An ex vivo protocol to generate mature human red blood cells from hematopoietic stem/progenitors is described. Additionally we describe an efficient lentiviral-delivery method to knockdown the transcription factor TAL1 in primary erythroid cells. The efficiency of lentivirus mediated gene delivery is demonstrated using GFP expressing viruses.
Erythropoiesis is a commonly used model system to study cell differentiation. During erythropoiesis, pluripotent adult human hematopoietic stem cells (HSCs) differentiate into oligopotent progenitors, committed precursors and mature red blood cells 1. This process is regulated for a large part at the level of gene expression, whereby specific transcription factors activate lineage-specific genes while concomitantly repressing genes that are specific to other cell types 2. Studies on transcription factors regulating erythropoiesis are often performed using human and murine cell lines that represent, to some extent, erythroid cells at given stages of differentiation 3-5. However transformed cell lines can only partially mimic erythroid cells and most importantly they do not allow one to comprehensibly study the dynamic changes that occur as cells progress through many stages towards their final erythroid fate. Therefore, a current challenge remains the development of a protocol to obtain relatively homogenous populations of primary HSCs and erythroid cells at various stages of differentiation in quantities that are sufficient to perform genomics and proteomics experiments.
Here we describe an ex vivo cell culture protocol to induce erythroid differentiation from human hematopoietic stem/progenitor cells that have been isolated from either cord blood, bone marrow, or adult peripheral blood mobilized with G-CSF (leukapheresis). This culture system, initially developed by the Douay laboratory 6, uses cytokines and co-culture on mesenchymal cells to mimic the bone marrow microenvironment. Using this ex vivo differentiation protocol, we observe a strong amplification of erythroid progenitors, an induction of differentiation exclusively towards the erythroid lineage and a complete maturation to the stage of enucleated red blood cells. Thus, this system provides an opportunity to study the molecular mechanism of transcriptional regulation as hematopoietic stem cells progress along the erythroid lineage.
Studying erythropoiesis at the transcriptional level also requires the ability to over-express or knockdown specific factors in primary erythroid cells. For this purpose, we use a lentivirus-mediated gene delivery system that allows for the efficient infection of both dividing and non-dividing cells 7. Here we show that we are able to efficiently knockdown the transcription factor TAL1 in primary human erythroid cells. In addition, GFP expression demonstrates an efficiency of lentiviral infection close to 90%. Thus, our protocol provides a highly useful system for characterization of the regulatory network of transcription factors that control erythropoiesis.
Part I. ex vivo erythropoiesis of human hematopoietic stem/progenitor cells
1. Isolation of CD34+ human hematopoietic stem/progenitor cells
Human CD34+ cell population, which contains a mixture of hematopoietic stem cells (HSCs) and early progenitors 8, is harvested from umbilical cord blood, peripheral blood mobilized with G-CSF (leukapheresis) or bone marrow (Figure 1-Step1). If using cord blood or peripheral blood, go directly to step 1.2. If using bone marrow, start at step 1.1. For reagents and equipment, see Table 1.
This step is done under sterile conditions (i.e. hood) and at RT unless indicated otherwise.
2. Erythroid differentiation in cell culture
CD34+ cells are cultured in supplemented IMDM medium with sequential supply of hydrocortisone and specific combinations of cytokines according to a four-step protocol adapted from 6 (Figure 1-Step 2). A series of tests is performed throughout the procedure to verify cell growth and erythroid differentiation.
The number of CD34+ cells required to start a culture is variable, depending on the scale of the experiment. Typically, for
gene expression studies we perform a small-scale culture starting with 0.1×106 CD34+ cells in 10 ml medium and maintaining small volumes by freezing extra
cells throughout the procedure. For studies where protein extraction is required, cultures are performed at a larger scale. Here we describe a large-scale
experiment where we start with ~7×106 CD34+ cells isolated from leukapheresis.
Sterile conditions are used throughout. See Tables 1, 2 and 3 for reagents.
Comments on the results:
Figure 1 outlines the protocol for ex vivo differentiation of human CD34+ cells to mature, hemoglobin-containing red blood cells. During erythroid differentiation, cells are highly proliferative as illustrated by the ~8,000-fold amplification shown on Figure 2 that was obtained using CD34+ cells isolated from leukapheresis. CD34+ cells isolated from bone marrow are slightly less proliferative with ~6,000-fold amplification while CD34+ cells isolated from cord blood can attain 50,000 fold amplification. Cells at particular stages of differentiation display recognizable sizes and morphologies that are illustrated on Figure 3. Notice the process of enucleation that occurs at the end of differentiation starting at Day 20. At Day 26, all cells have lost their nuclei (Figure 3).
Colony forming assays shown on Figure 4 are used to detect early and late hematopoietic progenitors during the first days of differentiation (Day 0 to Day 8). In addition to erythroid progenitors (CFU-E and BFU-E), we note that granulocyte-macrophage progenitors (CFU-GM) are significantly represented within the early CD34+ cells at Day 0. As differentiation proceeds, CFU-GM decrease significantly, and at Day 6 they have been completely replaced by erythroid progenitors BFU-E and CFU-E. It is also important to note that from Day 4 to Day 6, the early erythroid progenitors BFU-E are progressively replaced by their more differentiated counterparts CFU-E. By Day 10, erythroid cells have essentially lost their colony-forming capacity.
Finally, erythropoiesis is monitored by the expression of cell surface markers as shown in Figure 5. Note: 1) the progressive loss of the hematopoietic stem/progenitor marker CD34; 2) the progressive acquisition and loss of the progenitor marker CD36; 3) the progressive acquisition and loss of the erythroid progenitor and reticulocyte marker CD71; and 4) the increase of the erythroid precursor and erythrocyte marker GPA. Other markers of erythroid differentiation include hemoglobin synthesis starting at Day 8 (measured by benzidine staining-Figure 5) and cell enucleation starting at Day 20 as measured by the loss of the DNA stain LDS (Figure 5). Finally, erythroid differentiation can be measured by RT-qPCR of molecular markers such as the transcription factor TAL1 (Figure 6B) or the β-globin genes (data not shown).
Typically, 1 ml of cord blood / leukapheresis / bone marrow provides 0.35×105 / 0.5×105 / 1.1×105 CD34+ cells, which can theoretically produce 1.7×109 / 0.4×109 / 0.7×109 red blood cells with respective proliferation rates of 50,000 / 8,000 / 6,000. To obtain 50×106 red blood cells at the end of the differentiation procedure, you will need approximately 50 ml of culture medium.
Part II. Lentiviral mediated knockdown during ex vivo erythroid differentiation
Lentivirus is a tool of choice for gene transfer in human primary cells (see for example Amsellem et al. 9). Lentiviruses are often used to knockdown genes of interest in primary cells via delivery of targeted shRNA (see for example Laurenti et al. 10). In this protocol we target the transcription factor TAL1 in primary human erythroid cells. An shRNA with a scrambled sequence that does not correspond to any known gene is used as a negative control. GFP expression is used for evaluating transduction efficiency.
1. Preparation of lentiviral vector particles
To obtain lentiviral particles, human embryonic kidney 293T cells are cotransfected with three plasmids: one encompassing a second-generation virus packaging system coding for HIV-1 genes except env, vpr, vif, vpu and nef (i.e. psPAX2); one coding for the envelope of the vesicular stomatitis G-pseudotyped virus (i.e. pMD2G); and one lentiviral vector coding for the gene of interest (i.e. pBLOCK-it6-DEST(shTal1) coding for TAL1 shRNA, pBLOCK-it6-DEST(shScr) coding for scrambled shRNA or pWPIP coding for GFP) (Table 4). Lentiviral particles are collected from the cell culture supernatant of transfected 293T cells.
To avoid unwanted infection of cells, we perform all lentiviral experiments (preparation of viruses and infection) under a dedicated hood separate from other cell cultures. In addition, lentivirus-generated waste is discarded separately. Second generation lentiviruses described in this protocol are replication-incompetent and should be treated as Biosafety Level 2 organisms. Some institutions might have specific safety procedures regarding lentiviruses and we recommend that you follow established institutional guidelines.
Here we describe a protocol that will provide enough lentiviral particles to infect 1 to 2 millions primary erythroid cells at any given stage of differentiation. Volumes should be adjusted accordingly for scaling up or down. See Tables 4 and 5 for reagents.
2. Lentiviral infection of primary erythroid cells for gene delivery
3. Representative Results
Figure 1. Protocol outline for the isolation and ex vivo erythroid differentiation of human primary hematopoietic stem/progenitor cells. (Step1)
CD34+ cells (purity of 95 ± 3%) are isolated from cord blood, bone marrow or human peripheral blood mobilized with G-CSF (leukapheresis) using positive
immunomagnetic selection. (Step2) CD34+ cells are grown in liquid culture according to the indicated four-step protocol.
Figure 2. Cell amplification during ex vivo erythroid differentiation protocol. Results are in logarithmic scale. Representative results
obtained with CD34+ cells isolated from leukapheresis.
Figure 3. Cell morphological analysis during ex vivo erythroid differentiation. Cells are stained with May-Grünwald-Giemsa reagent at the indicated
time points during differentiation (magnification x 40). HSC: hematopoietic stem/progenitor cells; BFU-E: burst forming units –erythroid; CFU-E: colony forming
units –erythroid; Pro EB: pro-erythroblasts; Baso EB: basophilic erythroblasts; Poly EB: polychromatophilic erythroblasts; Ortho EB: orthochromatic erythroblasts;
RET: reticulocytes; RBC: red blood cells.
Figure 4. Colony-forming cell (CFC) assays to detect and quantify hematopoietic progenitors. At the indicated time points, 104 cells were
seeded in a semisolid methylcellulose medium in the presence of recombinant human cytokines: rh SCF, rh GM-CSF, rh G-CSF, rh Il-3, rh EPO to support optimal growth of
erythroid progenitors (BFU-E and CFU-E) and granulocyte-macrophage progenitors (CFU-GM). BFU: burst forming units; CFU: colony forming units. Typical picture of
BFU-E and CFU-E are also shown. Note the red-brown color due to hemoglobinization of BFU-E and CFU-E.
Figure 5. Kinetics of hematopoietic/erythroid markers during ex vivo erythroid differentiation. At the indicated time points, 2×105 cells were
harvested and analyzed for expression of indicated markers by FACS (CD34, CD36, CD71, GPA and LDS) or by microscopy (Benzidine). Bars represent percentages of positive
cells. CD34: transmembrane glycoprotein, present on human hematopoietic stem/progenitor cells; CD36: thrombospondin receptor, present on
progenitor cells; CD71: transferrin receptor, present on erythroid progenitors and reticulocytes; GPA: glycophorin A, present on
erythroid precursors and erythrocytes; Benzidine: hemoglobin stain; LDS: laser dye styryl, DNA stain.
Figure 6. Lentiviral-mediated gene transfer in human primary proerythroblasts. (A) Timeline for induction of TAL1 knockdown (KD) in primary
proerythroblasts using the lentiviral gene delivery system. (B) TAL1 knockdown was mediated by lentivirus-delivered anti-Tal1 shRNA (TAL1 KD)
or a scrambled control shRNA (Scr). Kinetics of TAL1 mRNA transcript level is measured by RT-qPCR. TAL1 transcript is expressed relative to 18S RNA. (C)
Lentivirus-delivered GFP was tested by microscopy (left panel) or FACS (right panel). 87% of cells were GFP positive.
4. Materials
Reagent | Source | Catalog number |
RPMI 1640 | Sigma | R8758 |
Collagenase B | Roche Applied Science | 11088807001 |
DNase I | Sigma | D5025-15KU |
Pre-Separation Filters (30μm nylon mesh) | Miltenyi Biotec Inc | 130-041-407 |
Ficoll-paque PLUS | GE Healthcare Bio-Sciences | 17-1440-03 |
EasySep 10x Red Blood Cells Lysis Buffer | StemCell Technologies | 20120 |
CD34 MicroBead Kit | Miltenyi Biotec Inc | 130-046-703 |
LS Mini-Macs columns | Miltenyi Biotec Inc | 130-042-401 |
autoMACS Running Buffer-MACS Separation Buffer | Miltenyi Biotec Inc | 130-091-221 |
MiniMACS Separator | Miltenyi Biotec Inc | 130-042-102 |
Mouse bone marrow stromal cells MS-5 (regularly test for mycoplasma contamination) | DSMZ | ACC 441 |
Trypan Blue Solution (0.4%) | Sigma | T8154 |
MEM Alpha+GlutaMAX | Gibco Invitrogen | 32571 |
Trypsin 0.05%/EDTA | ThermoScientific | J101521 |
May Grunwald Giemsa | Biostain Ready Reagents Ltd | RRSP54 |
Methocult Colony-Forming Cell (CFC) Assays for Human Cells | StemCell Technologies | 04044 |
Table 1.
Component | Stock solution | Volume 600 ml |
IMDM | Sigma I3390 | 443.4 ml |
Pen/Strept 1% (vol/vol) | Gibco 10378 | 6ml |
L-glutamine 4×10-3M | 200mM (Gibco 25030) | 12 ml |
Inositol 40μg/ml | 4mg/ml (Sigma I5125) preparation : Dissolve 28 mg powder in 7 ml IMDM and filter sterilize-stable for 1 week at 4°C. |
6ml |
Folic acid 10μg/ml | 1 mg/ml (Sigma F7876) preparation : Dissolve 7 mg powder in 7 ml pre-warmed IMDM and filter sterilize-stable for 1 week at 4°C. Pre-warm before using. |
6ml |
Monothioglycerol 1.6×10-4M | 0.16M (Sigma M6145) preparation : Dissolve 10 μl of 11.56M stock solution in 712 μl IMDM and filter sterilize-stable for 1 week at 4°C. |
0.6ml |
Ferrous nitrate 90 ng/ml | 18μg/ml ( Sigma 8508) preparation: Dissolve 1.8 mg in 1ml distilled water. Take 0.1 ml, add 9.9 ml IMDM and filter sterilize-stable for 1 week at 4°C. |
3ml |
Ferrous sulfate 900 ng/ml | 180 μg/ml ( Sigma F8633) preparation: Dissolve 18 mg in 1ml distilled water. Take 0.1 ml, add 9.9 ml IMDM and filter sterilize-stable for 1 week at 4°C. |
3ml |
Bovine serum albumine-Insulin-Transferrin (BIT) 20% (vol/vol) |
StemCell Technologies 9500 | 120 ml |
Table 2.
Component | Stock solution | Volume to add to 100 ml of supplemented IMDM (Table 2) |
Hydrocortisone (HC) 10-6 M | 10-4 M (Sigma H2270) preparation : Dissolve 5mg powder in 1.03ml IMDM. Take 0.1 ml, add 9.9 ml IMDM and filter sterilize. Stable for 1 week at 4°C. |
1ml |
Stem Cell Factor (SCF) 100ng/ml | 100 μg/ml (Cedarlane 4327-50) preparation: Dissolve 50 μg powder in 500 μl PBS+0.1% BSA and filter sterilize. Make aliquots of 100 μl and store at -20°C. |
100 μl |
Erythropoietin (EPO) 3 IU/ml | 100 μg/ml equivalent of 12,000UI/ml (Feldan FB03-10-290) preparation: Dissolve 50 μg powder in 500 ml PBS+0.1% BSA and filter sterilize. Make aliquots of 25 ml and store at -20°C. |
25 μl |
Interleukin 3 (IL3) 5 ng/ml | 100 μg/ml (Feldan FB03-10-330) preparation: Dissolve 10 μg powder in 100 μl PBS+0.1% BSA and filter sterilize. Make aliquots of 5 ml and store at -20°C. |
5 μl |
Table 3.
Reagent | Source | Reference |
293T cells (regularly test for mycoplasma contamination) |
ATCC | CRL-11268 |
DMEM/High Glucose | ThermoScientific | SH30243.01 |
Hexadimethrine bromide (Polybrene) | Sigma | H9268-5G |
OptiSeal ultracentrifugation tubes | Beckman Coulter | 361625 |
psPAX2 plasmid | Addgene (D. Trono laboratory) | 12260 |
pMD2.G plasmid | Addgene (D. Trono laboratory) | 12259 |
pBLOCK-it6-DEST (shTal1) plasmid | Tal1 shRNA sequence cloned into the pBLOCK-it6-DEST plasmid from Invitrogen | Palii et al. EMBO J. |
pBLOCK-it6-DEST (shScr) plasmid | Scrambled shRNA sequence cloned into the pBLOCK-it6-DEST plasmid from Invitrogen | Palii et al. EMBO J. |
pWPI plasmid | Addgene (D. Trono laboratory) | 12254 |
Table 4.
Buffer | Preparation | Final concentration |
2xHBS |
|
|
CaCl2 |
|
|
MS30 |
|
|
Table 5.
A number of methods have been used previously to differentiate erythroid cells in culture with variable degrees of success. For example, some protocols without co-culture on MS-5 allows expansion of erythroid cells but are not efficient in producing fully mature enucleated red blood cells 12-17. While other methods allow efficient enucleation, it is at the expense of proliferation 18,19. The method we have used here, initially developed by the Douay laboratory 6, combines an initial liquid culture step, which prolongs cell amplification during early stages of erythroid differentiation and thus provides us with large numbers of early progenitors (necessary for genomics and proteomics studies), and a second step of co-culture on MS-5 mesenchymal cells, which is important to maximize hemoglobin synthesis and to obtain complete enucleation of red blood cells. Two limitations of our method are 1) the still insufficient degree of amplification at the earliest stages of differentiation, which so far has prevented us to fully characterize the very early erythroid progenitors at the protein level and 2) the high cost of performing this protocol, particularly on a large-scale.
The quality of cell culture reagents and cytokines is essential for ex vivo erythroid differentiation. Notably, the use of BIT (Table 2) as a serum replacement is critical. Indeed in our experience the use of BIT eliminated batch-to-batch variations that we experienced previously (likely due to differences in reagent purity) when using separate sources of Bovine serum albumin, Insulin and Transferrin. Another critical point is the quality of the MS-5 layer used for co-culture of erythroid cells. Indeed, MS-5 cells must be 70% confluent when erythroid cells are added and must be changed every 3-4 days. If MS-5 cells are too confluent, erythroid cells will not grow and differentiate properly. In addition, unhealthy MS-5 cells tend to detach, making the harvesting of erythroid cells difficult.
Previous methods have used retrovirus for gene-delivery in hematopoietic stem/progenitor cells in culture. However, retroviruses infect only dividing cells and are therefore not efficient in infecting the mainly quiescent CD34+ cells isolated from cord blood, bone marrow or leukapheresis 20. The main advantage of lentiviral approaches for gene-delivery is that it allows an efficient infection of cells independently of their proliferation status 7,21. Our lentiviral-mediated gene delivery allows one to efficiently knockdown or over-express specific transcription factors that can modify cell fate. Appropriate phenotypic analyses (cellular and molecular) can be devised after infection to reveal particular phenotypes. In addition, growth conditions may be altered to favor particular cell-types after infection. Along these lines, in this protocol we have chosen not to use any method to select lentivirus-infected cells for two main reasons: 1) our infection efficiency is very high (Figure 6C) and 2) we are concerned that the methods used to select lentivirus-infected cells (e.g. cell-sorting or antibiotic) are biased towards surviving/proliferating cells that are able to express either the antibiotic-resistance gene or the fluorescent marker at high levels thereby concealing a potential apoptotic phenotype.
The authors have nothing to disclose.
We thank L. Douay and M.C Giarratana (Université Paris VI, France) for advice on the ex vivo erythroid differentiation, D. Allan and H. Atkins (OHRI, Canada) for providing blood samples (obtained under the Ottawa Hospital Research Ethics Board #2007804-01H), D. Trono (Ecole Polytechnique Federale de Lausanne) for providing the lentiviral vectors and F.J. Dilworth (OHRI, Canada) for critically reading the manuscript. This project was funded by a CIHR grant (MOP-82813) to M.B. C.G. P. is a recipient of an Ontario Research Fund Computational Regulomics Training Postdoctoral Fellowship. M.B. holds the Canada Research Chair in the Regulation of Gene Expression.