SEARCH

SEARCH BY CITATION

Keywords:

  • Hematopoietic stem cells;
  • In vivo marking;
  • Mice;
  • Lentiviral vector;
  • Adult stem cells;
  • Steady-state hematopoiesis

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

Hematopoietic stem cells (HSCs) generate all mature blood cells during the whole lifespan of an individual. However, the clonal contribution of individual HSC and progenitor cells in steady-state hematopoiesis is poorly understood. To investigate the activity of HSCs under steady-state conditions, murine HSC and progenitor cells were genetically marked in vivo by integrating lentiviral vectors (LVs) encoding green fluorescent protein (GFP). Hematopoietic contribution of individual marked clones was monitored by determination of lentiviral integration sites using highly sensitive linear amplification-mediated-polymerase chain reaction. A remarkably stable small proportion of hematopoietic cells expressed GFP in LV-injected animals for up to 24 months, indicating stable marking of murine steady-state hematopoiesis. Analysis of the lentiviral integration sites revealed that multiple hematopoietic clones with both myeloid and lymphoid differentiation potential contributed to long-term hematopoiesis. In contrast to intrafemoral vector injection, intravenous administration of LV preferentially targeted short-lived progenitor cells. Myelosuppressive treatment of mice prior to LV-injection did not affect the marking efficiency. Our study represents the first continuous analysis of clonal behavior of genetically marked hematopoietic cells in an unmanipulated system, providing evidence that multiple clones are simultaneously active in murine steady-state hematopoiesis. Stem Cells2012;30:1961–1970


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

Blood cell production by individual hematopoietic stem cells (HSCs) and progenitor cells in vivo has been thoroughly characterized by transplanting defined cell populations into syngeneic or xenogeneic hosts and monitoring their mature blood cell output. In contrast, very little is known about the clonal dynamics of the unperturbed hematopoiesis under steady-state conditions. It is unclear whether and how many individual clones permanently contribute to undisturbed hematopoiesis, and whether blood cell production is sustained by successive waves of active HSCs.

Numerous functional studies of HSCs and progenitor cells following transplantation clearly demonstrated heterogeneity within the transplantable stem cell compartment [1]. Long-term stem cells are characterized by extensive self-renewal potential and generate both myeloid and lymphoid progeny for more than 4–6 months after transplantation [2–4]. In contrast, the activity of short-term stem cells is restricted to the first 4 months post-transplantation indicating limited or no self-renewal in vivo [5, 6]. In a hierarchical structure, multipotent long-term HSCs give rise to short-term HSCs and lineage-restricted progenitors that in turn differentiate to mature myeloid or lymphoid blood cells [1]. Early transplantation experiments using genetically marked hematopoietic cells revealed sequential fluctuations of murine short-lived hematopoietic clones supporting a clonal succession model of stem cell activity [7–9]. In this model, at any given time, blood cell production is driven by a low number of HSC clones, whereas most other HSCs are quiescent but are recruited to replace exhausted clones [10]. In retrospect, limitations in the sensitivity of detection of active clones at that time may have contributed to this interpretation. Strong evidence for a lifelong contribution of individual self-renewing HSC to blood cell production after transplantation [11–13] favors an alternative model. Studies in cats [14], mice [15], and immunodeficient mice engrafted with human hematopoietic transplants [16] demonstrated that the initial phase of hematopoietic recovery after transplantation is driven by many short-lived clones, whereas long-term hematopoiesis derives from a smaller number of stable stem cells. Moreover, we and others were able to show by highly sensitive tracking of genetically marked stem cell clones that stable long-term HSCs contribute to highly polyclonal post-transplant hematopoiesis in large animals [17–19]. Computational modeling suggested that stochastic stem cell decisions may explain dynamical changes in the clonal fate of HSC [20]. In contrast, more recent experimental studies of single highly purified long-term HSCs in mice demonstrated distinct classes of HSCs with intrinsically fixed differences in their ability to self-renew and to differentiate into various blood cell lineages [21–23].

Notably, principles of stem cell kinetics measured after transplantation may not be directly assignable to unperturbed hematopoiesis at steady-state conditions. High oxygen concentrations during the transplantation procedure affect the intracellular homeostasis of HSCs. Accumulation of reactive oxygen species, telomere shortening, and proliferative stress during reconstitution significantly reduces the number and self-renewal capacity of quiescent stem cells [24–30]. Moreover, efficient HSC transplantation usually requires myeloablative treatment of the hosts by whole body irradiation or chemotherapy to improve stem cell engraftment, generally thought but not proven by opening up supportive HSC niches in the bone marrow (BM) cavity [31]. Inevitably, irradiation perturbs the BM environment [32] by vascular damage [33] and dramatic changes in the expression levels of growth factors [34] and endothelial adhesion molecules [35]. This may affect the number and cellular composition of subendosteal or perivascular stem cell niches [34, 36], thereby impairing their ability to support hematopoiesis [37].

Toward a better understanding of the clonal activity of adult HSCs and progenitor cells under steady-state conditions, here we used a direct in vivo genetic marking strategy to track the clonal contribution of individual stem and progenitor cells in murine steady-state hematopoiesis by injecting integrating green fluorescent protein (GFP) expressing lentiviral vectors (LVs) directly into the femur or into the tail vein of GFP immunotolerant mice. Monitoring of genetically marked hematopoietic cells for up to 24 months revealed remarkably stable long-term activity of individual multilineage stem cells with a very small clone size in steady-state hematopoiesis.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

LV Production

Replication-defective self-inactivating (SIN) pCCLsin.PPT.SFFV.eGFP.Wpre constructs were used to generate vesicular stomatitis virus (VSV-G)-pseudotyped lentiviral stocks as described with minor modifications [38]. Briefly, LVs were produced by calcium phosphate or polyethylenimine (PEI, Sigma-Aldrich, St. Louis, MO, www.sigmaaldrich.com) -mediated four-plasmid transfection of 293T cells in the presence of chloroquine diphosphate salt or PEI-transfection reagent (1:3 ratio, DNA/PEI). Twenty-four hours after transfection, viral supernatant was collected, concentrated by ultracentrifugation, and stored at −80°C. Titers of vector preparations were determined by transduction of HeLa cells with serial dilutions of vector supernatants followed by cytometric analysis 3 days after transduction. Final vector titers ranged between 108 and 109 transducing units (TU) per milliliter.

Mice

B6.Cg-Tg(Krt1-15-EGFP)2Cot/J and C57BL/6J mice were obtained from the Jackson Laboratory (Bar Harbor, ME, www.jax.org) and raised and housed at the pathogen-free animal facility of the German Cancer Research Center (DKFZ, Heidelberg, Germany) according to applicable laws and regulations following approval by the Institution's Animal Care and Ethical Committee. Mice were maintained in isolators and provided with autoclaved food and water containing 0.5 ml/l baytril (10%, BayerVital, Leverkusen, Germany, www.gesundheit.bayer.de).

In Vivo Marking

For in vivo marking of HSC and progenitor cells, GFP expressing LVs were injected intravenously (IV) or intrafemorally (IF) (n = 25). For IV marking, 200 μl of LV-suspension (2 × 108 TU/ml) were injected into the tail vein of the mice (n = 10). For IF injection, mice were narcotized with 1.7% isoflurane and a 27-gauge needle was inserted into the joint surface of the right tibia (IF, n = 5) or both tibias (2 × IF, n = 10) and 30 μl of vector suspension (vector titers: 2 × 109 TU/ml and 1.7 × 108 TU/ml, respectively) was injected into the BM cavity. Five mice from IV and 2 × IF cohorts were treated with 5-fluorouracil (5-FU, 150 mg/kg, GRY Pharma/Teva-Deutschland, Ulm, Germany, www.teva-deutschland.de) 4 days prior to lentiviral marking.

Blood and BM Analysis

At different time points after lentiviral injection for up to 24 months, 200 μl of peripheral blood (PB) was collected by puncture of the vena saphena. Erythrocytes were lysed twice with 0.15 M ammonium chloride solution and the cells were washed in Hank's solution (Sigma-Aldrich, St. Louis, MO, www.sigmaaldrich.com) containing 2% fetal bovine serum (FBS). After Fc receptor blocking (anti-mouse CD16/CD32), cells were stained with antibodies against murine Ter119, C-kit, Sca1, CD3, CD4, CD8, CD19, NK1.1, Ly6G, and CD11b (all antibodies: BD Biosciences, Heidelberg, Germany, www.bdbiosciences.com). Stem/progenitor (C-kit+/Sca1+), erythroid (Ter119+), T-lymphoid (CD3+/CD4+ or CD3+/CD8+), B-lymphoid (CD19+) subpopulations as well as natural killer cells (NK1.1+/CD8), granulocytes (CD11b+/Ly6G+), and monocytes (CD11b+/Ly6G) were purified using fluorescence-activated cell sorting (FACS) with AriaI or AriaII cell sorters (BD, Biosciences, Heidelberg, Germany, www.bdbiosciences.com). After sorting, cells were pelleted by centrifugation and stored at −80°C for later molecular analyses. All FACS measurements were validated by negative controls using blood samples from mock-injected GFP-negative BL6 or GFP-transgenic Krt-15 mice stained with the same antibody combinations.

At the end of the experiments, animals were sacrificed and BM cells from individual mice were harvested by flushing of femur and tibia with 3 ml of Hank's solution containing 2% FBS. BM cells were lysed with 0.15 M ammonium chloride solution and then processed as previously described for blood cells.

DNA Isolation

Frozen cell pellets were thawed, resuspended in 15 μl of lysis buffer (0.5% Tween 20 [Sigma-Aldrich, St. Louis, MO, www.sigmaaldrich.com]; 0.5% Nonidet P40 [AppliChem, Darmstadt, Germany, www.applichem.com]; 0.91 mg/ml Proteinase K [Qiagen, Hilden, Germany, www.qiagen.com]; 1× polymerase chain reaction (PCR) buffer [Qiagen, Hilden, Germany, www.qiagen.com]) and incubated for 30 minutes at 56°C. Subsequently, cell lysates were placed at 4°C and 1 μl was used for quantification of DNA-concentration by HPRT-PCR. In individual samples, DNA isolated from sorted cells was preamplified using the Illustra GenomiPhi V2 DNA Amplification Kit (GE Healthcare, Munich, Germany, www.gehealthcare.com) according to instructions of the manufacturer.

Linear Amplification-Mediated-PCR

Junction sequences of the LV and genomic DNA at the integration locus were detected by linear amplification-mediated-PCR (LAM-PCR) as previously described [39]. Briefly, 4 μl of lysed cells from sorted blood lineages was used as a template for 2–4 rounds of a 50-cycle linear PCR with a biotinylated vector-specific primer followed by immobilization on streptavidin-coupled paramagnetic beads and magnetic separation. After hexanucleotide random priming, restriction digestion with Tsp509I (New England Biolabs, Ipswich, UK, www.neb.com) and ligation of the linker, exponential PCRs were performed. Two percent of the first PCR product served as a template after an additional magnetic capture purification step for a second (semi-) nested PCR that enabled the visualization of the LAM-PCR amplicons. The LAM-PCR amplicons were separated on a high-resolution Spreadex gel (Elchrom Scientific, Cham, Switzerland, www.elchrom.com). For sequencing (GATC-Biotech, Konstanz, Germany, www.gatc-biotech.com), the LAM-PCR amplicons were isolated (QIAquick Gel Extraction Kit; Qiagen, Hilden, Germany, www.qiagen.com) and shotgun cloned into the TOPO TA vector (Invitrogen, Carlsbad, CA, www.invitrogen.com). For high-throughput pyrosequencing (GS FLX system; Roche Diagnostics, Mannheim, Germany, www.roche.de), samples were prepared according to the manufacture protocols. Sequences were mapped to the murine genome using the UCSC Blast-like alignment tool genome browser (NCBI37, July 2007/mm9).

Tissue Preparation

At the end of the observation period, mice were sacrificed and tissues were removed and immersed in 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, www.sigmaaldrich.com). The tissues were fixed overnight, embedded in paraffin blocks, and sectioned at 3- to 5-μm thickness. Sections were then mounted onto polylysine slides (Superfrost/Plus; Fisher Scientific, Schwerte, Germany, www.de.fishersci.com) and stored at room temperature. Hind limbs from genetically marked mice were fixed in 4% paraformaldehyde, decalcified for 5 hours in decalcifying solution (Sigma-Aldrich, St. Louis, MO, www.sigmaaldrich.com), and processed as described for other tissues.

Immunofluorescence

Paraffin-embedded tissues were deparaffinized in xylene (Carl Roth, Karlsruhe, Germany, www.carlroth.com) and rehydrated in 100%, 96%, and 70% ethanol. For antigen retrieval, slides were heated in a microwave (600 W) in Tris-EDTA Buffer (10 mM Tris Base, 1 mM EDTA, 0.05% Tween 20, pH 9.0) and then treated with permeabilization solution (0.1% Triton X-100; 0.1% sodium citrate; phosphate-buffered saline [PBS]). After permeabilization, sections were blocked with murine Fc Block (CD16/CD32; BD Biosciences, Heidelberg, Germany, www.bdbiosciences.com) and incubated for 1 hour with goat or rabbit anti-GFP (Abcam, Cambridge, UK, www.abcam.com), rabbit anti-human von Willebrand factor (vWF) (Dako, Glostrup, Denmark, www.dako.com), goat anti-mouse Vimentin, and with rat anti-mouse CD45 (BD Biosciences, Heidelberg, Germany, www.bdbiosciences.com) primary antibodies. The slides were washed in PBS (Gibco/Life Technologies, Darmstadt, Germany, www.lifetechnologies.com) containing 0.01% bovine serum albumin (BSA, PAA Laboratories, Pasching, Austria) and incubated for 1 hour at 37°C with Alexa488-conjugated donkey anti-goat, Alexa546-conjugated donkey anti-rabbit, and DyLight649-conjugated donkey anti-rat antibodies (Invitrogen, Carlsbad, CA, www.invitrogen.com, Jackson IR, Suffolk, UK www.jacksonimmuno.com). After rinsing in PBS/0.01% BSA, cell nuclei were counterstained with DAPI. Following the wash in PBS/0.01% BSA, the slides were rinsed in distilled water and mounted with fluorescence mounting medium (Dako, Glostrup, Denmark, www.dako.com). Staining was completely absent in identical tissue sections in which the primary antibody was absent. Fluorescent signals from single optical sections were acquired by 4-laser confocal microscope (Leica SP5).

Statistical Analysis

Data were analyzed using the unpaired two-tailed Student's t test as appropriate for the dataset. Significance of differences between groups was assessed using Student's t test for p values of <.05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

In Vivo Marked Hematopoietic Cells Contribute to Long-Term Myeloid and Lymphoid Hematopoiesis

To assess the clonal activity of individual stem and progenitor cells in adult steady-state hematopoiesis, highly concentrated enhanced GFP expressing LV vector was injected directly into the femur (IF) of mice to mark HSCs and progenitor cells in the BM (Fig. 1A). To avoid immune responses against GFP and subsequent clearance of transduced cells [40], the GFP-tolerant B6.Cg-Tg(Krt1-15-EGFP)2Cot/J (Krt-15) mouse strain [41] was used. Krt-15 mice express high levels of GFP solely in hair follicle stem cells. Accordingly, we found no GFP expression in PB, BM, and spleen cells or histological sections of lungs, heart, liver, and intestine of Krt-15 mice (data not shown).

thumbnail image

Figure 1. Stable GFP expression in PB of in vivo marked mice for up to 24 months. (A): Experimental design: hematopoietic stem and progenitor cells were genetically marked by injection of highly concentrated GFP expressing LVs IF (I) or intravenously (II) into GFP-tolerant B6.Cg-Tg (Krt1-15-EGFP) mice. At different time points after injection, blood was collected and myeloid and lymphoid cell populations were isolated by FACS. Hematopoietic activity of transduced cells was monitored by flow cytometry and integration site analysis via LAM-PCR of FACS-sorted cells. (B): Stable low level GFP expression in PB of in vivo marked mice. Small numbers of PB cells continuously expressed GFP marker gene for up to 24 months after injection of LV vectors in one or both femurs of the mice (IF and 2 × IF, respectively). Each line represents one individual marked animal (20,000 of PB cells analyzed, n = 10, y-axis is logarithmically scaled). Abbreviations: d, days; FACS, fluorescence-activated cell sorting; GFP, green fluorescent protein; IF, intrafemoral; L, lymphoid cells; LAM-PCR, linear amplification-mediated-polymerase chain reaction; LV, lentiviral vector; M, myeloid cells; PB, peripheral blood.

Download figure to PowerPoint

Flow cytometric analyses of murine PB showed that a small proportion of the PB cells in all IF-injected mice consistently expressed GFP during the whole observation period for up to 24 months (1–100 GFP+ cells per 2 × 104 PB cells analyzed, n = 10, Fig. 1B) indicating stable long-term marking of murine steady-state hematopoiesis. Marked cells contributed to myelopoiesis and lymphopoiesis in the mice. In all injected mice, GFP+ expressing C-kit+/Sca1+ progenitor cells, myeloid (CD11b+/Ly6G monocytes, Ly6G+/CD11b+ granulocytes, and Ter119+ erythroid cells) and lymphoid cells (CD3+/CD4+ or CD3+/CD8+ T-lymphocytes, CD19+ B-lymphocytes) were detectable at various time points after injection (Figs. 2, 3, Supporting Information Figs. S1, S2). In six out of 10 mice analyzed, natural killer (NK, NK1.1+/CD8) cells were found to express GFP. Even at later time points (>6 months after injection), GFP+ short-lived myeloid cells were consistently detected in the PB of all mice (Fig. 3, Supporting Information Fig. S2), indicating that long-term myeloid stem and progenitor cells were successfully marked by IF administration of LVs. Overall numbers and kinetics of GFP+ myeloid cells were similar to the levels and long-term dynamics of GFP+ lymphoid cells (Fig. 3, Supporting Information Fig. S2), suggesting remarkably stable sustained blood-forming activity of marked hematopoietic cells with both myeloid and lymphoid differentiation potential in vivo.

thumbnail image

Figure 2. Genetically marked cells contribute to myeloid and lymphoid hematopoiesis. (A): Representative fluorescence-activated cell sorting (FACS) profiles of lymphoid and myeloid contribution of GFP expressing cells in genetically marked mice at day 91 after intrafemoral injection of lentiviral vectors (LVs). CD19+, CD11b+/Ly6G, and CD11b+/Ly6G+ PB cells were gated for FACS. Gating for GFP+ cells shows that B-lymphoid, CD11b+/Ly6G, and CD11b+/Ly6G+ myeloid cells were marked. (B): GFP expressing erythroid (Ter119+) and C-Kit+/Sca1+ progenitor cells contribute to hematopoiesis at day 91 after intrafemoral injection of LVs. For FACS, Ter119+ and C-Kit+/Sca1+ cell populations were gated. (2 × 105 of living peripheral blood cells were analyzed). Abbreviation: GFP, green fluorescent protein.

Download figure to PowerPoint

thumbnail image

Figure 3. Long-term contribution of marked cells to myeloid and lymphoid hematopoiesis. Myeloid (CD11b+/Ly6G, CD11b+/Ly6G+, and Ter119+) and lymphoid (CD3+/CD4+, CD3+/CD8+, and CD19+) GFP+ cells detected in PB of marked mice at different time points after IF injection of lentiviral vectors. Similar activity of GFP-marked myeloid and lymphoid cells were obtained (y-axis is logarithmically scaled). Abbreviations: d, day; GFP, green fluorescent protein; IF, intrafemoral; Mo, months; PB, peripheral blood.

Download figure to PowerPoint

Intravenous Administration of LVs Preferentially Targets Short-Lived Cells

Although HSCs reside predominantly in the BM, it has been shown previously that even under steady-state conditions, HSCs and progenitor cells are continuously mobilized into the peripheral circulation and home back into the BM [42, 43]. We asked whether multilineage activity and long-term kinetics of individual HSCs and progenitor cells circulating in PB is different from those localized in BM. To address this, we genetically marked hematopoietic cells by IV injection of LV vectors in Krt-15 mice (Fig. 1A). The VSV-G pseudotyped LV particles have a broad tropism binding to lipid components of cellular membranes of all cell types [44]. Intravenous administration of such particles leads to preferential marking of circulating blood cells. After injection into the tail vein, the vector has to pass through the whole blood circulation including the lung capillaries before reaching the BM. It is therefore highly unlikely that functional vector reaches the BM and subsequently transduces BM resident HSCs. A low but stable percentage of PB cells (1-66 GFP+ cells per 2 × 104 PB cells analyzed) was found to continuously express GFP in all injected mice (Fig. 4A). Even though the initial percentage of IV-marked cells was similar to the levels of GFP+ cells after IF injection (p > .05), the kinetics of the marked cells were different. Whereas the percentage of GFP expressing cells in PB of IF-marked mice remained stable over the whole observation period for up to 24 months, a twofold decline in the levels of marked cells was detected 2 weeks after IV-marking indicating that predominantly short-lived more mature cells were transduced after IV vector injection (Fig. 4A). FACS analysis showed that GFP expressing cells equally contributed to myeloid and lymphoid hematopoietic lineages (Fig. 4B, Supporting Information Fig. S3). The long-term kinetics of IV-marked myeloid and lymphoid cells indicate that in addition to short-term hematopoietic cells, stem cells with long-term activity were also present in the circulation and marked by the LV.

thumbnail image

Figure 4. IV-marked cells contribute to hematopoiesis for up to 19 months. (A): IV administration of lentiviral vector (LV) preferentially targets short-lived cells. Twofold decrease of GFP-marked cells were detected in the first 2 weeks after IV administration of the LV vectors (n = 5). (B): Myeloid (CD11b+/Ly6G, CD11b+/Ly6G+, and Ter119+) and lymphoid (CD3+/CD4+, CD3+/CD8+, and CD19+) GFP+ cells contribute to hematopoiesis at different time points after intravenous injection of LVs. Abbreviations: d, day; GFP, green fluorescent protein; IV, intravenous; Mo, months; PB, peripheral blood.

Download figure to PowerPoint

Myelotoxic Pretreatment of Mice Does Not Affect the Kinetics and Lineage Contribution of Marked Cells

Treatment with 5-FU eliminates dividing hematopoietic cells as well as the majority of more committed progenitors and forces quiescent HSCs into the cell cycle to replace the 5-FU ablated hematopoiesis [45, 46]. We analyzed whether myelosuppressive 5-FU pretreatment of mice leads to a higher marking efficiency of hematopoietic cells or marking of clones with different kinetics. For this, 10 mice were treated with 150 mg/kg 5-FU at day 4 prior to IF or IV injection of LVs (n = 5 for each cohort) and kinetics and lineage distribution of vector-containing cells were compared to those of untreated mice. Serial analysis of PB cells revealed no differences in the numbers of GFP-marked cells or their consistent contribution to long-term hematopoiesis between 5-FU treated (Fig. 5A, 5B, Supporting Information Fig. S4) and untreated mice (Figs. 1B, 4A). Moreover, the kinetics of GFP-marked cells in myeloid (CD11b+/Ly6G, Ly6G+/CD11b+, and Ter119+) and lymphoid (CD3+/CD4+, CD3+/CD8+, and CD19+) lineages of 5-FU pretreated animals (Supporting Information Fig. S5) were very similar to untreated animals. Thus, pre-enrichment of hematopoietic progenitors with 5-FU did not result in preferential marking of a particular type of stem or progenitor cell and did not change the multilineage cell output.

thumbnail image

Figure 5. Myelotoxic treatment with 5-fluorouracil (5-FU) does not affect GFP marking efficiency. Numbers of GFP expressing cells in PB of animals pretreated with 5-FU at day 4 prior to marking by intrafemoral (A) or intravenous injection (B) of lentiviral vectors (n = 5 for IF- and IV-cohorts, respectively). Each line represents one individual marked animal. Abbreviations: d, days; GFP, green fluorescent protein; IF, intrafemoral; IV, intravenous; Mo, months; PB, peripheral blood.

Download figure to PowerPoint

Single Marked Clones Contribute to Multilineage Hematopoiesis Over Time

To assess the clonal contribution of individual marked stem and progenitor cells to long-term hematopoiesis, we monitored the LV-integration sites in highly purified PB cell lineages at different time points after in vivo injection using LAM-PCR (Fig. 1A). LAM-PCR analyses revealed that multiple clones contributed to myeloid and lymphoid long-term hematopoiesis (Fig. 6A–6D). Notably, the size of individual clones was very small. In line with the GFP expression results, we did not observe any differences in overall numbers of marked clones detected in murine PB of IF- and IV-injected animals. In addition, lineage contribution and long-term activity of genetically marked hematopoietic clones were similar. HSC pre-enrichment by 5-FU did not change the total number of transduced clones or their kinetics. Tracking the progeny of lentivirally marked hematopoietic clones in purified blood lineages over time revealed a variable lineage contribution of individual stem cell clones. Whereas some clones exclusively contributed to myeloid or lymphoid lineages (Fig. 6A–6D), other clones contributed to both lymphoid and myeloid lineages simultaneously or at different time points, indicating long-term activity of multipotent HSCs (Fig. 6A–6D). High-throughput sequencing of LAM-amplicons confirmed the genomic integration of the LV. Taken together, integration site analysis demonstrated stable contribution of uniquely marked stem cells to myeloid and lymphoid long-term hematopoiesis.

thumbnail image

Figure 6. Marked clones contribute to lymphoid and myeloid hematopoiesis for up to 9 months. Representative data from linear amplification-mediated-polymerase chain reaction (LAM-PCR) analysis of highly purified blood cell lineages of 2 × IF- (A) or IF-marked mouse (B, C,D) over time show different behavior of hematopoietic clones. Some clones exclusively contributed to myeloid (dashed arrows) or lymphoid hematopoiesis (filled block arrows). Other clones were found contributing to myeloid and lymphoid hematopoiesis simultaneously (white arrows) or at different time points (open block arrows). Numbers indicate individual marked clones contributing to myelo-lymphoid hematopoiesis over time. The two lanes at each time point represent aliquots of identical samples from two individual LAM-PCR reactions. Abbreviations: d, day; IC, internal control; M, 100 bp DNA marker; Mo, months.

Download figure to PowerPoint

Hematopoietic and Nonhematopoietic Cells Expressed GFP for up to 24 Months After Marking

To address the cell type and location of marked cells in murine BM as well as in other tissues, histological analyses of IF- and IV-injected mice were performed. Hematopoietic cells were identified by the expression of CD45, endothelial cells by vWF, and mesenchymal cells by Vimentin expression. Immunostaining of histological sections at 14.5–24 months after marking verified that GFP expressing hematopoietic cells were located in the BM (Fig. 7A), confirming our FACS and LAM-PCR data. The majority of GFP-marked cells were arranged in clusters of 2–13 cells that were distributed in the BM cavity as well as localized very close to subendosteal region of BM (Fig. 7B). Some of the GFP expressing hematopoietic cells were in close contact to the endosteal surface (Fig. 7A). Nonhematopoietic BM stromal and endothelial cells expressed the GFP marker protein even at 24 months after LV injection. In addition, GFP expressing CD45+ and CD45 cells were found in the spleen, liver, lungs, and kidney of IF- or IV-injected mice (kupffer cells, hepatocytes, mesenchymal, and endothelial cells) (Fig. 7C–7E). IV-injected animals showed at least fourfold higher numbers of GFP-marked cells in the liver and a twofold increased amounts of GFP+ cells in the spleen as compared to IF-injected mice. The numbers of GFP expressing cells were very similar in lungs as well as in kidney of IF- or IV-injected mice. Thus, the numbers and distribution of GFP expressing cells demonstrated very stable long-term marking of hematopoietic and nonhematopoietic cells by in vivo LV injection.

thumbnail image

Figure 7. Hematopoietic and nonhematopoietic cells express GFP marker protein for up to 21.5 months after in vivo injection of LV vectors. Representative immunostaining of histological sections of in vivo marked mice. Bone marrow, spleen, and liver sections were stained with anti-GFP antibody to detect genetic marker protein, with anti-vWF and with anti-CD45 antibodies for analysis of endothelial and hematopoietic cells. (A): GFP-marked CD45+ hematopoietic cells (filled arrow) located in the endosteal region of murine BM 14.5 months after intravenous (IV)-injection of lentiviral vectors. (B): Clusters of CD45-negative GFP expressing cells in the BM of intrafemoral (IF)-marked mice 21.5 months after LV injection. (C): High numbers of CD45+ hematopoietic (filled arrow) and vWF+ endothelial (open arrow) GFP expressing cells were detected in spleen of IV-marked mice at 14.5 months after LV injection. (D): GFP-marked CD45+ kupffer cell (filled arrow) in the liver of IF-marked mice at 21.5 months after LV-injection. (E): GFP+ hepatocytes (filled arrow) in the liver of IF-injected mice at 21.5 months after in vivo marking. Abbreviations: GFP, green fluorescent protein; vWF, von Willebrand factor.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

Our results for the first time provide insights into the clonal dynamics of in vivo marked HSC and progenitor cell clones with myeloid and lymphoid differentiation potential in the undisturbed murine steady-state hematopoiesis.

We used a highly sensitive genetic marking approach that allowed us to visualize the activity of individual stem and progenitor cells without transplantation. A potential disadvantage of this strategy is that genetic marking by integrating vectors is not always inert to the target cell. Retroviral vectors may alter the fate of HSCs by frequently integrating in the vicinity of transcriptional start sites and subsequent transcriptional activation of neighboring genes [47–49]. This may have dramatic consequences for affected hematopoietic cell clones: insertional mutagenesis driven clonal dominance to overt leukemia has been shown in mice [50] and humans [51–53]. In our study, we minimized the risk of integration site driven gene activation using a SIN LV to clonally mark individual HSC and progenitor cell clones in vivo. In contrast to γ-retroviral vectors, LVs preferably integrate into gene coding regions and are thereby less likely to activate adjacent genes [18, 54, 55]. Moreover, the U3 region of the long terminal repeats of SIN vectors is truncated, resulting in their inactivation after genomic integration [56]. Consequently, in our study, transduced HSCs stably contributed to myeloid and lymphoid hematopoiesis without any sign for in vivo expansion of genetically marked cells, indicating that the clonal marking itself did not mediate clonal selection or dominance. Another systematic bias in the analysis of stem cell kinetics after genetic marking may be introduced by immune responses against marker gene expressing cells [40, 57]. To avoid such transgene-specific immune responses, we used a GFP-tolerant mouse strain expressing GFP exclusively in hair follicle bulge cells [41]. Stable long-term expression of GFP in hematopoiesis for up to 2 years after in vivo marking was observed without any evidence for an extinction of transgene expressing cell clones.

In our study, a remarkably small but stable proportion of transduced cells contributed to steady-state hematopoiesis and this was highly reproducible in all experiments done. As VSV-G pseudotyped LV unselectively binds to all cells, it is not surprising that only few HSC clones were marked in individual mice. Unexpectedly, at all time points, the contribution of each clone to blood formation was very small: 1–10 marked clones detected by LAM-PCR generated only 0.01%–0.23% of all PB cells in individual mice. The proportion of marked hematopoiesis was very similar after IV administration of LV vectors with or without preconditioning with 5-FU, besides the preferential transduction of short-living hematopoietic progenitors present in the PB [42]. Assuming that a representative proportion of all HSCs was transduced by in vivo injection, a surprisingly high number of more than 400 to several thousand HSC clones simultaneously carried steady-state hematopoiesis. In contrast, only a limited number of ex vivo manipulated stem cell clones have been reported to maintain hematopoiesis following transplantation [7, 8, 13]. Pretransplantation conditioning may contribute to this apparent difference, as stroma and endothelial cell damage may alter the HSC niche function leading to a limited, only partial refilling of the stem cell compartment [27, 32–35, 58]. Moreover, limitations in the sensitivity of clonal marker detection may have led to an underestimation of the number of active HSCs in earlier mouse transplantation studies [7, 8, 15, 16] by restricting the analysis to large or dominant clones [39, 59]. In non-human primates, recent data indicate that more than 500 marked clones contribute to post-transplant hematopoiesis [17, 18, 60]. Moreover, monitoring of distinct hemoglobin types produced in aggregated embryonic chimeras suggested that several hundred clones simultaneously contribute to erythropoiesis [61].

Recently, transplantation studies of highly purified single long-term HSCs in mice revealed substantial differences in their differentiation programs [12, 23, 62]. Whether lineage-biased differentiation of these classes of HSCs also occurs under steady-state conditions is currently unknown. In our study, myeloid cells expressing GFP in the periphery were detected in 15 out of 25 mice at ≥12 months after in vivo marking and individual marked clones generated lymphoid and myeloid progeny indicating that at least a proportion of all long-term stem cells contributed to both lymphopoiesis and myelopoiesis. Due to the very small clone size, it was not possible to dissect the lineage contribution and kinetics of each individual marked stem cell clone. It remains an open question whether active clones with lineage-biased differentiation programs contribute to the undisturbed hematopoiesis, but given the very little clonal contribution of individual stem cells in a highly polyclonal system, this question may be difficult or even impossible to address experimentally.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

In summary, we assessed the clonal contribution of HSCs and progenitor cells to steady-state hematopoiesis over time by an in vivo marking strategy. Our data demonstrate the stable long-term contribution of individually marked hematopoietic cells with myeloid and lymphoid differentiation potential to murine steady-state hematopoiesis. Our data indicate that an unexpected high number of individual stem cell clones may contribute to murine steady-state hematopoiesis. In vivo marking will further allow to directly analyze the response of extrinsic stress stimuli such as blood loss or chemotherapy to individual HSC clones.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

We gratefully acknowledge Steffen Schmitt (Core facility flow cytometry, DKFZ) for cell sorting and the support by the DKFZ Light Microscopy Facility. We thank Norma Howells and Silke Hamzaoui-Nord for animal care. The lentiviral constructs for this study were kindly provided by Dr. L. Naldini (San Raffaele Telethon Institute for Gene Therapy—“Vita-Salute San Raffaele” University Medical School, Milano, Italy). This work was supported by grants from the German Cancer Aid (Deutsche Krebshilfe, Project No. 107217) and German Research Foundation (DFG, SFB 873).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. Disclosure of Potential Conflicts of Interest
  10. REFERENCES
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
SC_12-0104_sm_supplFigure1.pdf289KFigure S1. Mock transduced Krt-15 mice do not express GFP in peripheral blood. Representative FACS profiles of PB cells of mock transduced control mice at day 86 after IF-injection. (20,000 living PB cells analyzed)
SC_12-0104_sm_supplFigure2.pdf349KFigure S2. Multilineage contribution of marked cells to lonG−term hematopoiesis. Myeloid (CD11b+/Ly6G−, CD11b+/Ly6G+, Ter119+) and lymphoid (CD3+/CD4+, CD3+/CD8+, CD19+) GFP+ cells detected in PB of marked mice at different time points after IF injection of lentiviral vector. Similar activity of GFP marked myeloid and lymphoid cells were obtained. (PB= peripheral blood, d= day, Mo= months, IF= intrafemoral, y-axis is logarithmically scaled)
SC_12-0104_sm_supplFigure3.pdf176KFigure S3. Myeloid (CD11b+/Ly6G−, CD11b+/Ly6G+, Ter119+) and lymphoid (CD3+/CD4+, CD3+/CD8+, CD19+) GFP+ cells contribute to hematopoiesis at different time points after intravenous injection of lentiviral vector. (PB= peripheral blood, d= day, Mo= months, IV= intravenous, y-axis is logarithmically scaled)
SC_12-0104_sm_supplFigure4.pdf275KFigure S4. Reduced numbers of CD11b+/Ly6G+ granulocytes in peripheral blood of Krt-15 mice treated with 5-Fluorouracil. Representative FACS profiles of PB cells of 5-FU untreated control mice (A) as compared to 5-FU treated mice at day 7 after injection of 5-FU (B). Reduced numbers of CD11b+/Ly6G+ granulocytes were obtained in PB of 5-FU injected mice. (20,000 living PB cells analyzed)
SC_12-0104_sm_supplFigure5.pdf514KFigure S5. Myeloid and lymphoid marked cells contribute to lonG−term hematopoiesis in 5-FU pretreated animals. Myeloid (CD11b+/Ly6G−, CD11b+/Ly6G+, Ter119+) or lymphoid (CD3+/CD4+, CD3+/CD8+, CD19+) blood cells were detected at different time points after intrafemoral or intravenous injection of lentiviral vector in animals treated with 5-FU prior to marking. Mice were injected intravenously with 150mg/kg of 5-FU at day 4 prior to LV-marking. (PB= peripheral blood, d= day, Mo= months, IF= intrafemoral, IV= intravenous, y-axis is logarithmically scaled)

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.