Validated gene transfer and expression tracers are essential for elucidating functions of mammalian genes. Here, we have determined the suitability and unintended side effects of enhanced green fluorescent protein (EGFP) and DsRed-Express fluorescent protein as expression tracers in long-term hematopoietic stem cells (HSCs). Retrovirally transduced mouse bone marrow cells expressing either EGFP or DsRed-Express in single or mixed dual-color cell populations were clearly discerned by flow cytometry and fluorescence microscopy. The results from in vivo competitive repopulation assays demonstrated that EGFP-expressing HSCs were maintained nearly throughout the lifespan of the transplanted mice and retained long-term multilineage repopulating potential. All mice assessed at 15 months post-transplantation were EGFP positive, and, on average, 24% total peripheral white blood cells expressed EGFP. Most EGFP-expressing recipient mice lived at least 22 months. In contrast, Discosoma sp. red fluorescent protein (DsRed)-expressing donor cells dramatically declined in transplant-recipient mice over time, particularly in the competitive setting, in which mixed EGFP- and DsRed-expressing cells were cotransplanted. Moreover, under in vitro culture condition favoring preservation of HSCs, purified EGFP-expressing cells grew robustly, whereas DsRed-expressing cells did not. Therefore, EGFP has no detectable deteriorative effects on HSCs, and is nearly an ideal long-term expression tracer for hematopoietic cells; however, DsRed-Express fluorescent protein is not suitable for these cells.
Hematopoietic stem cells (HSCs) are rare cells found in bone marrow, fetal liver, and umbilical cord blood. They undergo self-renewal and differentiation to continuously replenish all types of mature blood cells and, therefore, maintain the blood and immune systems [1, 2]. Hematopoietic stem and progenitor cells are much pursued and promising targets for gene therapy applications to treat a variety of blood and immunological disorders . After transplantation, genetically modified HSCs are able to completely or partially reconstitute the recipient's hematopoietic and immune systems, which can persist in the recipient for life [4, –6]. Phenotypic attributes and functional properties of murine and human HSCs and progenitor cells are well-characterized by numerous elegant in vitro and in vivo studies [7, , , , –12]. Yet, we still know very little about the molecular components and networks underlying various functions of HSCs and progenitor cells, as well as mechanisms by which these functions are regulated. The genomes of several organisms including human and mouse have been completely sequenced. The remaining challenges are to elucidate and define precise functions of the vast majority of human and mouse genes and to figure out how these encoded proteins work together to perform certain cellular functions within a given cell. At present, assessing the contributions of many different genes to hematopoiesis and determining how they function in HSCs and progenitor cells often requires knockdown or introduction of the gene of interest or its derivative mutants into these cells along with a marker gene used to trace transduced cells over time. Moreover, to study multicomponent cellular machinery as well as protein-protein interaction networks, it is necessary to introduce multiple genes or their derivative mutants along with multiple gene transfer and expression tracers into target cells so that these genes can be tracked unambiguously and simultaneously.
An ideal expression tracer should meet the following criteria: It should (a) be easy to track and nontoxic to cells of interest; (b) not perturb the cell metabolism, nor interfere with various cellular functions; (c) be insensitive to environmental effects that could confound interpretation of experimental results ; and (d) be able to be used in conjunction with other tracers for multicomponent tracking. Fluorescent proteins have several important advantages over other types of gene transfer and expression tracers. They are genetically encoded and easy to track, and make it possible to image specific molecules and/or cells in space and time within living cells and organisms [14, –16]. Recent advances have made available a repertoire of fluorescent proteins with unique characteristics and distinct spectral properties. These fluorescent proteins, especially the optimized green fluorescent protein (GFP) originally discovered in the jellyfish Aequorea victoria, have become very useful and versatile tools. In the past decade, there has been a widespread use of enhanced GFP (EGFP) and various GFP transgenic mice in a broad range of biological and biomedical applications, including visualizing gene expression, localizing and tracing GFP-tagged proteins, color-coding particular cells in complex tissues, and real-time tracking of tumor growth and metastasis [17, –19]. Discosoma sp. red fluorescent protein (DsRed) was first cloned from the reef coral Discosoma sp. and has attracted interest as a gene transfer and expression tracer. It is genetically and spectrally distinct from GFP, so it could potentially be used along with EGFP for simultaneously tracking multiple proteins or cells. DsRed-Express is a human codon-optimized and genetically engineered variant of wild-type DsRed fluorescent protein. Its coding sequence contains nine amino acid substitutions that were identified through both directed and random mutagenesis. These substitutions result in substantially improved solubility of the protein, faster maturation rate, and reduced level of residual green emission compared to wild-type DsRed [20, , –23].
Although EGFP has been used widely as a gene transfer and expression tracer, and other fluorescent proteins have occasionally been used in this capacity as well, their suitability for long-term marking of HSCs and their potential unintended side effects on the functions of these cells have not been systematically investigated. In this study, we have performed a series of in vitro and in vivo experiments to determine whether EGFP and DsRed-Express fluorescent protein are suitable for use as gene transfer and expression tracers in HSCs and progenitor cells.
Materials and Methods
Retroviral Constructs and Generation of Retroviral Vectors
MIEG3, a murine stem cell virus (MSCV)-based bicistronic retroviral construct expressing EGFP was used as described [24, 25]. Expression of EGFP is via the internal ribosome entry site (IRES) derived from the encephalomyocarditis virus. M3Red, a bicistronic retroviral construct expressing DsRed-Express fluorescent protein, contains a backbone identical to that of MIEG3, and was created as follows. Briefly, pIRES2-DsRed-Express plasmid (Clontech, Mountain View, CA, http://www.clontech.com) was digested with EcoR I and Dra I restriction enzymes, and 1,361-base pair (bp) fragment isolated and ligated to the purified MIEG3 EcoR I/Hinc II fragment. The resulting M3Red bicistronic retroviral construct expresses DsRed-Express fluorescent protein through IRES. DNA sequences around the cloning sites of the M3Red construct were verified by DNA sequencing, and the integrity of its backbone was confirmed by restriction mapping. Thus, the DNA sequences of MIEG3 and M3Red constructs are identical except for the EGFP and DsRed-Express coding regions. Ecotropic retroviral vectors were produced by the Indiana University Vector Production Facility (Indianapolis, IN). The MIEG3 and M3Red retroviral vectors were generated side-by-side by transfecting the Phoenix-Eco packaging cell line (ATCC, Manassas, VA, http://www.atcc.org) with the corresponding constructs. Replication competent retrovirus (RCR) was analyzed using the direct S+/L− assay as previously described . The RCR testing showed no detectable RCR in the Phoenix-Eco MIEG3 and M3Red retroviral supernatants.
Low-density bone marrow (BM) mononuclear cells were isolated from 6- to 8-week-old female C57BL/6J mice. All animal procedures were approved by the Animal Care Committee at Indiana University School of Medicine. Isolated BM cells were cultured with a modified cytokine cocktail, and transduced twice on RetroNectin-coated surfaces with the MIEG3 or M3Red retroviral vectors as described previously . Briefly, BM mononuclear cells were prestimulated for 2 days in α-minimum essential medium (MEM) containing 20% fetal bovine serum (FBS; HyClone, Logan, UT, http://www.hyclone.com), 4 mM l-glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin, 100 ng/ml murine stem cell factor (SCF), 100 ng/ml murine FLT3-ligand (Flt3-L), and 100 ng/ml murine thrombopoietin (Tpo). All cytokines were purchased from R&D Systems, Inc. (Minneapolis, MN, http://www.rndsystems.com). Prestimulated cells were incubated in RetroNectin-coated (Takara Mirus Bio, Madison, WI, http://www.takaramirusbio.com) six-well plates with either MIEG3 or M3Red Phoenix-Eco supernatants plus 100 ng/ml SCF, 100 ng/ml Flt3-L, and 50 ng/ml Tpo. After incubation at 37°C for 30 minutes, plates were centrifuged (400g) for 5 minutes at 20°C, and incubated at 37°C for an additional 5.5 hours. Transduction was repeated once the next day. Proportions of EGFP+ and DsRed-Express+ cells were determined by flow cytometry at various time points after transduction.
Competitive Repopulations Assays
Recipient 6–8-week-old B6.SJL-PtprcaPep3b/BoyJ (CD45.1) mice (female) were exposed to 1,100 rads of whole body lethal irradiation administered in two doses of 700 and 400 rads 5 hours apart 1 day before transplantation. Donor BM mononuclear cells isolated from C57BL/6J mice (CD45.2) were prestimulated and transduced as described herein. After transduction, cells were maintained in α-MEM containing 25 ng/ml of murine SCF overnight. Then, 6.75 × 105 to 1.1 × 106 MIEG3-transduced, M3Red-transduced, or mixed MIEG3/M3Red donor cells plus 4–5 × 105 freshly isolated BoyJ BM cells as competitors were transplanted into each of the lethally irradiated recipient mice by tail vein injection (0.2 ml of cell suspension per mouse). Analyses of contributions of the EGFP and DsRed expressing donor cells (EGFP+ CD45.2 and DsRed-Express+ CD45.2) as well as total donor cells (CD45.2) to mature blood cells in recipient mice were carried out at 1 month after transplantation and then at approximately 2-month intervals. Blood samples from tail bleeds were treated with ammonium chloride buffer (155 mM NH4Cl, 10 mM KHCO3, and 0.1 mM EDTA) to lyse red blood cells, and then each sample was split into two equal aliquots. One set was stained with PerCP-Cy5.5-conjugated antimouse CD45.2 and R-phycoerythrin (PE)-conjugated antimouse CD45.1; the other was stained with PerCP-Cy5.5-conjugated antimouse CD45.2 and PE-conjugated antimouse CD11b alone or CD11b plus Gr-1 antibodies (BD Biosciences Pharmingen, San Jose, CA, http://www.bdbiosciences.com). For transplant experiments including M3Red transduced cells, blood samples from all groups were stained with allophycocyanin (APC)-conjugated antimouse CD45.2 or CD45.1 antibodies (eBioscience, San Diego, CA, http://www.ebioscience.com/). After staining, cells were washed and analyzed by flow cytometry.
In Vitro Cell Growth Assay
BM mononuclear cells were transduced as described herein, and cells expressing EGFP or DsRed-Express sorted by flow cytometry. A portion of the BM mononuclear cells was treated in parallel as transduced cells, except that no retroviral supernatants were added during transduction period. These cells were used as untransduced controls. Triplicates of untransduced, sorted EGFP+, sorted DsRed-Express+, and mixed EGFP+/DsRed-Express+ cells were cultured at 37°C in α-MEM with 20% FCS plus one of the following cytokine mixtures: (a) 100 ng/ml SCF, 50 ng/ml Flt3-L, and 50 ng/ml Tpo; or (b) 100 ng/ml SCF and 5% pokeweed mitogen mouse spleen conditioned medium. The assay was set up in 24-well plates with 20,000 cells per well in 1.5 ml of growth medium at the beginning. At various time points after the initiation of culture, an aliquot was taken from each sample, a predetermined number of polystyrene microbeads (15 μm; Polysciences, Inc., Warrington, PA, http://www.polysciences.com) was added, and the mixture was analyzed by flow cytometry to simultaneously determine total numbers of cells and percentages of EGFP- and DsRed-expressing cells. Cells and beads were displayed as well-separated populations on a fluorescence-activated cell sorting (FACS) plot of forward scatter versus side scatter, which allows us to determine the ratio of beads to cells and thereby calculate the total number of cells in a given well .
Transduced Murine Bone Marrow Cells Expressing EGFP and DsRed-Express Fluorescent Protein Can Be Easily Distinguished by Fluorescent Microscopy and by Flow Cytometry
To evaluate whether EGFP and DsRed-Express fluorescent protein are suitable for use as expression tracers in HSC and progenitor cells, we made a pair of MSCV-based bicistronic retroviral constructs expressing either EGFP or DsRed-Express (Fig. 1A). We chose DsRed-Express because it is an improved variant of the Discosoma sp. red fluorescent protein with substantially enhanced solubility and maturation time , and its emission spectrum has minimal overlap with EGFP, which allows accurately discerning between EGFP- and DsRed-expressing cells [18, 22, 23]. In addition, it was the best available DsRed mutant at the time we started our experiments. Efforts were made to ensure the DNA sequences of these two retroviral constructs were identical except for the EGFP and DsRed-Express coding regions. This is important because identical backbone sequences allow us to unambiguously attribute the observed functional differences between EGFP- and DsRed-expressing cells to either the fluorescent proteins themselves or other mechanisms specifically induced by EGFP or DsRed-Express fluorescent protein. To experimentally determine whether we could track transduced BM cells expressing EGFP or DsRed-Express simultaneously and accurately in a mixed cell population, mouse BM cells were isolated and transduced with MIEG3 or M3Red retroviral vectors, and single and mixed cell populations were examined by confocal fluorescent microscopy as well as flow cytometry. The results demonstrated that transduced BM cells expressing EGFP or DsRed-Express in a mixed population can be easily distinguished by either method (Fig. 1B, 1C). When these two types of cells in a mixed population were analyzed by flow cytometry, relatively low compensation values were sufficient because the emission spectra of EGFP and DsRed-Express have minimal overlap.
Retrovirally Transduced Hematopoietic Stem Cells Expressing EGFP Are Stably Maintained in Recipient Mice for Nearly Their Lifespan, and All Recipient Mice Displayed Long-Term Lymphoid and Myeloid Multilineage Reconstitution
We utilized retroviral transduced mouse BM cells in a competitive repopulation assay [28, , , –32] to determine whether EGFP is suitable for use as a gene transfer and long-term expression tracer in HSCs and progenitor cells. We chose to transplant relatively low numbers of donor cells to recipient mice (6.75 × 105 total C57BL/6J donor cells per recipient mouse containing approximately 1.7 × 105 MIEG3-transduced cells) because it is more relevant to the clinical BM transplantation setting. Results shown in Figure 2 demonstrated that retrovirally transduced HSCs expressing EGFP were stably maintained and functioned normally in the recipient mice for nearly their lifespan. All recipient mice were engrafted by EGFP-expressing donor cells, and these mice appeared normal. Figure 2A shows representative FACS chimeric analyses of a peripheral-blood sample collected from a recipient mouse 6 months after transplantation. Due to relatively fewer donor stem cells injected, the resulting percentage of EGFP+CD45.2+ cells between individual recipient mice varied substantially (Fig. 2B); however, the mean percentage of EGFP+CD45.2+ cells for the entire experimental group remained steady (n = 6; Fig. 2C). This is expected on the basis of a series of elegant studies by Harrison et al [28, 33, 34], in which they demonstrated that transplanting fewer normal HSCs results in substantial fluctuation of engraftment levels between individual recipients, and that the amount of variation decreases as the number of stem cells increases, although mean engraftment levels remain the same. Our last chimeric analyses of peripheral-blood samples performed 15 months after transplantation showed that, on average, 24% total peripheral white blood cells in recipient mice expressed EGFP. This initial donor population before injection contained 25.2% EGFP-positive cells, and all five mice assessed at 15 months were EGFP positive. The majority of mice (four of six) lived for 22 months and appeared normal at sacrifice.
It is well established in the clinical laboratories that lymphoid, myeloid, and erythroid cell populations from BM and peripheral blood can be separated on a single FACS plot of CD45 (pan-leukocyte marker) fluorescence versus right-angle light scatter [35, –37]. Although this method may not precisely quantify the subpopulations of cells within a given lineage, it does provide fairly accurate estimates of percentages of cells among different lineages. To determine whether the EGFP-expressing stem cells retained their capability for long-term multilineage reconstitution, we analyzed and compared lymphoid and myeloid peripheral-blood cells repopulated by untreated CD45.1 competitor cells, and untransduced and transduced EGFP+ CD45.2 donor cells (all three cell populations were from the same recipient mouse) in recipient mice using CD45 fluorescence versus right-angle light scatter. Our results demonstrated that all recipient mice (n = 6) displayed long-term lymphoid and myeloid multilineage reconstitution by EGFP-expressing stem cells (Fig. 3). There was no significant difference in repopulating and differentiating potentials among the three types of cells analyzed (Fig. 3C). It is also apparent that a few detected nucleated erythroid progenitor cells were EGFP positive (Fig. 3B, arrow), suggesting that EGFP-expressing stem cells are capable of reconstituting the erythroid lineage as well. We refrained from quantifying the erythroid cell population because erythropoietin or other erythroid cytokines were not present during prestimulation and transduction, and the majority of the mature red blood cells were lysed prior to antibody staining. Furthermore, when quantitative real-time polymerase chain reaction (PCR) analyses were performed on an independent group of mice 7 months after transplantation, the results demonstrated that the average MIEG3 copy number was approximately 0.17 copies per genome, which mirrors the percentage of EGFP-expressing cells (14%) determined by flow cytometry (data not shown). Because by this time, the engraftment was well-established and the self-renewal and differentiation of the engrafted stem cells reached steady state, this indicates that, under the experimental conditions, most EGFP-expressing cells contain approximately one copy of MIEG3 retroviral vector per cell. Taken together, our results indicate that retrovirally transduced HSCs expressing EGFP are stable in transplant-recipient mice, and retain their long-term multilineage reconstitution potential. Introduction of EGFP into HSC did not have detectable deteriorative effects on these cells, and, therefore, EGFP is nearly an ideal long-term expression tracer for these cells. On the basis of these experiments, we propose that EGFP can now be used as a standard for evaluating other gene transfer and expression tracers for HSC.
In a Competitive Environment, Hematopoietic Stem Cells Expressing EGFP Are Stably Maintained; However, the Percentage of DsRed-Expressing Donor Cells Dramatically Declines in Recipient Mice over Time
To directly study the cooperative effects of two or more genes on stem and progenitor cells or to elucidate multicomponent protein machinery, it is necessary to use multiple expression tracers in a single cell or mouse. Having established that EGFP is nearly an ideal expression tracer for HSCs and progenitor cells, we next examined the suitability of DsRed-Express fluorescent protein as an expression tracer for these cells, and whether it could be used in combination with EGFP for multicolor tracking of these cells in competitive repopulation stem cell assay. We separately transduced donor BM mononuclear cells with MIEG3 or M3Red retroviral vectors, and then mixed a portion of the transduced cells together so that the resulting population contained fixed percentages of EGFP- and DsRed-expressing cells (ratios = 1.38–1.9). Untransduced control cells, MIEG3 or M3Red singly transduced cells, or the mixed MIEG3/M3Red cells were combined with freshly isolated BoyJ competitor cells and transplanted into lethally irradiated recipient mice. We performed two totally independent transduction and transplantation experiments, and the results are summarized in Figure 4 and Table 1. Our results demonstrate that, although HSCs expressing EGFP are stably maintained in the recipients, the percentage of donor cells expressing DsRed-Express fluorescent protein in recipient mice decreases over time. This decrease was more dramatic in a competitive setting, in which mixed EGFP- and DsRed-expressing donor cells were cotransplanted into a single recipient (Fig. 4B). In such a competitive environment, EGFP-expressing cells outcompeted DsRed-expressing cells in almost every recipient mouse analyzed (13 of 14; one mouse exhibited approximately equivalent levels of relatively low engraftment by EGFP- and DsRed-expressing cells). It has been established that, when equal number of normal bone marrow cells that contain a relatively low number of stem cells is transplanted into each of the recipient mice, the engraftment in individual recipients varies substantially (large standard deviations), but the mean levels remain constant [28, 33, 34]. Thus, decreased mean engraftment levels in recipient mice transplanted with DsRed-Express donor cells indicate that DsRed-Express fluorescent protein directly or indirectly impairs viability or self-renewal of HSCs. In mice transplanted with M3Red singly transduced cells, the percentage of DsRed-expressing cells steadily declined during the first 3 months post-transplant (10 of 11 mice), and although this decrease continued over time in the majority of the recipient mice, sometimes a few adapted DsRed-expressing stem cell clones were able to expand in some recipient mice thereafter (4 of 11). Collectively, these results demonstrate that DsRed-Express is not a suitable expression tracer for the long-term HSC since it interferes with the functionality of these cells.
Table Table 1. Averaged engraftment by untransduced, transduced EGFP- or DsRed-expressing cell, and mixed EGFP/DsRed cell populations
The EGFP-Expressing Cells Grow Robustly, Whereas DsRed-expressing Cells Do Not Under In Vitro Culture Condition Favoring Preservation of HSC, and Retrovirus Preferentially Transduces BM HSC and Progenitor Cells
To complement and further corroborate our in vivo competitive repopulation findings, we conducted a series of in vitro culture experiments using purified EGFP- and DsRed-expressing cells. Low-density mononuclear BM cells were transduced, and cells expressing EGFP and DsRed-Express fluorescent protein were sorted by FACS. The initial purities of sorted EGFP- and DsRed-expressing cells were, respectively, 99% and 95.2%. Then untransduced control cells, purified EGFP+ and DsRed-Express+ cells, and mixed EGFP+/DsRed-Express+ cells were cultured in triplicate under two different culture conditions; one favors the preservation and growth of stem and primitive progenitor cells (SCF, Flt3-L, and Tpo), whereas the other promotes the proliferation and differentiation of primitive progenitor cells (SCF plus 5% pokeweed mitogen mouse spleen conditioned medium). Each individual culture was initiated with 20,000 cells, and at various time points after starting the culture, the total number of cells and the percentages of EGFP- and DsRed-expressing cells were determined by flow cytometry. Results shown in Figure 5 demonstrated that, under the culture condition favoring preservation and growth of HSCs (Fig. 5A), cells expressing EGFP grew robustly, whereas cells expressing DsRed-Express did not. Similar to our in vivo findings, this difference was also apparent in the competitive setting, in which EGFP- and DsRed-expressing cells were cocultured. This was evident by a steady increase in ratios of EGFP- versus DsRed-expressing cells (Fig. 5A, numbers above the graph bars). In contrast, under culture conditions similar to the standard colony assays that favor proliferation and differentiation of hematopoietic progenitor cells, both EGFP- and DsRed-expressing cells grew quite well, and there was no detectable difference between these two types of cells (Fig. 5B). These results complement and provide additional support for our in vivo findings, and strongly suggest that DsRed-Express fluorescent protein interferes with viability and/or growth of hematopoietic stem cells. Therefore, this variant of DsRed fluorescent protein (DsRed-Express) is not a proper expression tracer for HSCs.
In addition, our results also directly demonstrated that retrovirally transduced BM cells were more enriched for HSCs and progenitor cells compared to untransduced control cells, indicating that retrovirus preferentially transduces BM stem and progenitor cells. This is apparent when comparing population growth kinetics of MIEG3 transduced cells with that of untransduced cells in Figure 5A, as well as population growth kinetics of all types of transduced cells with that of untransduced cells in Figure 5B.
In this study, we have systematically evaluated whether two fluorescent proteins, EGFP and DsRed-Express, are suitable as expression tracers for HSC and progenitor cells using the competitive repopulation assay as well as the in vitro cell growth assay. The results established that retrovirally transduced mouse hematopoietic stem cells expressing EGFP are maintained nearly throughout the lifespan of the transplant-recipient mice, and retain their long-term multilineage repopulating potential. Moreover, our results directly demonstrated that the retrovirally transduced BM cell population was more enriched for HSCs as compared to untransduced control cells, and that transduced BM stem and progenitor cells expressing EGFP grew robustly in culture. Therefore, EGFP itself has no detectable deteriorative effects on HSCs, and is nearly an ideal long-term expression tracer for hematopoietic cells. We propose that EGFP can be used as a standard gene expression tracer for HSCs and progenitor cells. In contrast, the percentage of donor cells expressing DsRed-Express fluorescent protein in transplant-recipient mice decreased over time, suggesting that DsRed-Express fluorescent protein directly or indirectly interferes with the viability and/or self-renewal of HSC. This decrease was more dramatic in a competitive setting, in which mixed EGFP- and DsRed-expressing donor cells were cotransplanted into a single recipient. Similar and complementary results were obtained from the in vitro growth assay using FACS-purified EGFP- and DsRed-expressing cells. Thus, DsRed-Express fluorescent protein is not a suitable expression tracer for long-term hematopoietic stem cells.
We do not yet know exactly how DsRed-Express fluorescent protein decreases viability and/or inhibits the growth of HSC. This most likely results from the obligate tetramerization of the DsRed-Express protein, as well as its tendency to form even higher order of oligomers, although at reduced levels compared to wild-type DsRed protein . It has been reported that embryonic stem cells and mice expressing wild-type or earlier variants of DsRed fluorescent protein could not be established and maintained due to nonviability of the cells, possibly as a result of the formation of obligate tetramer or higher-order aggregates by DsRed proteins [39, –41]. Moreover, high-resolution imaging of embryonic stem cells and cells in chimeras that express DsRed. T3, a variant similar to DsRed-Express, reveals punctate staining in a perinuclear/Golgi region . On the basis of numerous functional and gene expression studies of hematopoietic as well as embryonic stem cells, stem cells in general have relatively large transcriptomes, and their gene expression networks and intracellular molecular machinery are highly orchestrated and regulated. One could envision that aggregation of a highly expressed protein in the cytoplasm not only disrupts normal localization and protein-protein interactions of some of the critically important proteins but also pose many spatial obstacles that interfere with intracellular trafficking and sorting. In addition, it has been shown recently that, when the monomeric/weak dimeric form of GFP was converted to an aggregation-prone protein through C-terminal addition of a short peptide, the aggregated GFP became cytotoxic. Expression of this aggregating GFP variant in Caenorhabditis elegans and mammalian HEK293 cells results in formation of stable perinuclear deposits, and induction of its expression in C. elegans muscle leads to rapid paralysis, demonstrating its in vivo toxicity . Furthermore, aggregating proteins or peptides have been shown to be associated with many neurodegenerative diseases, and transgenic expression of these proteins in cultured cells, in C. elegans, and in mice leads to the formation of intracellular aggregates and associated toxicity [42, –44].
Although we can not completely rule out the possibility that the decrease in cells expressing DsRed-Express in transplanted mice was caused by retroviral silencing, this is very unlikely on the basis of the following. First, the two retroviral constructs used here, MIEG3 and M3Red, have the same sequences except for the coding regions for EGFP and DsRed-Express, and the same target cell population was used for transduction. Therefore, one would expect a similar degree of silencing for MIEG3 and M3Red retroviral vectors. Because it is known that retroviral silencing is usually associated with the viral sequences and/or certain promoters and regulatory regions , it is unlikely that DsRed-Express coding sequence is particularly prone to silencing. Moreover, when we cultured the FACS-purified EGFP- and DsRed-expressing cells under in vitro conditions favoring preservation and growth of HSCs and progenitor cells, the cells expressing EGFP grew robustly, whereas cells expressing DsRed-Express did not. Over time, the total number of DsRed-expressing cells was far less than that of cells expressing EGFP (Fig. 5A). This experiment demonstrates that, under the ex vivo conditions, DsRed-Express impairs the viability or growth of hematopoietic stem and progenitor cells, and silencing is not a major contributor to the decrease in cells expressing DsRed-Express because the starting cell populations are purified so that any subsequently silenced cell populations (EGFP or DsRed-Express) can be readily identified. Although there was a slight increase in the number of DsRed-Express negative cells in the sorted DsRed-Express population over time as compared to that in the sorted EGFP population, this increase has little impact on the total number of cells. Finally, if the DsRed-expressing HSCs were as good as EGFP-expressing HSCs and the decrease in number of DsRed-expressing cells resulted from silencing, one would expect a similar degree of decrease in DsRed-expressing cells in mice transplanted with DsRed-Express cells alone and with mixed EGFP/DsRed-Express cells, which is not what we observed. Taken together, the evidence strongly suggests that gene silencing of M3Red retroviral vector does not play a major role in the reduction of cells expressing DsRed-Express in transplanted mice. On the other hand, we would also like to point out that a small initial population of silenced cells can become dominant over the long term in recipient mice due to in vivo selection.
Although transgenic mice expressing GFP have been available for sometime now, and are excellent and valuable resources for cell tracking [46, –48], we feel that it is not a trivial issue to evaluate whether EGFP is a suitable gene expression tracer for long-term hematopoietic stem cells. First, although transgenic GFP cells or mice are exceptional for cell tracking, they cannot be used as gene transfer and expression tracers. Secondly, continuous expression of GFP in cells or mice allows selection of the most adapted and tolerated cell or mouse populations over generations. This is quite different from the situation in which a gene of interest is introduced into cells de novo, and both its immediate and long-term effects are investigated afterwards. We noticed that, sometimes, a few donor stem cell clones expressing DsRed-Express protein were able to adapt and expand in a portion of transplant-recipient mice approximately 4 months post-transplantation. Clearly, expression of DsRed-Express protein in these stem cells was well tolerated, whereas that is not the case for the majority of the stem cells expressing the same protein. It is possible that, perhaps after directed selection of transfected embryonic stem cells, transgenic mice expressing DsRed-Express protein could also be made.
Proteins that fluoresce at the red or far-red wavelengths are proteins of choice for in vivo imaging applications because mammalian tissues have higher transparency and display lower background autofluorescence at these longer wavelengths, which permits high-resolution, deeper tissue imaging [15, 16]. Moreover, DsRed-based fluorescent proteins are genetically and spectrally distinct from other available fluorescent proteins; thus, they could be used in conjunction with other fluorescent proteins for multicolor tracking and imaging as well as for fluorescence resonance energy transfer experiments [15, 39, 49]. It is very important to note that we have studied only one of the improved DsRed variants, DsRed-Express; thus, one cannot and should not conclude that expression of any other further improved variants of Discosoma sp. red fluorescent protein will have the same effects on hematopoietic stem cells. In fact, continuous efforts to engineer and optimize red fluorescent proteins have led to development of the next generation of DsRed-based monomeric or tandem dimeric fluorescent proteins, namely mCherry, DsRed-monomer (Clontech, Mountain View, CA, http://www.clontech.com), and tdTomato [13, 22, 50, 51]. These newly developed DsRed variants seem to retain most of their desired properties without apparent cytotoxic side effects. They represent very good candidates for gene transfer and expression tracers in HSCs and progenitor cells. However, whether any one or all of them are suitable as gene transfer and expression tracers for long-term repopulating hematopoietic stem cells or other types of cells will have to be tested and verified experimentally, perhaps using EGFP as a standard. Nevertheless, a variety of spectrally distinct fluorescent proteins will undoubtedly become powerful and invaluable tools for exploring multicomponent protein complexes and spatiotemporal dynamics of intracellular protein networks, and for noninvasive monitoring of biochemical processes and/or cells in real time in a whole organism. One could even foresee that, once it is proven that certain fluorescent proteins are noncytotoxic and noninterfering, their encoding genes can be knocked in to genomic regions of genes of interest. Then, one could visualize in real time where, when, and how these gene products participate in various biochemical and biological processes during development as well as in normal and disease states.
The authors indicate no potential conflicts of interest.
This work was supported by a project in PO1 HL 053586, and U.S. Public Health Service Grants RO1 HL 56416 and DK 53674 from the National Institutes of Health (H.E.B.). We thank the Indiana University Vector Production Facility, the Indiana Center for Biological Microscopy core facility, and the Flow Cytometry Facility.