We typically use cord blood CD34+ cells in our protocols, as these cells have been shown to give the most robust engraftment in immunodeficient mice (17
). However, whether these fetal/infant cells are representative of the bulk of human leukemia, which occurs in the adult, remains an open question, and it will be necessary to perform comparative experiments to answer these questions. Our system is focused on retroviral constructs rather than lentiviral, and it remains to be determined whether the strong viral promoters that are present in these constructs are contributing to the phenotypes obtained, presumably through retroviral insertional activation of endogenous oncogenes. It will also be interesting to examine whether lentiviral transduction of quiescent, noncycled human HSPC will increase the transduction frequency of the most primitive cells. The optimal viral envelope for use in human CD34+ experiments is also a variable that has been analyzed in some studies, with varying and sometimes contradictory results (18
). As more labs become proficient in the growth, transduction, and transplantation of human CD34+ cells, it is likely that these questions will be answered.
The specific conditions that are used for the transduction of the human CD34+ cells depend upon the ultimate use of the cells upon transduction. If the cells are to be propagated in vitro, and will not be injected into immunodeficient mice, it is less critical that the cytokines IL-3 and IL-6 are excluded from the prestimulation mix. Including these cytokines will increase cell yield and transduction efficiency, and typically does not negatively impact on the overall expansion and proliferation of the cells in vitro. The choice of immunodeficient mouse, and the route of delivery as well as the age at which to transplant, depends upon the availability of the strains and the expertise of the lab. Intravenous injection of 6- to 8-week-old NOD/SCID or NOG mice (both commercially available strains) is the most popular approach, but the use of newborn pups, cranial facial vein injection, intrafemoral injection, and the newer strains of immunodeficient mice (e.g., NOD/SCID-SGM3 mice for myeloid biased grafts) are gaining popularity, and only time will tell which approach will be superior for studying normal human hematopoiesis as well as leukemogenesis in the mouse.
3.1. Isolation of Human CD34+ Cells
- Mix the whole blood in an equal volume of sterile PBS in a 50-mL tube, layer slowly onto 0.5 volumes of Ficoll-Paque PLUS solution, spin at 250 × g for 30 min to 1 h at 18–20°C. The majority of red blood cells and granulocytes will be pelleted while mononuclear cells including CD34+ cells will form a white interface just above the ficoll. Serum is left above the interface.
- Harvest the mononuclear cells (MNC) from the interface, transfer to a 50-mL tube, wash cells in at least 3 volumes of selection buffer, centrifuge at 500 × g for 15 min at 18–20°C.
- Resuspend the pellet in 5 mL of selection buffer and transfer to a 15-mL tube. Rinse the original 50-mL tube with an additional 5 ml of selection buffer and transfer to the 15-mL tube to make 10 mL of total volume.
- Take 10 µL of MNC and mix with 90 µL of counting solution (see Note 6).
- Spin at 700 × g for 8 min, aspirate the supernatant, resuspend the MNC in selection buffer at the density recommended by the manufacturer of the selection kit.
- CD34+ cells are positively selected from this population using either the Miltenyi Macs CD34 Microbead Kit or the Stem Cell Technologies CD34 EasySep procedure. Both protocols are successful and give comparable yields and purity of CD34+ cells (). Each protocol is followed according to the manufacturer’s recommendations (see Note 7).
Fig. 1 Overview of the model system for retroviral transduction of human CD34+ cells. (a) Human CD34+ cells are purified by magnetic selection, and purity is confirmed by flow cytometric analysis. (b) A transduction of human CD34+ cells is shown, with a non-transduced (more ...)
- After selection of CD34+ cells, count the cells and centrifuge at 700 × g for 10 min to pellet them. Aspirate the supernatant. Cells can either be used immediately as described in Subheading 3.3 below, or viably frozen by resuspending the pellet in 300–500 µL of hetastarch solution 1 and transferring to a cryovial. An equal volume of hetastarch solution 2 is then added dropwise (105 to 2 × 106 cells per vial).
3.2. Virus Preparation and Titer
- Either the 293T or Phoenix cell line is used for transient virus production using a plasmid transfection procedure. The “strain” of Phoenix cell line used depends on the species of the cells that will be targeted (see Note 8). Virus production is best done in batches that are titered and aliquoted for future use. 10-cm tissue culture dishes are coated with poly-l-lysine by adding 3 mL of solution to the dish, swirling to cover the whole bottom, and transferring to subsequent dishes. 5 million cells are then plated to each dish (day 1). To ensure even distribution, do not swirl the media, but rock the dishes twice side-to-side and twice to-and-fro. The dishes are placed in the incubator overnight (37°C, 5% CO2).
- Virus production is initiated by transfecting the cells in the 10-cm dish with three plasmids:
All three plasmids are used in the transfection, even if a Phoenix cell line already contains the gag/pol or envelope plasmids as stable integrants. The quality of the plasmid used in the transfection is critical for maximal uptake by the cells. Only DNA from a maxiprep should be used for virus production. Different transfection protocols can be used, but one of the most reliable and cost-efficient is the calcium phosphate method (see Note 9).
- Retroviral vector 12 µg
- Gag/Pol vector 10 µg
- Envelope vector 3 µg
- Transfect the cells the afternoon of day 2. The cells should be approximately 75% confluent. All solutions should be at room temperature. For calcium phosphate transfection, a final volume of 1-mL transfection mix is needed per 10-cm dish. The volumes can be scaled up if multiple dishes are transfected with the same retroviral vector. Add plasmids to ddH2 O to a final volume of 438 µL and mix by brief pipeting. Next, add 62 µL of 2 M CaCl2 and mix. Then add this solution dropwise to a Falcon 2054 polystyrene tube containing 500 µL of 2× HBS (pH 7.10). The solution should appear somewhat cloudy after this step. Some investigators prefer to vortex while adding the DNA mixture, others prefer to bubble air through the 2× HBS during the addition. Results seem to be comparable using the two techniques in our experience.
- Remove dishes from the incubator and slowly add 1 mL of the DNA–CaPO4 mixture solution directly on top of the cells. Use a 1-mL pipetman with the tip beneath the media to discharge the mixture while moving the tip around, blanketing the entire cell layer. Return the dishes to the incubator. Some investigators prefer to incubate the mixture for 15–45 min before adding to the cells, but we find that this increases the size of the precipitate which decreases the transduction efficiency. Very fine black particles (visible at 20× power under a microscope) are the best precipitate for transfection.
- On the morning of Day 3 (12–16 h after initiating transfection), remove dishes from the incubator and aspirate the media by gently tipping the dish at a 45° angle and touching the tip of the aspirating device to the side of the dish. Slowly add 6–7 mL of warmed collection media to the side of the angled dish, so as not to disturb the cell monolayer. The media that is used depends on the target cells that will be transduced and should be the optimal media for these cells. Ideally, virus-containing supernatant is now collected at 12-h intervals, for a total of 4–5 collections. Fresh warm collection media is replaced each time, gently, to prevent disruption of the cell monolayer. Virus supernatant can be kept overnight on ice at 4°C to allow concentration of the supernatant in one step. Producer cells should be visualized under the fluorescent microscope, if a fluorescent marker protein is used (i.e., EGFP). At least 50% of cells should be transfected to justify proceeding with the collections (see Note 10).
- Combine virus collections, filter through a 0.45- µm filter (0.2 µm can also be used with no loss of virus and less risk of producer cell contamination), and concentrate using a protein purification column with a molecular weight cutoff of 100-kD. The specific fold concentration depends on the needs of the investigator and the transgene used, but can range between 5 and 100× or more. These filters can be re-used for the same virus supernatant. Aliquot and freeze concentrated virus at −80°C. Keep a very small aliquot for determining viral titer. Depending on the envelope used, viral titer will drop by one-half for each freeze-thaw cycle. (see Note 11).
- To titer the virus, we use the HT1080 adherent cell line (for a human-tropic virus). These cells are available from ATCC.
- Remove sub-confluent HT1080 cells from the plate with trypsin–EDTA and plate in six-well plates at 2.5 × 105 cells per well, in DMEM 10% FBS (Day 1).
- The following day, dilute viral supernatant by taking 30 µL of virus and mixing with 3 mL of DMEM 10% FBS. This is a 10−2 dilution. Make tenfold serial dilutions by taking 300 µL of the 10−2 dilution and adding to 2.7 mL of DMEM 10% FBS (10−3), repeating this three more times until a 10−6 dilution is reached. Aspirate the media from the six-well plate of HT1080 cells, add 2 mL of the 10−6 dilution to a single well, and continue for each of the remaining dilutions. Use a new pipet and add 2 mL of DMEM 10% FBS to the final well of the six-well plate (no-virus control well). Add 2 µL of an 8 mg/mL polybrene solution to each well (final concentration is 8 µg/mL). Incubate overnight at 37°C in a 5% CO2 incubator.
- On day 3, aspirate the medium from each well. Add 2 mL of complete media. If the retrovirus contains a drug-selectable cassette, add the working solution of drug in complete media at this time (see Note 12).
- After 10–14 days, fix the colonies of cells that have formed by adding ice-cold methanol for 5–10 min and then rinse once with distilled deionized water. Stain the colonies with an aqueous solution of 0.4% methylene blue for one minute, wash twice with water and dry the plate upside down. Calculate the titer. The best dilution for calculations will be one that gives between 10 and 100 colonies.
- For viral vectors that contain a fluorescent marker, collect cells two days after the end of transduction (day 5) and analyze by flow cytometry.
- Viral particle number is calculated by multiplying the percentage marked cells by the total number of target cells present at the time of incubation with virus (day 2; a replicate well is included, and this is counted at the time of incubation with virus) (It is important that the dilution that is used for the calculation of titer has given a transduction efficiency of less than 37%, to ensure single hit kinetics). Multiplying this number by the dilution factor for that well (e.g., 100 for a 10−2 dilution) will give the titer of the virus per mL. We have found that the viral titer should not be less than 105/mL for successful transduction of human CD34+ cells.
3.3. Transduction of Human CD34+ Cells
- To thaw frozen human CD34+ cells, agitate the vial rapidly in a 37°C water bath until only a small piece of ice remains. Add 2 mL of room temperature HBSS + 2% BSA to the vial, mix, and transfer to a 15-mL tube. Rinse the vial and cap with HBSS + 2% BSA and increase the volume in the 15-mL tube to 10 mL. Centrifuge at 4°C, 500 × g for 10 min, flick the tube with index finger to disperse the cell pellet, and resuspend in prestimulation media to give a final concentration of 106 cells per mL (see Note 13).
- Culture the cells for 1.5–2 days to ensure that a population of cells are replicating (essential for transduction with retrovirus).
- Prepare a RetroNectin-coated plate the day before transduction. Incubate a six-well nontissue culture treated plate with RetroNectin solution (2 h, room temperature), then PBS containing 2% BSA (30 min, room temperature), and finally with HBSS containing 2.5% HEPES (quick wash). Although plates can be stored for months at 4°C, we have found that a fresh treatment is superior to stored plates. The primary solution of RetroNectin can be used in a secondary (and tertiary) treatment, and these plates can be used for transduction of less critical cell lines if desired.
- Count the cells. The cell number should be approximately equal to the number when cells were thawed (see Note 14).
- Precoat the RetroNectin well with 2–4 mL of unconcentrated viral supernatant or 1 mL of concentrated supernatant. Centrifuge the plate for 45 min at 2,000 × g at room temperature. Aspirate the solution and repeat the spin with additional virus supernatant. After the second treatment, add approximately 1 million prestimulated cells in an equal volume of fresh prestimulation media to the viral supernatant in the well, and incubate at 37°C for 4–8 h (during the day), with 8 µg/mL polybrene (see Note 15).
- At the end of the day, carefully aspirate the majority of the media (cells should be predominantly attached to the Retro-Nectin-coated plastic) and add 2 mL of fresh prestimulation media. Allow the cells to recover in the incubator overnight.
- The next morning, add 2–4 mL (unconcentrated) or 1 mL (concentrated) of virus and polybrene. Do not centrifuge. At the end of the day, remove the cells from the plate by collecting the media and detaching the cells from the plate using a non-enzymatic cell dissociation buffer. Centrifuge at 500 × g for 7 min and resuspend in the appropriate media at a density of 106 cells per mL.
- Two days later, cells can be analyzed by flow cytometry (if a fluorescent co-marker was present in the retroviral construct; ) to determine the transduction frequency, or drug selection can begin if a drug-selectable construct was used (see Note 16).
3.4. In Vitro Culture of Transduced Cells
- The choice of culture medium will depend on the specific lineage that is desired and the oncogene that is used in the transduction. For culture under myeloid conditions, a complex mixture of cytokines is most likely to allow the greatest diversity in terms of myeloid potential for the cells. After transduction, cells can be cultured in a five-cytokine cocktail of SCF, MDGF, Flt3L, IL-6, and IL-3. The specific concentrations to use can be determined empirically and may differ depending on the transgene. We have found that 10 ng/mL of each is sufficient for the AML1-ETO, CBFB-SMMHC, and MLL-AF9 fusion protein-expressing long-term cultures (12, 21, 22). Typically the control-transduced CD34+ cells will proliferate for 8–12 weeks under these conditions, while the oncogene-transduced cultures grow for varying lengths of time, depending on the specific oncogene.
- Count cells weekly and seed at 4 × 105 cells/mL, in a volume that will give the desired number of cells after the 1-week expansion period (see Note 17). During the first 3–4 weeks, cells will expand approximately tenfold each week, and supplementation of the medium may be needed on day 4 or 5 to prevent depletion of the medium. Under these conditions, the control cultures will maintain a population of CD34+ cells of approximately 2–10%, depending on the specific cord blood. The remainders of the cells are at various stages of myelopoiesis. A layer of adherent cells will slowly form over time, and the suspension cells can be moved to a new well each week if desired. Toward the end of the proliferative period (weeks 7–12 depending on the cord blood), cells will double only once or twice per week, and the majority of the cells will be monocyte/macrophage. Expression of the fluorescent marker, if present in the retroviral construct, can be monitored during this time, to determine effects of the transgene on proliferation and/or differentiation. An example of a myeloid culture of cells expressing AML1-ETO in shown in .
Fig. 2 Overview of the in vitro analysis of retrovirally transduced human CD34+ cells. (A) Representative long-term myeloid cultures (6 weeks) of control CB cells and cells expressing the AML1-ETO oncogene (AE). This phenotype is preserved throughout the 6–8 (more ...)
- For expansion of B-lymphoid cells in vitro, the use of a stroma coculture will give the best growth and the most reliable and reproducible results. We use the MS-5 murine stroma cell line (23) (see Note 18). Immediately after transduction, seed 1 × 105 cells onto a monolayer of MS-5 stroma cells formed in a T25 culture flask with 5 mL of B-cell media. Some cells will invade the stroma layer and grow beneath the stroma as “phase-dark” cells, meaning that they will appear as dark, nontranslucent cells through the phase-contrast microscope. Under B-cell growth conditions, these cells will typically form organized areas known as cobblestone areas, and will also appear dispersed throughout the monolayer in addition to growing in loosely organized areas. An example of a typical cobblestone area forming cells is shown in . The majority of cells will grow as suspension cells or weakly attached to the topside of the stroma. The dilution to make during weekly passage of cells will depend upon the individual capacity of each cord blood preparation, which demonstrates large proliferative variations under B-cell growth conditions (see Note 19). For the first 2–3 weeks of growth, only myeloid cells (CD33+) will be present in the suspension. At weeks 3–4, a small population of CD19+ cells will form, which can become the majority population by 5–6 weeks of culture. Most of the CD19+ B-cells will co-express CD10, and a percentage (10–50%) will co-express CD20, but in our hands the B-cells that develop under these conditions do not express surface Ig ().
- The methylcellulose assay is a convenient and powerful way to determine the effects of a particular transgene on hematopoietic differentiation and proliferation. For the most quantitative assay, transduced cells should be sorted or drug selected prior to use in a methylcellulose assay. Typically, after transduction, cells are washed with IMDM without cytokines or FBS and counted. Then, 8,000 cells are deposited into a 15-mL tube in a final volume of 800 µL IMDM (enough cells to give triplicate methylcellulose plates with 2,000 cells each, and allowing for pipetting error by calculating for four plates). The cells are mixed with 3.2 mL of methylcellulose media while vortexing at high speed. The methylcellulose is easiest to measure and manipulate using a 3-mL syringe with a 16 gauge needle (fill the syringe with methylcellulose solution from the bottle, discharge to clear the air void, and completely fill the syringe by slowly withdrawing the plunger fully. This will give a volume of approximately 3.2 mL). Rest the syringe and needle in the falcon tube for approximately 5 min to allow the air bubbles to rise.
- Draw up 3 mL of methylcellulose/cell suspension, and add 1 mL to each of triplicate 35-mm dishes that contain 50 µL of 20× cytokine cocktail (see Note 20). Tilt and rotate the dish to distribute the solution over the entire bottom. Incubate the dishes for 2 weeks in a humidified chamber to prevent drying of the methylcellulose media (see Note 21). Minimize disturbance or movement of the dishes during this time.
- After two weeks, count the colonies and score the colony type. We typically will score three types of colonies, including granulocyte/macrophage (GM), burst forming uniterythroid (BFU-E), and granulocyte/erythroid/macrophage/megakaryocyte (GEMM) colonies (). Transgene expression in human CD34+ cells could affect the size, the number as well as the specific type of colonies present. For example, expression of the CBFB-SMMHC oncogene causes a G0/G1 arrest, and the number of colonies that are present after 2 weeks are severely decreased (). The expected number of colonies from 2,000 input cells varies greatly depending on the particular cord blood as well as the timing of the experiments. Numbers could range from 50–400 colonies per dish. The best growth/differentiation will typically occur when colony number is around 100–200 per dish.
3.5 Injection of Transduced Cells into Immunodeficient Mice
- If cells are going to be used for injection into immunodeficient mice, the culture medium for prestimulation should have only minimal cytokines to preserve the primitive nature of the cells as much as possible. The cytokines SCF, TPO, and Flt3L should be sufficient to promote cell cycle entry and survival; IL-3 should be avoided, since this cytokine has been found to promote differentiation and loss of SCID repopulating potential.
- Cells should be injected into animals as soon after thaw as possible. Our usual approach would be to prestimulate for 1.5 days, transduce cells for 0.5–1.0 days (1–2 incubations with virus, 4–6 h each) and then immediately inject into animals, so that only two or three days pass from initial thaw. A small aliquot can be retained in vitro to check verify transduction and calculate efficiency.
- The best characterized animal model for xenograft of human hematopoietic cells is the NOD/SCID mouse. A number of variants are now available, including the NOD/SCID-β2M−/−, NOD/SCID-IL2Rγc−/− (NOG), and NOD/SCID-SGM3 mice. The latter two strains are just becoming widely used, and limited information is available in the literature for these mice. The NOG mouse should be highly immunocompromised and possess no residual NK activity and may be defective for dendritic cell and macrophage function as well (24). The SGM3 mouse is transgenic for the myeloid-promoting cytokines SCF, GM-CSF, and IL-3 and skews the human graft towards myeloid differentiation, but has also been documented to promote loss of the normal human graft, possibly due to mobilization and differentiation of the primitive human cells in the mouse bone marrow (25, 26).
- The number of human cells that must be injected to ensure a reliable hematopoietic graft will vary depending on the mouse strain and sub-strain that is used. A good rule of thumb for regular NOD/SCID mice, with injection of prestimulated, cycling cells, is to use 300,000 cells per mouse, calculated based on the starting number of the CD34+ population. For example, if 9 × 105 CD34+ cells are prestimulated, and after 4 days have doubled in number to 1.8 × 106, this would be enough cell number for three mice (900,000 starting cells/300,000 = 3). If the cells tripled in this time to 2.7 × 106, this would still be enough cell number for only three mice. (see Note 22).
- Mice (6–8 weeks old on the day of injection) should be prepared for injection by feeding with doxycycline-treated chow for 1 week before irradiation, and continued on this chow for 1 week after irradiation (we have recently found that we now achieve better results using Bactrim-treated chow for 2 weeks post-irradiation, with doxycycline-treated chow for 1 week before irradiation. It is possible that this treatment needs optimization for individual animal facilities, and procedures may need to be altered if mouse loss becomes unacceptable after irradiation). This minimizes the loss of mice due to radiation illness. The dose of radiation needs to be determined empirically for each colony, and the dose may need to be recalculated as the colony ages. We have found that our colony gave very good results using 375 Gy initially, but we now use 300 Gy (2 years after establishing the colony) with similar survival numbers (approximately 10–20% of mice will be lost in each experiment due to radiation illness).
- Mice are irradiated up to 24 h in advance of injection. For intravenous injection, a volume of 300 µL works well (cells in a PBS solution). For intrafemoral injection, a volume of 25 µL or less should be used. (see Note 23).
- To inject cells intravenously, place mice under a heat lamp for several minutes. This will allow easier visualization of the tail vein. Injection is easiest with immobilized mice, either by containment in a mouse restraint device or by anesthetization with isoflurane. After cleaning the tail with an alcohol wipe, inject the cells into the tail vein with an insulin syringe taking special care to avoid injection of any air bubble (this will kill the mouse). You will be able to see the clear PBS-cell mixture travel through the vein. If the needle is not in the vein, the plunger will not easily be depressed and a bump will form in the tail of the mouse. Attempt the initial injection midway down the tail so that if this occurs, another attempt can be made closer to the body of the mouse.
- To inject intrafemorally, anesthetize the mouse with isoflurane. Give the mouse painkiller such as Buprenex (buprenorphine) by injecting under the skin of the back. Place the mouse on its back, and insert the snout of the mouse into a tube for constant delivery of isoflurane during the procedure. Clean the leg with an alcohol wipe and Betadine. Insert a 25 gauge needle (5/8 in.) attached to a 1-mL syringe into the femur (see Note 24). Remove the needle slowly, while keeping the leg steady and remaining fully focused on the location of the hole. Insert the insulin syringe containing the cells and slowly inject the mixture into the hole. Move the mouse back to its cage where it will regain consciousness within 5 min.
For measurement of a graft due to the most primitive human cells, it is recommended that the mice be monitored at 9–12 weeks post injection. It is often difficult to measure human cells in the peripheral blood in these mice, but depending on the size of the graft it is possible. We use the CD45, CD33 and CD19 antibodies to determine total cell graft, myeloid and B-lymphoid populations respectively. To block non-specific staining of murine hematopoietic cells that express FcR, the use of 1 µL of Fc block is recommended per 1 × 106 cells. For surface staining of PB, red cells should be lysed in an ammonium chloride solution, and the residual red blood cells should be excluded from the viable cell gate during flow to determine an accurate percentage of human white blood cells in the mouse.
- When the experiment is to be ended, the bone marrow, spleen and peripheral blood of the animal should be processed and analyzed for human cells. The four long bones of the hind legs are removed, and either crushed using mortar and pestle or flushed using an insulin syringe and IMDM media after cutting the ends of the bones with scissors. Either way, cells are then filtered through a 40 µm filter, and red blood cells are lysed. The spleen or other organs can be crushed through a 40 µm filter, using the flat end of a 3 mL syringe, and red cells are then lysed. Cells are stained for surface markers to determine the lineage composition of the human graft (). Expression of a leukemia fusion gene can specifically affect the composition of the graft, as shown by expression of the inv16 oncogene CBFBSMMHC, which leads to a myeloid-dominated graft with very few CD19+ cells and a decreased number of CD34+ cells (). Methylcellulose assays using human-specific cytokines can be performed to determine the progenitor activity, using the protocol described earlier. Cells can be viably frozen for later use. To show self-renewal of the human cells using the most stringent criteria, a secondary transplant is performed, following the same procedures as for the primary transplant. Secondary mice are analyzed at 6–12 weeks for the presence of human cells..
Fig. 3 Overview of the in vivo analysis of retrovirally transduced human CD34+ cells. (A, B) Analysis of the bone marrow of a NOD/SCID-β2M−/− mouse that was injected with control transduced (MIGR1) and CBFB-SMMHC-transduced (inv16) cells. (more ...)