|Home | About | Journals | Submit | Contact Us | Français|
The use of primary human cells to model cancer initiation and progression is now within the grasp of investigators. It has been nearly a decade since the first defined genetic elements were introduced into primary human epithelial and fibroblast cells to model oncogenesis. This approach has now been extended to the hematopoietic system, with the first described experimental transformation of primary human hematopoietic cells. Human cell model systems will lead to a better understanding of the species and cell type specific signals necessary for oncogenic initiation and progression, and will allow investigators to interrogate the cancer stem cell hypothesis using a well-defined hierarchical system that has been studied for decades. The molecular and biochemical link between self-renewal and differentiation can now be experimentally approached using primary human cells. In addition, the models that result from these experiments are likely to generate highly relevant systems for use in identification and validation of potential therapeutic targets as well as testing of small molecule therapeutics. We describe here the methodologies and reagents that are used to examine the effects of leukemia fusion protein expression on primary human hematopoietic cells, both in vitro and in vivo.
Many different systems are now available for modeling leukemia, but by far the most popular model is the mouse, using either genetically engineered mice or a bone marrow transduction and transplantation system (1)(refer to Chapter “Retroviral/Lentiviral Transduction and Transformation Assay”). However, it is becoming increasingly clear that there are important differences between murine and human cells with respect to cellular transformation (2–4). Species-specific differences in ras signaling as well as in telomerase and telomere length regulation limit the conclusions that can be reached when using murine cells to model human cancer (5–7). The inbred nature of the mouse strains used for these studies further complicates the extrapolation of results to humans. For these reasons, some labs are now using human cells to study oncogenesis, and the controlled transformation of primary human fibroblasts and epithelial cells has been previously reported (8, 9). Similar studies have been pursued using hematopoietic stem and progenitor cells (HSPCs) to model preleukemia (10–14), and recently the experimental transformation of a primary human HSPC has been accomplished in the laboratory (15).
The strategy that is used in these studies with primary human HSPC necessarily involves the introduction of oncogenes using retroviral or lentiviral delivery systems. The specific HSPC that is used, whether the cord blood, bone marrow or mobilized CD34+ (or lineage negative) cell, may alter the results that are obtained, since it is now recognized that these cells differ significantly in their properties (16). In addition, many delivery options are available using different viral promoters, marker genes, and viral envelopes. These options give great flexibility to the investigator for pursuing different experimental approaches as well as in using combinations of genes together in the same experiment. The in vitro and in vivo systems in which to examine the phenotype upon oncogene expression are also expanding. As we learn more about the specific signals that are important in self-renewal divisions of the human HSPC, we strive to mimic these signals in vitro, to expand normal or preleukemic cells for analysis and therapeutic purposes. The strains of immunodeficient mice are continually improving, allowing greater sensitivity for xenograft analysis and ultimately for in vivo drug treatment studies. With the defined transformation of primary human cells of hematopoietic origin now in hand, we can use these systems to study the cell type specific signals involved in oncogenesis and identify the critical pathways that can be therapeutically targeted in these cancers.
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 (18–20). 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.
1The quality of the whole blood sample should be considered prior to proceeding with the selection protocol. We have found that a significant drop in yield occurs at volumes smaller than 80 mL and we generally do not proceed with these samples. Additionally, the procedure should begin within 48 h of collection. The samples can be stored at room temperature in the presence of anticoagulants. Samples with obvious clumps or a very dark color should be avoided.
2Solutions should be prepared without including the BSA. Solutions can then be stored at 4°C for months. Withdraw an aliquot of each solution and add BSA to each immediately before use.
3The choice of virus collection media will depend on which culture condition is preferred (serum or serum-free) for culture of the CD34+ cells after transduction. Phoenix-GP producer cells produce adequate viral particles in either virus collection media.
4The response of CD34+ cells to FBS culture conditions can vary significantly depending on the lot of serum. For this reason, it is important to carefully test several lots before purchase. We generally perform methylcellulose colony assays of UCB CD34+ cells with several lots of serum before purchasing a bulk quantity of a satisfactory lot (most colonies, with good growth and multilineage differentiation, with performance at least as good as the current or past lots).
5A typical stock BME solution is 14.4 M. To prepare aliquots of 500× concentrated BME, dilute 35 µL of stock BME into 10 mL water and sterilize with a 0.2- µm syringe filter. This yields a 10−2 M solution (500×, add 1 mL to a 500-mL bottle IMDM) that can be aliquoted and stored at −20°C. Solutions are generally stable for 1 month at 4°C, wrapped in foil.
6The acetic acid/trypan blue solution will allow easy counting of nucleated cells. The red blood cells will be lysed in this solution. Depending on the number of cells in the preparation, a larger or smaller dilution of the suspension may be needed.
7The yield and quality of cells from these procedures can vary significantly, due to sample differences. From an average UCB of 100 mL, the expected yield of CD34+ cells should be between 0.5 and 1.5 million cells. Occasionally far more CD34+ cells can be recovered, but typical recovery is approximately 50% of the CD34+ cells present in the MNC fraction. The purity of the selected cells should be confirmed by flow cytometry and the cell counts adjusted to represent the actual number of CD34+ cells. We routinely achieve purities of greater than 90% using these methods (Fig. 1a). A small aliquot of the selected cells can be tested for growth and viability by culturing in myeloid media and monitoring cell counts or by performing methylcellulose assays.
8The Phoenix cells, based on the 293T cell line, are available through ATCC in an agreement with Dr. Garry Nolan. The cells can be obtained with only the gag and pol genes (Phoenix-GP), or additionally with the ecotropic or amphotropic envelopes (Phoenix-E or Phoenix-A). We typically use Phoenix-GP and transfect additional gag/pol helper plasmid, as this has been shown to be the limiting construct in viral preparations. Full details on these cell lines are available upon purchase from ATCC.
9Many methods of transfection exist. The amounts of DNA required by these methods vary. The DNA amounts presented here are optimized for the calcium phosphate precipitation method. In using different methods of transfection we have found that the ratio of the different plasmids is an important factor and should be preserved if the total amount of DNA is altered.
10The fluorescent intensity of EGFP may not be impressive on the first day after transfection, but by day 2 it should be readily visible, especially from the control cells transfected with empty vector. It is possible that the transgene in the viral construct will severely diminish the intensity of the EGFP. We have found that this is transgene-specific but often occurs when large genes or oncogenes are used. In this case, the intensity of the EGFP will not correlate with the viral titer.
11The rule of thumb is that approximately 50% of viral titer is lost for each freeze/thaw cycle when using amphotropic or ecotropic envelopes. We have not formally tested this. When using the feline endogenous virus (RD118) or the vesicular stomatitus virus envelope (VSV-G), there is little to no loss of virus titer. These two envelopes are also reported to allow concentration of virus particles by ultracentrifugation without loss of viral titer.
12For HT1080, use the following concentrations for commonly used drugs. G418 for neomycin resistance, 800 µg/mL, (10–14 days, with a media change at day 5); hygromycin B, 500 µg/mL, (7–10 days); puromycin, 1 µg/mL (4–7 days).
13Media is IMDM with 10% heat-inactivated FBS, 50 U/mL penicillin, 50 U/mL streptomycin, 2 mM l-glutamine, 10−4 M β-mercaptoethanol. Alternatively, if serum-free conditions are preferred, the FBS is substituted with the BIT supplement from Stem Cell Technologies, supplied as a 5× concentrate. If the transduced cells will be used in vivo (in immunodeficient mice), the human cytokines SCF, TPO (MDGF), and FLT3L are included at 100 ng/mL. For in vitro applications only, the addition of human IL-3 at 20 ng/mL and IL-6 at 100 ng/mL will increase growth and transduction efficiency but will also promote differentiation.
14Some cells will die during the prestimulation period, and some cells will begin to divide. This typically results in approximately an equal viable population relative to the starting number on Day 0. If the number is lower than 50% of the viable cell number on Day 0 it is often not worth proceeding with the transduction, since the cells that do recover are usually less hardy and long-lived and may not represent the expected population of human CD34+ cells.
15Ensure that cells have adhered to the RetroNectin-coated plastic after approximately 15 min of incubation. This is essential for good transduction by all pseudotypes of virus with the exception of VSV-G; if the virus is pseudotyped with VSV-G, no RetroNectin is used for transduction (27, 28).
16Working concentrations for human CD34+ cells for some common drugs are: G418 (for neomycin), 800 µg/mL (will take approximately 1.5 weeks; fresh drug should be added at day 4 or 5); hygromycin B, 300 µg/mL (will take approximately 1 week); puromycin, 0.5 µg/mL (will take 3–5 days for selection). Nontransduced cells should be incubated with drug, and selection is complete when all of these cells are dead. As a note of caution, it is possible that some transgenes will have deleterious effects on human CD34+ cells, and selection with drug resistance may be impossible. A fluorescent marker is more useful in this case; cells can be sorted after transduction and analyzed for these effects.
17We find the cells grow best when we use six-well plates, for reasons we have not determined. Cultures can reach a density of 2 million cells/mL without significant loss of viability, and a volume of up to 8 mL can be used for each well of the plate.
18The MS-5 cell line is grown in α-MEM medium with 10–20% FBS. Cells are never allowed to become confluent. A large number of viably frozen vials should be made to ensure a reliable stock of cells that are at an early passage. Over time in vitro, the phenotype/morphology of the cells can become more fibroblastoid, and these cultures are less able to support hematopoiesis and also tend to lose contact inhibition. The MS-5 cells do not need to be irradiated for use as a feeder layer. If cells will not be used for experiments within 2–3 weeks, it is better to discontinue the culture and thaw a new vial rather than to continue passage of cells during this time.
19A typical cell split would include a gentle agitation of the flask to loosen the hematopoietic cells that are weakly attached to stroma cells, followed by removal of 50% of the volume with replacement by fresh media. This procedure is frequently referred to as demi-depopulation. Care must be taken to ensure that the stroma does not loosen, which becomes more likely with each passing week.
20The cytokines to use in a methylcellulose assay that will permit the broadest range of colony types includes 10 ng/mL G-CSF, 20 ng/mL each of SCF, IL-3 and IL-6, and 6 U/mL of Epo.
21Large 500 cm2 square tissue culture dishes can be used to create a humidified chamber. Fill both halves of two 60-mm culture plates with sterile water and place one piece into each corner of the dish (use each lid and the bottom separately). This chamber can hold approximately twenty 35-mm methylcellulose cultures. The chambers are sterilized and reused.
22There is great variability in the engraftment of different cord blood preparations and also large variability mouse to mouse from the same cord blood. This biological variability is impossible to control at this time, and means that a large number of mice must be used for experiments.
23We have found that the intrafemoral route, although significantly more time-consuming and requiring more practice, results in a significantly superior graft. It has also been shown that the cell number can be markedly reduced given the superiority of this route of injection (29, 30). It is likely that with the newly available NOG mice, and the intrafemoral route of injection, the number of CD34+ cells that are needed for each mouse will be less than tenfold what is currently recommended.
24While looking at the mouse knee, locate the highest point (just above the white connective tissue of the joint; it is easier for some investigators if the hair at the knee joint is removed, by simply plucking it off with gloved index finger and thumb). Insert the needle just above that point and slightly to the outside of the leg. The inside of the bone will have a gritty texture that can be sensed by moving the needle up and down slightly. Before attempting this procedure on live mice, it is best to obtain the skill by practicing with sacrificed mice. Try inserting a colored dye into the femur; e.g., trypan blue. Dissection will confirm injection and allow a certain amount of trial and error that is required to successfully and consistently perform the procedure.