Slopes and R2 values of standard curves are shown for a representative experiment. These values were calculated separately for each experiment, based on a cDNA dilution series. ce were performed using SDS v2.3 software. f, Agarose gel analysis of qPCR amplicons. qPCR reactions were run on a 3% agarose gel to check for reaction specificity: nonspecific amplicons would be visible as additional bands. The expected product size is shown above the gel; the ladder sizes are indicated on the right. g, Adult bone marrow cells were sorted into HSPC and progenitor fractions. mRNA levels for CHOP and ERDJ4 were measured by qPCR. Results are shown as mean±s.e.m. of n = 5 bone marrow samples. ****P<0.0001.

  2. Extended Data Figure 2: Differential response of HSPCs and progenitors to ER-stress-inducing agents. (355 KB)

    a, b, HSPC and progenitor fractions were sorted and plated in the presence of (a) thapsigargin or (b) tunicamycin. mRNA was isolated after 0.5, 1, 6, 16 and 40h and expression levels of GRP78, ERDJ4, GADD34 and ATF4 were assessed by qPCR. The DMSO-treated controls were the same between a and b. Data are shown as mean±s.e.m. of n = 3 cord blood samples, P value was calculated based on fold change of treated over DMSO control cells and indicates differential response between HSPCs and progenitors. c, Adult bone marrow HSPCs and progenitors were sorted and plated in the presence of tunicamycin. After 16h, mRNA was isolated and expression levels of CHOP, ERDJ4 and GRP94 were assessed by qPCR. Data are shown as mean±s.e.m. of n = 5 bone marrow samples.

  3. Extended Data Figure 3: Survival of HSCs is lower compared to progenitors after tunicamycin, but not thapsigargin treatment. (216 KB)

    a, Thapsigargin has similar toxicity for sorted HSC and progenitor fractions. Sorted HSCs and progenitors were plated in TSGF6 culture conditions in the presence of thapsigargin or DMSO control. Symbols represent viable cell counts of individual samples where fractions are connected by a black line; the blue line indicates mean±s.e.m. of n = 7 cord blood samples. b, c, Reduced clonogenic capacity of sorted HSCs compared to progenitors after tunicamycin treatment. Total colony number is shown in Fig. 2c. Here, data are separated into colony types based on morphological appearance. Data are shown as mean±s.e.m. of n = 4 cord blood samples. G, granulocyte; M, macrophage; GM, granulocyte/macrophage; BFU, erythroid burst forming unit; mix, multilineage. d, HSCs have lower survival compared to progenitors after tunicamycin treatment, even after cell cycle induction. Sorted HSC and progenitor fractions were plated in TSGF6 culture conditions with double cytokine concentrations for 72–96h to induce G0 exit of the HSC fraction33. Then, cells were plated in the presence of tunicamycin. Viable cell counts as a percentage of DMSO controls are shown. Symbols represent individual samples where fractions are connected by a black line; the blue line indicates mean±s.e.m. of n = 5 cord blood samples at 0.6μgml−1 and n = 3 cord blood samples at 3μgml−1 tunicamycin. e, f, Increased apoptosis of HSCs compared to progenitors after tunicamycin treatment. e, Cord blood cells were plated with tunicamycin and stained for primitive surface markers, Annexin-V and Sytox. Representative flow plots are shown. f, Sorted HSCs and progenitors were plated in the presence of tunicamycin. The percentage of viable Annexin-V cells after 40h compared to DMSO controls is shown as mean±s.e.m. of n = 4 cord blood samples. **P<0.01, ***P<0.001, ****P<0.0001.

  4. Extended Data Figure 4: ATF4 reporter enables visualization of increased ATF4 translation after tunicamycin treatment. (338 KB)

    a, ATF4 reporter validation. Two upstream ORFs (uORFs) that are 5′ of the ATF4 coding sequence in the ATF4 mRNA ensure more efficient translation of ATF4 when eIF2α phosphorylation levels are high13, 14. A bidirectional lentiviral vector was constructed that gives constitutive expression of TagBFP to mark transduced cells. In the other direction, the SFFV promoter drives expression of the 5′ end of the ATF4 mRNA which fuses with a GFP reporter gene 3′ of the termination codon of uORF2. HeLa cells were transduced with pSMALB-ATF4.5rep (referred to as ATF4 reporter) and treated with tunicamycin. After 30h, GFP fluorescence was read out by flow cytometry. Histogram plots show n = 2 technical duplicates (two black lines for DMSO control, two red lines for tunicamycin treatment). b, c, Reporter fluorescence depends on uORFs. HeLa cells were transduced and treated with tunicamycin. As expected, ATF4–GFP translation is (b) repressed in the negative control that has a mutated uORF1 start codon and (c) constitutively high in the positive control with mutated start codons for both uORFs13. Histogram plots show n = 2 technical duplicates. d, ATF4 reporter-transduced cord blood cells were treated with tunicamycin and increasing doses of the PERK inhibitor GSK2606414. The transgene ratio (TGR) is shown as mean±s.e.m. of n = 6 cord blood samples (except at 600nM, n = 3 cord blood samples). *P<0.05, **P<0.01, ****P<0.0001.

  5. Extended Data Figure 5: Modulation of UPR-associated genes affects haematopoietic stem and progenitor cells in vivo. (371 KB)

    a, Analysis of haematopoietic stem and progenitor cell frequencies in Chop–/–mice. Flow cytometry was performed on mouse bone marrow (Extended Data Table 2). Bars show the absolute cell production in each population from wild-type or Chop–/– mice. Data are shown as mean±s.d. of n = 5 mice per group. b, Viability analysis of stem and progenitor cell populations in Chop–/– mice. The percentage of viable Annexin-V7-AAD cells within the HSC-enriched LSK and LinSca-1c-Kit+ progenitor fractions was assessed by flow cytometry. Data are shown as mean±s.d. of technical duplicates of n = 5 mice per group. c, ERDJ4OE cells show increased survival after tunicamycin treatment. The haematopoietic TEX cell line28 was transduced with SFFV-Ctrl or SFFV-ERDJ4OE lentiviral vectors and plated in the presence of 0.6μgml−1 tunicamycin (SFFV refers to lentiviral promoter driving transgene expression). After 48h, the number of transduced cells compared to DMSO-treated controls was determined by automated counting of GFP+ cells. Data are shown as mean±s.d. of n = 3 independent experiments, P value was calculated using a paired t-test. d, Tunicamycin-induced apoptosis is reduced by ERDJ4OE. Cells from c were analysed for Annexin-V and cleaved caspase-3 expression by flow cytometry. Data are shown as mean±s.d. of n = 3 independent experiments, P values were calculated using paired t-tests. e, ERDJ4OE endows cord blood cells with a competitive advantage over untransduced cells. Three cord blood pools (Exp. 1–3) were transduced with PGK-Ctrl or PGK-ERDJ4OE lentiviruses and injected into 5 mice each. Dashed line indicates GFP% after transduction (day 0); solid line indicates median GFP% of the human CD45+ graft in the injected femur of xenografted mice (20weeks). Every symbol represents one mouse. f, Similar expansion of PGK-Ctrl and PGK-ERDJ4OE transduced cord blood cells in vitro. Three cord blood pools (Exp. 1–3) were transduced with PGK-Ctrl or PGK-ERDJ4OE lentivirus and expanded for 10days in liquid culture. Total population doublings of transduced GFP+ cells is shown. g, ERDJ4OE increases HSC output. After liquid culture, GFP+ cells from f were sorted and injected at high and low cell doses, indicated below the x axis. Total human CD45+GFP+ engraftment in the injected femur after 10 weeks is shown. P values were calculated using the Mann–Whitney U-test. Every symbol represents one mouse, line shows median. *P<0.05, **P<0.01, ***P<0.001.

  6. Extended Data Figure 6: Lineage differentiation, progenitor cell frequencies, homing, and serial transplantability are maintained following ERDJ4OE. (497 KB)

    a, PGK-ERDJ4OE-transduced cord blood maintains multilineage differentiation potential in vivo. Left: gating scheme to assess differentiation of the human graft in mouse bone marrow. Representative flow plots show quantification of CD45+CD19+ B cells, CD45+CD33+ monocytes and granulocytes, and CD45GlyA+ erythroid cells within the GFP+ graft. Right: the differentiation of transduced cord blood cells was assessed in the peripheral blood (PB) at 10 and 20weeks and in the injected (RF) and non-injected (LF) femur at 20weeks after transplantation. Results are shown as mean±s.e.m. of n = 15 mice representing n = 3 cord blood samples. b, ERDJ4OE does not cause aberrant expansion of stem or progenitor cell fractions. To assess the distribution of human stem and progenitor cells, lineage+ and mouse cells were depleted from xenografted mouse bone marrow. The remaining human lineage cells were analysed by flow cytometry. Left: gating scheme to assess differentiation into HSC, MPP, MLP, CMP/MEP and B/NK/GMP fractions (Extended Data Table 1). Right: the frequency of human stem and progenitor cells within the human CD45+GFP+ graft was assessed 20weeks after transplantation of transduced cord blood cells. Results are shown as mean±s.e.m. of n = 3 cord blood samples. c, Homing capacity to the non-injected bone marrow is not altered by ERDJ4OE. Transduced cord blood cells were expanded for 12days in liquid culture conditions and 1–1.6×106 cells were transplanted per mouse. After 19h, mice were euthanized to assess human CD45+GFP+ cell homing to the non-injected femur. Results were normalized to transduction efficiency. Every symbol represents one mouse, results of n = 3 cord blood samples are shown with 2 mice per group each; line shows median. d, Frequency of functional human HSCs in vivo is maintained with ERDJ4OE. Cord blood cells were transduced and injected into primary mice. After 10weeks, mice were killed and transduced GFP+ cells were sorted from their bone marrow. Thirty thousand to one million cells were re-transplanted into secondary mice for serial LDA. After 10weeks, the bone marrow of secondary mice was assessed for human CD45+GFP+ engraftment; mice were scored as positive if the engraftment level was >0.01%. Data from n = 3 cord blood samples was pooled.

Extended Data Tables

  1. Extended Data Table 1: Surface marker phenotypes to separate human stem and progenitor cell subsets (158 KB)
  2. Extended Data Table 2: Surface marker phenotypes to separate mouse stem and progenitor cell subsets (91 KB)
  3. Extended Data Table 3: Primer sequences used for quantitative RT–PCR (429 KB)

Additional data