- Email: [email protected]
Spontaneous switch from Aγ- to β-globin promoter activity in a stable transfected dual reporter vector
Blood Cells, Molecules, and Diseases 34 (2005) 174 – 180 www.elsevier.com/locate/ybcmd
Spontaneous switch from Ag- to h-globin promoter activity in a stable transfected dual reporter vector Giovanni Migliaccioa,*, Angela Di Baldassarreb, Cristina Di Ricob, Antonella Di Noiaa, Betty Nakamotoc, Hua Caoc, Eva Skarpidic, Anna Rita Migliacciod a
Gene and Cell Therapy, Department of Cell Biology and Neuroscience, Istituto Superiore di Sanita`, Viale Regina Elena 299, I-00161 Rome, Italy b Department of Biomorphology, University bG D’AnnunzioQ, Chieti, Italy c Department of Medicine, University of Washington, Seattle, WA 98195, USA d Department of Hematology, Oncology and Molecular Medicine, Istituto Superiore Sanita`, Rome Submitted 28 July 2004; revised 21 November 2004 Available online 20 January 2005 (Communicated by Dr. G. Stamatoyannopoulus, 13 December 2004)
Abstract Here it is analyzed the expression of a mini locus dual reporter construct composed by a micro-LCR and by the promoters for Ag- and hglobin gene, each one linked to a different Luciferase, in stably transfected GM979 cells for as long as 1–2 years from transfection. The transfected GM979 cells rapidly (within 1 month) evolved into a stable population which expresses constant levels of reporters for more than a year of continuous bulk culture. No silencing of the inserted construct was observed over time. In contrast, after 1 month, the reporter activity (both from Ag- and h-promoter) expressed per cell increased over time. The analysis of the Luciferase contained in single cell clones indicated that the higher reporter activity was due to increased gene expression per cell rather than to clonal selection of the most expressing clones. Since the activity driven by the h-promoter increased 10-fold more than that driven by the Ag one, the ratio between Ag-driven/(Agdriven + h-driven) reporter activity in the cells decreased after 1 month and became similar to the g/(g + h) globin mRNA ratio expressed by adult erythroid cells. Moreover, although both cells from early and late bulk culture responded to incubation with butyric acid, a known inducer of fetal globin gene expression, by increasing the reporter activity driven by the Ag-promoter, only cells from late bulk culture decreased, as normal primary erythroblasts do, the activity of the reporter driven by the h-promoter. These results suggest that the rapid changes in activity driven by the Ag- and h-globin promoters occurring during the first month after transfection may represent a novel in vitro model to study epigenetic regulation of the Ag- and h-promoter during the fetal to adult hemoglobin switch and confirm GM979 cells stably transfected with the dual reporter construct as a reliable assay for automated screening of chemical inducers of fetal globin gene activation. D 2004 Elsevier Inc. All rights reserved. Keywords: Hb switch; GM979 cell line; Epigenetic regulation; Dual Luciferase reporter assay
Introduction Thalassemia is a severe genetic disease caused by coinheritance of defective a- or h-globin alleles resulting in deficient globin expression. In the case of h-thalassemia, the disease is due to deletion or abnormal regulation of the
* Corresponding author. Fax: +39 06 4990 2530. E-mail address: [email protected] (G. Migliaccio). 1079-9796/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bcmd.2004.11.005
h-globin gene. The resulting inefficient h-globin expression leads to accumulation of unpaired, insoluble a-globin chains in red cells. The consequent accelerated destruction of these cells induces ineffective erythropoiesis and severe anemia. The development of transfusion therapies coupled with administration of iron-chelating agents has greatly ameliorated the treatment and the quality of life of hthalassemia patients. The overall treatment of this disorder, however, is still unsatisfactory. In fact, while the incidence of the disease is steadily decreasing in western countries, it
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
remains high in economically challenged countries that do not have easy access to the technologies required to properly manage these complex therapies. The observation that some of the homozygote carriers of h-globin gene deletions are asymptomatic because of persistent expression in adult life of the fetal-specific h-like genes, the gglobin genes (High Persistence of Fetal Hemoglobin Syndrome, HPFH), has inspired numerous studies aimed to identify chemical substances capable to pharmacologically re-activate g-globin expression in patients with hthalassemia. One of the pharmacological agents identified so far is hydroxyurea. Phase I–II clinical trials with hydroxyurea on h-thalassemia [1] and sickle cell [2,3] patients have been successfully completed, and the drug is presently considered to be introduced in the current medical practice for these disorders. Hydroxyurea, however, is a compound not devoid of counter indications [4]. Therefore, the search for additional and less toxic agents is still in progress [5]. The ideal candidate for pharmaceutical treatment of hthalassemia is a cheap chemical that could be administered orally, would be effective within reasonable doses, and would present no or little counter indications. The efficiency of the search for these agents is strictly dependent on the availability of a surrogate assay to measure the potentiality to reactivate g-globin gene expression in vivo. Such surrogate assays should be rapid, cheap, reliable, and suitable to be adapted for automated screening. Assays of induction of g-globin gene expression that use as target cells human erythroblasts and that determine the cellular globin content by HPLC represent the tests that would most reliably mimic the possible effects a chemical might induce in vivo [6]. These assays, however, as precise and indicative of clinical efficacy as they might be, are cumbersome and not easily adaptable for automatic screening. More suitable for this purpose would be surrogate assays based on murine erythroleukemia cell lines transfected with reporter constructs containing regions of the human g-globin promoter linked to appropriate reporter genes. The accuracy of these assays depends on how closely the promoter regions that drive the expression of the reporter gene mimic the physiology of g-globin gene activation in primary erythroid cells. Recently, it has been developed an assay based on the murine erythroid GM979 cell line stably transfected with a construct containing a human mini globin locus composed by a micro-LCR and by the promoters for Ag- and h-globin, each one linked to a different Luciferase [7,8]. The ratio between the expression of the two Luciferases reflects the relative activity of the two promoters and can be measured easily in to be automated assays. The suitability of this assay, however, for automated screening of chemical inducers of globin gene expression depends on how stable and consistent is the expression of the reporter in transfected cells over time. Here it is characterized the expression over time of the mini locus dual reporter construct in stably transfected
175
GM979 cells, from the first month up to 1–2 years from transfection. The results obtained indicate that GM979 ALCRhprRluc AgprFluc cells undergo an initial period of stabilization after which they express high levels of Luciferase activity driven both by the Ag- and hpromoter. The Ag-driven/(Ag-driven + h-driven) reporter activity ratio, that was similar to the fetal-globin ratio at early bulk cultures, became similar to adult globin gene ratio at later time points. Butyric acid, a known fetal hemoglobin inducer, increased expression of the reporter from the Ag-promoter both in early and in late bulk cultures but decreased that from the h-promoter only in late bulk cultures. As such, GM979 ALCRhprRluc AgprFluc cells from late bulk cultures represent a reliable assay for automated screening of chemical inducers of globin gene activation.
Materials and methods Establishment of the GM979 lLCRbprRluc AcprFluc cell line The murine erythroleukemic GM979 cell line had been established from fetal liver cells [9] and was chosen for this study because its phenotype includes expression both of minor and major Hb. The GM979 ALCRhprRluc AgprFluc cell line was obtained by lipofecting GM979 cells with a ALCRhprRluc AgprFluc plasmid that contains a 3.1-kb ALCR cassette including the DNAse I hypersensitive core of the 5V hypersensitive sites HS1, HS2, HS3, and HS4, linked to 315-bp of the human h-globin promoter and 1.4-kb of the A g-globin promoter driving the Renilla (R) and the Firefly (F) Luciferase gene, respectively [8]. A separate plasmid containing the NeoR gene was also included in the cotransfection procedure. A stably transfected cell line was selected on the basis of neomicine resistance, but no formal clonal selection was attempted. Passage of GM979 lLCRbprRluc AcprFluc cells in bulk culture and isolation of single cell-derived clones The GM979 ALCRhprRluc AgprFluc cell line was maintained in RPMI 1640 (Gibco, USA) supplemented with 10% of heat inactivated fetal calf serum (FCS, Hyclone, Logan UT) and 400 Ag/mL of G418 (Gibco) and incubated at 378C in a humidified incubator containing 5% CO2 in air. To avoid random clonal selection, cells were routinely sub-cultured every 4–5 days by seeding at least 5 105 cells/mL in each passage. No changes were noticed in the growth characteristics of the GM979 ALCR hprRluc AgprFluc cells over more than 1 year of continuous growth in bulk culture. Single cell-derived clones of GM979 ALCRhprRluc AgprprFluc cells were obtained by standard limiting dilution procedures. Briefly, GM979 ALCRhprRluc AgprFluc cells were diluted in complete medium at a concentration of 10
176
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
cells per 300 AL, and 100 AL of this suspension was plated in each well of 96-multiwells plates (a total of 6 plates). The plates were then incubated at 378C in a humidified incubator containing 5% CO2 in air. After 1 week, the wells were analyzed under an inverted microscope to identify those in which cell proliferation had occurred. After two additional weeks, the cells contained in each positive well were harvested, counted, and passed into individual flasks where they were cultured until they reached a number sufficient for performing the Dual Luciferase assay (usually 4 weeks) (see later). Dual Luciferase assay The Ag-driven Firefly (Ag-F) and h-driven Renilla (h-R) Luciferase activities were measured with the Dual Luciferase Reporter Assay System (Promega, Madison, WI, USA), as described by the manufacturer. Briefly, triplicate samples, containing 2 105 to 1 106 cells each, were centrifuged and washed twice in phosphate-buffered saline. The cell pellet was then lysed with 100 AL of Lysing Buffer (Dual Luciferase Kit, Promega) and the cellular debris removed from the supernatants by centrifugation. Aliquots of 20 AL of the supernatants were added to the Renilla and Firefly Luciferase substrate solution and the amount of light emitted by the sample measured with a Lumat LB9507 Luminometer (EG&G BERTHOLD, USA). The numeric values for h-R were doubled to compensate for the intrinsic lower luminosity, and both Luciferase activities were expressed in Arbitrary Fluorescence Units (AFU). As the translation efficiencies, as well as the mRNA and protein half-life, of the two Luciferase genes are all similar, the ratio between the Ag-F and h-R AFU corresponds not only to the amount of Ag-F and h-R proteins produced by the cells, but also to the ratio between the transcription rates driven by the human A g and h-globin promoters. The results are presented both as amount of Luciferase activity per cells (either Ag-F or h-F AFU/total number of cells in the sample) and as ratio of reporter activities [Ag-F AFU/(Ag-F AFU + h-F AFU)].
Results Increased Luciferase activity per cell, both from Ac- and b-promoter, in GM979 lLCRbprRluc AcprFluc cells over time The amount of Ag-F and h-R Luciferase activity expressed over time by GM979 ALCRhprRluc AgprFluc cells is presented in Fig. 1. The activity of both the Ag-driven Firefly and the h-driven Renilla expressed by the cells progressively increased during the culture (Fig. 1). The amount of the h-driven Renilla Luciferase, however, increased much more than that of the Ag-driven Firefly Luciferase. In fact, the average activity of Ag-F Luciferase per cell increased by 2-fold (from 0.05 F 0.02 AFU at 2 months to 0.10 F 0.06 AFU at 11 months) while that of the h-R Luciferase increased by 8-fold (from 4.3 F 1.7 AFU at 2 months to 35.96 F 19.16 AFU at 9 months). After 14 months of continuous bulk culture, the average Luciferase activity per cell had further increased to 0.28 F 0.14 AFU for Firefly and 58.13 F 20.95 AFU for Renilla, with an average
Fetal Hb (HbF) induction Butyric (0 up to 1.5 mM) and proprionic (0 and 5 mM) acids were added to three separate pools of 2 106 GM979 ALCRhprRluc AgprFluc cells resuspended in fresh culture medium. After 4 days of incubation at 378C and 5% CO2, the number of alive cells was counted by Trypan Blue exclusion. Identical numbers of live cells for test point were dissolved in Lysing Buffer and their Ag-F and h-R activity measured with the Dual Luciferase Assay, as described above. Statistical analysis Statistical analysis of the results was obtained with the Origin 6.1 software (Microcal, Northampton, NJ, USA).
Fig. 1. (A and B) Luciferase activity driven by the human h-globin (A) and A g- (B) promoters in bulk cultures of GM979 ALCRhprRluc AgprFluc cells over time. Data are reported as AFU per alive cell. The activity per cell of both Luciferases significantly increased ( P b 0.001) over time in more than 1 year of bulk culture. Activity driven by the h-promoter, however (r = 0.7), increased 4-fold more than that driven by the Ag-promoter (r = 0.52).
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
increase with respect to the early time points after transfection of approximately 5- and 10-fold, respectively. In both cases, the changes in Luciferase activity expressed by the cells over time fitted a linear regression curve (Figs. 1A and B, P b 0.0001 for both regressions), indicating a direct proportionality between the content of Ag-F and h-R Luciferases per cell and the time the cell line had spent in bulk culture. The different kinetics of the changes in Ag-F and h-R Luciferase activity expressed by GM979 ALCRhprRluc AgprprFluc cells over time were reflected by changes in the Ag-F/ (Ag-F + h-R) ratio (Fig. 2). The most striking change in the A gF/AgF + hR ratio was observed during the first 2 months after transfection. The Ag-F/(Ag-F + h-R) ratio, that had ranged from 3% to 8% at early time points, became as low as 0.6% from 100 days of bulk culture thereon (Fig. 2). The kinetics of the changes observed in the Ag-F/(Ag-F + h-R) ratio over time were best fitted with a curve of exponential decay of first grade (Fig. 2). These results indicate that the changes in Ag-F/(Ag-F + hR) ratio expressed by GM979 ALCRhprRluc AgprFluc cells over time in bulk culture are not due to decreased transcription driven from the human Ag-promoter (and therefore to its selective silencing) but rather to differential activation of the expression driven by the Ag- and hpromoter during the first 50 days of culture. Luciferase activities in single cell-derived clones of the GM979 lLCRbprRluc AcprFluc cell line In order to clarify if the changes in the Ag-F and h-R activities produced by the cells over time were a consequence of clonal selection, we analyzed Ag-F and h-R activity expressed by single cell-derived clones obtained
177
Fig. 3. Relative Ag- vs. h-driven Luciferase activity expressed by single cell-derived clones obtained from bulk cultures at early (2 months, dots) and late (11 months, squares) time points of bulk culture. Each dot represents the activity expressed by an individual clone. A positive linear relationship ( P b 0.0001) was observed between the relative Firefly and Renilla Luciferase activity per cell expressed by clones isolated at 11 months, but not in those obtained at 2 months of bulk culture.
from the GM979 ALCRhprRluc AgprFluc cell line at early (within 1 month) and late (12 months) time points of bulk culture (Fig. 3 and Table 1). No difference was found in the cloning efficiency of GM979 ALCRhprRluc AgprFluc cells at the two time points. In both cases, only one-fifth of the wells contained proliferating clones after 3 weeks of culture. The Ag-F and h-R Luciferase activity expressed by single cell-derived clones obtained from early and late time points of bulk culture ranges over a large spectrum of values (Fig. 3). There is no overlap in the relative activity of Ag-F and h-R Luciferase expressed by the clones obtained at the two time points. At the early stage of bulk culture, the average activity per cell for both Firefly and Renilla Luciferases ranges in intensity over more than four log with no correlation between the relative amount of Ag-F and h-R expressed by each clones (Fig. 3, Early). In the clones obtained at 12 months, the range of activities of the two Luciferases is still spread over more than 4 logs. However, there is a significant ( P b 0.004) positive linear correlation Table 1 Mean (FSEM) Firefly and Renilla Luciferase activity, and the corresponding Ag-F/(Ag-F + h-R) ratio, expressed by single clones derived by limiting dilution from GM979 ALCRhprRluc AgprFluc cells at early (2 months) and late (11 months) times of bulk culture A
Fig. 2. Ag-F/(Ag-F + h-R) ratio in bulk cultures of GM979ALCRhprRluc AgprgprFluc cells over time. The Ag-F/(Ag-F + h-R) Luciferase activity is plotted against time from transfection. The exponential decay curve of first grade that significantly (Chi2 = 0.0002) fits the distribution of the experimental points is indicated as dotted line.
Early time Clones Bulk culture Late time Clones Bulk culture
g-Firefly
h-Renilla
A
g-F/(Ag-F + hR)
0.15 F 0.11 0.06 F 0.02
2.99 F 1.07 6.54 F 2.6
0.06 F 0.04 0.03 F 0.01
0.02 F 0.01 0.27 F 0.03
12.9 F 4.6 55.4 F 4.0
0.007 F 0.001 0.005 F 0.001
The levels of activity expressed by the corresponding cells in bulk cultures are presented as well.
178
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
between the relative activity expressed from the two promoters by each single clone (Fig. 3). Although the amount of Ag-F and h-R Luciferase expressed by cells from the bulk cultures is higher than the average amount of Luciferase expressed by single clones (Table 1), there is no significant difference in the Ag-F/(AgF + h-R) ratio measured in cells in bulk culture and those expressed by the corresponding single clones (Table 1). These results indicate that during the time in bulk culture the cells coordinately increase the expression of both reporter genes. Differential effects of butyric acid on Ac-and b-reporter activity in early and late GM979 lLCRbprRluc AcprFluc cells The modifications observed in the steady-state level of hand Ag-globin promoter activity in GM979 ALCRhprprRluc AgprFluc cells over time prompted us to investigate whether compounds known to induce g-globin gene expression would activate the Ag-globin promoter in these cells at any stage of the culture. Since previous studies had shown that short chain fatty acids specifically induce g-globin transcription, we treated GM979ALCRhprRluc AgprFluc cells obtained at early and late stage of the bulk culture for 4 days with increasing doses of butyric acid and compared the level of Ag-F and h-R activities expressed in the different conditions (Table 2). Incubation with butyric acid of cells at the early stage of the culture (b2 months) strongly increased A g-F activity. The maximal effects were exerted at the concentration of 1 mM. Incubation with butyric acid scarcely affected the h-R activity expressed by these cells (Table 2). As a result, butyric acid induced an evident, concentrationdependent increase of the Ag-F/(Ag-F+h-R) ratio in early GM979ALCRhprRluc AgprFluc cells (Table 2, Ag-F/(Ag-F + h-R)). On the other hand, GM979 ALCRhprRluc AgprFluc cells at 11 months of bulk culture had a clearly different response to butyrate. At this point in time, butyric acid modestly increased Ag-F activity, while the drug had a strong inhibitory effect on the h-R activity that, however, remained still much higher than that expressed by cells at early time points. As a
result, although butyric acid increased the Ag-F/(Ag-F + h-R) ratio also in GM979 ALCRhprRluc AgprFluc cells at 11 months of bulk culture, the effect was lower than that induced in cells which had been in culture only for 2 months. The reduced response to butyric acid of the Ag globin promoter in late (11 months) bulk culture is not due to silencing of the promoter. In fact, the amount of Firefly Luciferase produced in untreated late (11 months) cells is higher than that produced by early (2 months) bulk culture cells (Figs. 1 and 3). Furthermore, late bulk culture cells increased by 40-fold (0.09 vs. 3.00, P b 0.05) Firefly Luciferase activity upon treatment with propionic acid (5 mM), another compound known to induce HbF [10] (Table 3). Both propionic and butyric acid have an inhibitory effect on cell proliferation as demonstrated by the reduced number of cell doublings occurring during the 4 days of incubation with the inducer (Tables 2 and 3).
Discussion To provide for a fast and reliable assay for chemical inducers of human HbF, the murine GM979 cell line was transfected with a dual Luciferase construct (ALCRhprRluc AgprFluc) [8] that allows direct comparison of the reporter activity driven by the human Ag- and h-globin promoters [7,11]. Here we show that the resulting GM979 ALCRhprRluc AgprFluc cell line maintains the ability to produce both Firefly and Renilla Luciferases over a period of more than 1 year of continuous bulk culture (Figs. 1 and 2). Over the same time period, it maintains its ability to respond to butyric acid by increasing the Luciferase driven by the Ag-promoter (Table 2). Changes were, however, observed in the pattern of reporter activity expressed by GM979 ALCRhprRluc AgprFluc cells over time. The Ag-F/(Ag-F + h-R) ratio decreased shortly after transfection, switching from a fetal-neonatal-like ratio (8–2%), that had been observed in the first 2 weeks of culture, to an adult-like ratio (0.1–0.01%) after the first 2 months of bulk culture. Thereafter, it remained stable for more than 1 year in culture (Fig. 1 and data not shown).
Table 2 Effects of increasing concentrations of butyric acid on Ag-driven (Ag-Firefly) and h-driven (h-Renilla) Luciferase activity, as well as on the Ag-F/(Ag-F + h-R) ratio, in GM979 ALCRhprRluc AgprFluc cells at 2 (Early phase) and 11 (Late Phase) months of bulk culture Early phase
Late phase
Butyric acid
A
Controls 0.5 mM 1 mM 1.5 mM Controls 0.5 mM 1 mM 1.5 mM
0.06 0.68 1.61 1.66 0.17 0.30 0.33 0.52
g-Firefly F F F F F F F F
0.01 0.10* 0.13* 0.16* 0.04 0.09* 0.07* 0.13*
h-Renilla 3.71 F 11.24 F 5.12 F 4.91 F 58.19 F 42.95 F 21.20 F 16.53 F
0.74 1.94* 2.1* 2.5* 19.27 12.39* 5.0* 4.2*
A
g-F/(Ag-F + h-R)
0.02 0.07 0.24 0.25 0.002 0.003 0.009 0.017
F F F F F F F F
0.01 0.02 0.02* 0.02* 0.0001 0.001 0.001* 0.003*
Cell doublings 3.77 1.95 0.89 0.46 3.66 1.45 0.29 1.36
F F F F F F F F
0.04 0.22* 0.13* 0.14* 0.09 0.11* 0.12* 0.32*
The effects of butyric acid on cellular proliferation are expressed as number of cell doublings during the 4 days of incubation (a negative number indicate cell death). Data are presented as mean F SEM of at least 4 separate experiments performed in triplicate. * P b 0.05.
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
179
Table 3 Effects of propionic acid (5 mM) on Ag-driven (Ag-Firefly) and h-driven (h-Renilla) Luciferase activity, as well as on the Ag-F/(Ag-F + h-R) ratio, in GM979 ALCRhprRluc AgprFluc cells at 11 (Late Phase) months of bulk culture Controls Propionic acid
Ag-Firefly
h-Renilla
0.09 F 0.01 3.00 F 0.09*
21.5 F 5.40 79.8 F 49.1
A
g-F/(Ag-F + h-R)
0.003 F 0.001 0.027 F 0.006*
Cell number (106) 1.37 F 0.07 0.087 F 0.01
The cell number recovered after the 4 days of incubation is also presented, as a measure of toxicity. Data are presented as mean F SEM of at least 3 separate experiments. * P b 0.05.
Unlike most of the MEL cell lines, the murine GM979 cell line co-expresses both murine embryonic (qy) and adult (hmaj) globin genes and allows expression of human fetal globin genes in hetero-specific hybrids constructed with nonerythroid human cells [12]. These observations have suggested that these cells contain trans-activating factors that allow activation of silenced human g-globin genes in humanmouse hybrids [12] and that are able to keep the human gglobin genes in an inducible state. However, the relatively high expression of the human g-globin genes in the heterospecific hybrids was transient and disappeared after a few weeks. The g-globin expression in these clones, however, remained inducible by chemical agents, such as butyric acid, indicating that the endogenous regulatory elements, although favoring an adult globin expression phenotype, do not limit gene expression driven by the g-globin promoter [13]. In normal primary erythroid cells, as well as in murine cell lines carrying the complete human h-globin locus, the switch observed over time from a fetal to an adult Hb pattern implies an autonomous silencing of the g-promoter and competitive transcription from the h-globin one. In GM979 ALCRhprRluc AgprFluc cells, the decrease in the Ag-F/(Ag-F + h-R) ratio observed in late bulk culture is not mediated by silencing of the Ag-promoter. In fact, the amount of both Ag-F and h-R Luciferases per cell increases steadily during the 10 months of bulk culture (Figs. 1A and B). However, while the average amount of h-R per cell increases over time by 10-fold, that of the Ag-F increases only 3-fold. It is, therefore, a different rate in the increase of promoter activity over time, rather than gene silencing, that confers an adult-like globin ratio to the ratio between the Luciferase activities driven by the two promoters at the late bulk cultures. A possible explanation of this phenomenon could be progressive selection over time in culture of those clones that preferentially express an adult-type of globin gene regulation. However, the amount of both Ag-F and h-R Luciferases expressed per cell by single clones isolated from the bulk culture at early and late time point is highly heterogeneous and spans a range of four AFU logs (Fig. 3). No decrease in the overall variability of the Luciferase activity expressed by the single clones was observed over time. The fact that the average values of the Ag-F and h-R Luciferases per cell and of the Ag-F/(Ag-F + h-R) ratio expressed by the single clones are not significantly different from those found in cells from the corresponding bulk culture indicates that the clones are representative of the total population (Table 1).
The continuous treatment of GM979ALCRhprRlucgprFluc cells with the selecting agent G418 allows for selection of cells in which the transgene is inserted in stable regions of the DNA and prevents that the reporter is lost by deletion [14]. Since the reporter and the NeoR vector are physically separated, treatment with G418 does not select for high level of transcription from the reporter constructs [8]. Mechanism endogenous to GM979 cells, possibly similar to those that affected the gene expressed by the hetero-specific hybrids, might underlay the increased expression of the h-driven activity in these cells over time. There is no correlation between the relative amount of A g-F and h-R Luciferase activities expressed by single GM979ALCRhprRluc AgprFluc clones isolated from early bulk cultures (Fig. 3). Instead, clones derived from late bulk culture show a consistent and highly significant ( P b 0.004) positive correlation between the Ag-F and h-R Luciferase activity they express (Fig. 3). The heterogeneous Ag-F/(AgF + h-R) ratios observed in clones obtained at early time suggest that the transcription driven from the Ag- and hglobin promoter is regulated in an independent manner in GM979 ALCRhprRluc AgprFluc cells at early points of time. This variability also suggests that early on after transfection, the Ag-globin promoter is regulated in a fetal fashion and does not obey the endogenous regulatory mechanism(s) present in GM979 cells. A similar delay in enforcement of the epigenetic mechanisms has been found with the hybrids between fetal human erythroblasts and MEL cells. In these experiments, the synthesis of the human globins would switch to the adult phenotype in a few weeks of culture [15]. The progressive switching of the Ag-F/(Ag-F + h-R) to an adult-like range of values could reflect the impact of epigenetic mechanisms, which would normally silence the A g-F and increase the h-R promoter activity. The minimal correspondence to the normal structure of the globin locus in the artificial construct arrangement could explain why there is only a partial reduction of the Ag-F promoter activity over time. The response to the HbF inducer butyric acid was stronger in the early stage of the bulk culture, where it was mediated mainly by a selective increase of the Ag-F production (Table 2). However, in primary cells, the reactivation of a silenced human Ag-globin gene by a chemical inducer is contemporary with a reduction of transcription from the h-globin gene [16]. This balance is correctly reproduced in late bulk cultures, when after induction with butyric acid the lower A
180
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
increase in Ag-F is accompanied by a decrease of the h-R transcription. It should be remembered that butyric acid is not only an inhibitor of histone-deacetylases but also a posttranslational regulator of human g-globin mRNA [17], and thus is only partially active in an artificial model of HbF reactivation based on reporter genes. In effect, GM979 cells from late bulk culture strongly (by 40-fold) increased Ag-F activity in response to propionic acid, another known inducer of HbF expression [10]. The ALCR construct has been shown to be strictly required for the continuous activity of both adult and fetal globin gene promoters, either in vivo and in vitro, and to confer a partial position-independent and copy number-dependent expression to the transgene [18]. The change in regulation of the activity of the Ag- and h-globin promoters in GM979 ALCRhprRluc AgprFluc cells in early and late bulk culture indicates the presence of other mechanisms, not linked to the insertion site and to the LCR, which might regulate the activity of these promoters and their response to histonedeacetylase inhibitors, like butyric and propionic acid. In conclusion, the GM979 ALCRhprRluc AgprFluc cell line switches soon after transfection from a fetal to an adultlike Ag-F/(Ag-F + h-R) ratio. Such a switch is linked to the acquisition of a differential response to histone-deacetylases inhibitors, like butyric acid. These data confirm GM979 ALCRhprRluc AgprFluc cells from late bulk culture as a model for the identification of new HbF inducers. Moreover, the rapid change observed in the first weeks after transfection could represent a new model to assay LCR independent mechanism(s) of regulation of the Ag- and hglobin promoters.
Acknowledgments The authors gratefully acknowledge Prof. Francesco Antonio Manzoli for continuous support. This study was supported by, Progetti di ricerca di Interesse Nazionale 2001 and 2002 from the Ministry of Health, MIUR 60% Grant 2002, Grant E.1172 from Telethon Foundation, Progetti FIRB 2002 and 2003, National Project on Stem Cells, Institutional Funds from Istituto Superiore di Sanita`, and grant NIH-HL20899.
References [1] M. Bradai, M.T. Abad, S. Pissard, F. Lamraoui, L. Skopinski, M. de Montalembert, Hydroxyurea can eliminate transfusion requirements in children with severe beta-thalassemia, Blood 102 (2003) 1529 – 1530.
[2] M. Maier-Redelsperger, D. Labie, J. Elion, Long-term hydroxyurea treatment in young sickle cell patients, Curr. Opin. Hematol. 6 (1999) 115 – 120. [3] R.E. Ware, B. Eggleston, R. Redding-Lallinger, W.C. Wang, K. Smith-Whitley, C. Daeschner, B. Gee, L.A. Styles, R.W. Helms, T.R. Kinney, K. Ohene-Frempong, Predictors of fetal hemoglobin response in children with sickle cell anemia receiving hydroxyurea therapy, Blood 99 (2002) 10 – 14. [4] W.C. Wang, R.W. Helms, H.S. Lynn, R. Redding-Lallinger, B.E. Gee, K. Ohene-Frempong, K. Smith-Whitley, M.A. Waclawiw, E.P. Vichinsky, L.A. Styles, R.E. Ware, T.R. Kinney, Effect of hydroxyurea on growth in children with sickle cell anemia: results of the HUG-KIDS Study, J. Pediatr. 140 (2002) 225 – 229. [5] G.F. Atweh, A.N. Schechter, Pharmacologic induction of fetal hemoglobin: raising the therapeutic bar in sickle cell disease, Curr. Opin. Hematol. 8 (2001) 123 – 130. [6] J.A. Ho, C.V. Pickens, M.P. Gamcsik, O.M. Colvin, R.E. Ware, M.P. Gamscik, In vitro induction of fetal hemoglobin in human erythroid progenitor cells, Exp. Hematol. 31 (2003) 586 – 591. [7] H. Cao, G. Stamatoyannopoulos, M. Jung, Induction of human {gamma} globin gene expression by histone deacetylase inhibitors, Blood 103 (2004) 701 – 709. [8] E. Skarpidi, G. Vassilopoulos, Q.L. Li, G. Stamatoyannopoulos, Novel in vitro assay for the detection of pharmacologic inducers of fetal hemoglobin, Blood 96 (2000) 321 – 326. [9] Y. Ikawa, M. Aida, Y. Inoue, Isolation and characterization of high and low differentiation-inducible Friend leukemia lines, Gann 67 (1976) 767 – 770. [10] E. Liakopoulou, Q. Li, G. Stamatoyannopoulos, Induction of fetal hemoglobin by propionic and butyric acid derivatives: correlations between chemical structure and potency of Hb F induction, Blood Cells Mol. Diseases 29 (2002) 48 – 56. [11] E. Skarpidi, H. Cao, B. Heltweg, B.F. White, R.L. Marhenke, M. Jung, G. Stamatoyannopoulos, Hydroxamide derivatives of shortchain fatty acids are potent inducers of human fetal globin gene expression, Exp. Hematol. 31 (2003) 197 – 203. [12] G. Zitnik, P. Hines, G. Stamatoyannopoulos, T. Papayannopoulou, Murine erythroleukemia cell line GM979 contains factors that can activate silent chromosomal human gamma-globin genes, Proc. Natl. Acad. Sci. U. S. A. 88 (1991) 2530 – 2534. [13] G. Zitnik, K. Peterson, G. Stamatoyannopoulos, T. Papayannopoulou, Effects of butyrate and glucocorticoids on gamma- to beta-globin gene switching in somatic cell hybrids, Mol. Cell. Biol. 15 (1995) 790 – 795. [14] A.R. Migliaccio, C. Bengra, J. Ling, W. Pi, C. Li, S. Zeng, M. Keskintepe, B. Whitney, M. Sanchez, G. Migliaccio, D. Tuan, Stable and unstable transgene integration sites in the human genome: extinction of the Green Fluorescent Protein transgene in K562 cells, Gene 256 (2000) 197 – 214. [15] T. Papayannopoulou, M. Brice, G. Stamatoyannopoulos, Analysis of human hemoglobin switching in MEL x human fetal erythroid cell hybrids, Cell 46 (1986) 469 – 476. [16] R.D. Smith, J. Li, C.T. Noguchi, A.N. Schechter, Quantitative PCR analysis of HbF inducers in primary human adult erythroid cells, Blood 95 (2000) 863 – 869. [17] R.S. Weinberg, X. Ji, M. Sutton, S. Perrine, Y. Galperin, Q. Li, S.A. Liebhaber, G. Stamatoyannopoulos, G.F. Atweh, Butyrate increases the efficiency of translation of {gamma}-Globin mRNA, Blood. [18] Q. Li, K.R. Peterson, X. Fang, G. Stamatoyannopoulos, Locus control regions, Blood 100 (2002) 3077 – 3086.
Spontaneous switch from Ag- to h-globin promoter activity in a stable transfected dual reporter vector Giovanni Migliaccioa,*, Angela Di Baldassarreb, Cristina Di Ricob, Antonella Di Noiaa, Betty Nakamotoc, Hua Caoc, Eva Skarpidic, Anna Rita Migliacciod a
Gene and Cell Therapy, Department of Cell Biology and Neuroscience, Istituto Superiore di Sanita`, Viale Regina Elena 299, I-00161 Rome, Italy b Department of Biomorphology, University bG D’AnnunzioQ, Chieti, Italy c Department of Medicine, University of Washington, Seattle, WA 98195, USA d Department of Hematology, Oncology and Molecular Medicine, Istituto Superiore Sanita`, Rome Submitted 28 July 2004; revised 21 November 2004 Available online 20 January 2005 (Communicated by Dr. G. Stamatoyannopoulus, 13 December 2004)
Abstract Here it is analyzed the expression of a mini locus dual reporter construct composed by a micro-LCR and by the promoters for Ag- and hglobin gene, each one linked to a different Luciferase, in stably transfected GM979 cells for as long as 1–2 years from transfection. The transfected GM979 cells rapidly (within 1 month) evolved into a stable population which expresses constant levels of reporters for more than a year of continuous bulk culture. No silencing of the inserted construct was observed over time. In contrast, after 1 month, the reporter activity (both from Ag- and h-promoter) expressed per cell increased over time. The analysis of the Luciferase contained in single cell clones indicated that the higher reporter activity was due to increased gene expression per cell rather than to clonal selection of the most expressing clones. Since the activity driven by the h-promoter increased 10-fold more than that driven by the Ag one, the ratio between Ag-driven/(Agdriven + h-driven) reporter activity in the cells decreased after 1 month and became similar to the g/(g + h) globin mRNA ratio expressed by adult erythroid cells. Moreover, although both cells from early and late bulk culture responded to incubation with butyric acid, a known inducer of fetal globin gene expression, by increasing the reporter activity driven by the Ag-promoter, only cells from late bulk culture decreased, as normal primary erythroblasts do, the activity of the reporter driven by the h-promoter. These results suggest that the rapid changes in activity driven by the Ag- and h-globin promoters occurring during the first month after transfection may represent a novel in vitro model to study epigenetic regulation of the Ag- and h-promoter during the fetal to adult hemoglobin switch and confirm GM979 cells stably transfected with the dual reporter construct as a reliable assay for automated screening of chemical inducers of fetal globin gene activation. D 2004 Elsevier Inc. All rights reserved. Keywords: Hb switch; GM979 cell line; Epigenetic regulation; Dual Luciferase reporter assay
Introduction Thalassemia is a severe genetic disease caused by coinheritance of defective a- or h-globin alleles resulting in deficient globin expression. In the case of h-thalassemia, the disease is due to deletion or abnormal regulation of the
* Corresponding author. Fax: +39 06 4990 2530. E-mail address: [email protected] (G. Migliaccio). 1079-9796/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bcmd.2004.11.005
h-globin gene. The resulting inefficient h-globin expression leads to accumulation of unpaired, insoluble a-globin chains in red cells. The consequent accelerated destruction of these cells induces ineffective erythropoiesis and severe anemia. The development of transfusion therapies coupled with administration of iron-chelating agents has greatly ameliorated the treatment and the quality of life of hthalassemia patients. The overall treatment of this disorder, however, is still unsatisfactory. In fact, while the incidence of the disease is steadily decreasing in western countries, it
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
remains high in economically challenged countries that do not have easy access to the technologies required to properly manage these complex therapies. The observation that some of the homozygote carriers of h-globin gene deletions are asymptomatic because of persistent expression in adult life of the fetal-specific h-like genes, the gglobin genes (High Persistence of Fetal Hemoglobin Syndrome, HPFH), has inspired numerous studies aimed to identify chemical substances capable to pharmacologically re-activate g-globin expression in patients with hthalassemia. One of the pharmacological agents identified so far is hydroxyurea. Phase I–II clinical trials with hydroxyurea on h-thalassemia [1] and sickle cell [2,3] patients have been successfully completed, and the drug is presently considered to be introduced in the current medical practice for these disorders. Hydroxyurea, however, is a compound not devoid of counter indications [4]. Therefore, the search for additional and less toxic agents is still in progress [5]. The ideal candidate for pharmaceutical treatment of hthalassemia is a cheap chemical that could be administered orally, would be effective within reasonable doses, and would present no or little counter indications. The efficiency of the search for these agents is strictly dependent on the availability of a surrogate assay to measure the potentiality to reactivate g-globin gene expression in vivo. Such surrogate assays should be rapid, cheap, reliable, and suitable to be adapted for automated screening. Assays of induction of g-globin gene expression that use as target cells human erythroblasts and that determine the cellular globin content by HPLC represent the tests that would most reliably mimic the possible effects a chemical might induce in vivo [6]. These assays, however, as precise and indicative of clinical efficacy as they might be, are cumbersome and not easily adaptable for automatic screening. More suitable for this purpose would be surrogate assays based on murine erythroleukemia cell lines transfected with reporter constructs containing regions of the human g-globin promoter linked to appropriate reporter genes. The accuracy of these assays depends on how closely the promoter regions that drive the expression of the reporter gene mimic the physiology of g-globin gene activation in primary erythroid cells. Recently, it has been developed an assay based on the murine erythroid GM979 cell line stably transfected with a construct containing a human mini globin locus composed by a micro-LCR and by the promoters for Ag- and h-globin, each one linked to a different Luciferase [7,8]. The ratio between the expression of the two Luciferases reflects the relative activity of the two promoters and can be measured easily in to be automated assays. The suitability of this assay, however, for automated screening of chemical inducers of globin gene expression depends on how stable and consistent is the expression of the reporter in transfected cells over time. Here it is characterized the expression over time of the mini locus dual reporter construct in stably transfected
175
GM979 cells, from the first month up to 1–2 years from transfection. The results obtained indicate that GM979 ALCRhprRluc AgprFluc cells undergo an initial period of stabilization after which they express high levels of Luciferase activity driven both by the Ag- and hpromoter. The Ag-driven/(Ag-driven + h-driven) reporter activity ratio, that was similar to the fetal-globin ratio at early bulk cultures, became similar to adult globin gene ratio at later time points. Butyric acid, a known fetal hemoglobin inducer, increased expression of the reporter from the Ag-promoter both in early and in late bulk cultures but decreased that from the h-promoter only in late bulk cultures. As such, GM979 ALCRhprRluc AgprFluc cells from late bulk cultures represent a reliable assay for automated screening of chemical inducers of globin gene activation.
Materials and methods Establishment of the GM979 lLCRbprRluc AcprFluc cell line The murine erythroleukemic GM979 cell line had been established from fetal liver cells [9] and was chosen for this study because its phenotype includes expression both of minor and major Hb. The GM979 ALCRhprRluc AgprFluc cell line was obtained by lipofecting GM979 cells with a ALCRhprRluc AgprFluc plasmid that contains a 3.1-kb ALCR cassette including the DNAse I hypersensitive core of the 5V hypersensitive sites HS1, HS2, HS3, and HS4, linked to 315-bp of the human h-globin promoter and 1.4-kb of the A g-globin promoter driving the Renilla (R) and the Firefly (F) Luciferase gene, respectively [8]. A separate plasmid containing the NeoR gene was also included in the cotransfection procedure. A stably transfected cell line was selected on the basis of neomicine resistance, but no formal clonal selection was attempted. Passage of GM979 lLCRbprRluc AcprFluc cells in bulk culture and isolation of single cell-derived clones The GM979 ALCRhprRluc AgprFluc cell line was maintained in RPMI 1640 (Gibco, USA) supplemented with 10% of heat inactivated fetal calf serum (FCS, Hyclone, Logan UT) and 400 Ag/mL of G418 (Gibco) and incubated at 378C in a humidified incubator containing 5% CO2 in air. To avoid random clonal selection, cells were routinely sub-cultured every 4–5 days by seeding at least 5 105 cells/mL in each passage. No changes were noticed in the growth characteristics of the GM979 ALCR hprRluc AgprFluc cells over more than 1 year of continuous growth in bulk culture. Single cell-derived clones of GM979 ALCRhprRluc AgprprFluc cells were obtained by standard limiting dilution procedures. Briefly, GM979 ALCRhprRluc AgprFluc cells were diluted in complete medium at a concentration of 10
176
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
cells per 300 AL, and 100 AL of this suspension was plated in each well of 96-multiwells plates (a total of 6 plates). The plates were then incubated at 378C in a humidified incubator containing 5% CO2 in air. After 1 week, the wells were analyzed under an inverted microscope to identify those in which cell proliferation had occurred. After two additional weeks, the cells contained in each positive well were harvested, counted, and passed into individual flasks where they were cultured until they reached a number sufficient for performing the Dual Luciferase assay (usually 4 weeks) (see later). Dual Luciferase assay The Ag-driven Firefly (Ag-F) and h-driven Renilla (h-R) Luciferase activities were measured with the Dual Luciferase Reporter Assay System (Promega, Madison, WI, USA), as described by the manufacturer. Briefly, triplicate samples, containing 2 105 to 1 106 cells each, were centrifuged and washed twice in phosphate-buffered saline. The cell pellet was then lysed with 100 AL of Lysing Buffer (Dual Luciferase Kit, Promega) and the cellular debris removed from the supernatants by centrifugation. Aliquots of 20 AL of the supernatants were added to the Renilla and Firefly Luciferase substrate solution and the amount of light emitted by the sample measured with a Lumat LB9507 Luminometer (EG&G BERTHOLD, USA). The numeric values for h-R were doubled to compensate for the intrinsic lower luminosity, and both Luciferase activities were expressed in Arbitrary Fluorescence Units (AFU). As the translation efficiencies, as well as the mRNA and protein half-life, of the two Luciferase genes are all similar, the ratio between the Ag-F and h-R AFU corresponds not only to the amount of Ag-F and h-R proteins produced by the cells, but also to the ratio between the transcription rates driven by the human A g and h-globin promoters. The results are presented both as amount of Luciferase activity per cells (either Ag-F or h-F AFU/total number of cells in the sample) and as ratio of reporter activities [Ag-F AFU/(Ag-F AFU + h-F AFU)].
Results Increased Luciferase activity per cell, both from Ac- and b-promoter, in GM979 lLCRbprRluc AcprFluc cells over time The amount of Ag-F and h-R Luciferase activity expressed over time by GM979 ALCRhprRluc AgprFluc cells is presented in Fig. 1. The activity of both the Ag-driven Firefly and the h-driven Renilla expressed by the cells progressively increased during the culture (Fig. 1). The amount of the h-driven Renilla Luciferase, however, increased much more than that of the Ag-driven Firefly Luciferase. In fact, the average activity of Ag-F Luciferase per cell increased by 2-fold (from 0.05 F 0.02 AFU at 2 months to 0.10 F 0.06 AFU at 11 months) while that of the h-R Luciferase increased by 8-fold (from 4.3 F 1.7 AFU at 2 months to 35.96 F 19.16 AFU at 9 months). After 14 months of continuous bulk culture, the average Luciferase activity per cell had further increased to 0.28 F 0.14 AFU for Firefly and 58.13 F 20.95 AFU for Renilla, with an average
Fetal Hb (HbF) induction Butyric (0 up to 1.5 mM) and proprionic (0 and 5 mM) acids were added to three separate pools of 2 106 GM979 ALCRhprRluc AgprFluc cells resuspended in fresh culture medium. After 4 days of incubation at 378C and 5% CO2, the number of alive cells was counted by Trypan Blue exclusion. Identical numbers of live cells for test point were dissolved in Lysing Buffer and their Ag-F and h-R activity measured with the Dual Luciferase Assay, as described above. Statistical analysis Statistical analysis of the results was obtained with the Origin 6.1 software (Microcal, Northampton, NJ, USA).
Fig. 1. (A and B) Luciferase activity driven by the human h-globin (A) and A g- (B) promoters in bulk cultures of GM979 ALCRhprRluc AgprFluc cells over time. Data are reported as AFU per alive cell. The activity per cell of both Luciferases significantly increased ( P b 0.001) over time in more than 1 year of bulk culture. Activity driven by the h-promoter, however (r = 0.7), increased 4-fold more than that driven by the Ag-promoter (r = 0.52).
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
increase with respect to the early time points after transfection of approximately 5- and 10-fold, respectively. In both cases, the changes in Luciferase activity expressed by the cells over time fitted a linear regression curve (Figs. 1A and B, P b 0.0001 for both regressions), indicating a direct proportionality between the content of Ag-F and h-R Luciferases per cell and the time the cell line had spent in bulk culture. The different kinetics of the changes in Ag-F and h-R Luciferase activity expressed by GM979 ALCRhprRluc AgprprFluc cells over time were reflected by changes in the Ag-F/ (Ag-F + h-R) ratio (Fig. 2). The most striking change in the A gF/AgF + hR ratio was observed during the first 2 months after transfection. The Ag-F/(Ag-F + h-R) ratio, that had ranged from 3% to 8% at early time points, became as low as 0.6% from 100 days of bulk culture thereon (Fig. 2). The kinetics of the changes observed in the Ag-F/(Ag-F + h-R) ratio over time were best fitted with a curve of exponential decay of first grade (Fig. 2). These results indicate that the changes in Ag-F/(Ag-F + hR) ratio expressed by GM979 ALCRhprRluc AgprFluc cells over time in bulk culture are not due to decreased transcription driven from the human Ag-promoter (and therefore to its selective silencing) but rather to differential activation of the expression driven by the Ag- and hpromoter during the first 50 days of culture. Luciferase activities in single cell-derived clones of the GM979 lLCRbprRluc AcprFluc cell line In order to clarify if the changes in the Ag-F and h-R activities produced by the cells over time were a consequence of clonal selection, we analyzed Ag-F and h-R activity expressed by single cell-derived clones obtained
177
Fig. 3. Relative Ag- vs. h-driven Luciferase activity expressed by single cell-derived clones obtained from bulk cultures at early (2 months, dots) and late (11 months, squares) time points of bulk culture. Each dot represents the activity expressed by an individual clone. A positive linear relationship ( P b 0.0001) was observed between the relative Firefly and Renilla Luciferase activity per cell expressed by clones isolated at 11 months, but not in those obtained at 2 months of bulk culture.
from the GM979 ALCRhprRluc AgprFluc cell line at early (within 1 month) and late (12 months) time points of bulk culture (Fig. 3 and Table 1). No difference was found in the cloning efficiency of GM979 ALCRhprRluc AgprFluc cells at the two time points. In both cases, only one-fifth of the wells contained proliferating clones after 3 weeks of culture. The Ag-F and h-R Luciferase activity expressed by single cell-derived clones obtained from early and late time points of bulk culture ranges over a large spectrum of values (Fig. 3). There is no overlap in the relative activity of Ag-F and h-R Luciferase expressed by the clones obtained at the two time points. At the early stage of bulk culture, the average activity per cell for both Firefly and Renilla Luciferases ranges in intensity over more than four log with no correlation between the relative amount of Ag-F and h-R expressed by each clones (Fig. 3, Early). In the clones obtained at 12 months, the range of activities of the two Luciferases is still spread over more than 4 logs. However, there is a significant ( P b 0.004) positive linear correlation Table 1 Mean (FSEM) Firefly and Renilla Luciferase activity, and the corresponding Ag-F/(Ag-F + h-R) ratio, expressed by single clones derived by limiting dilution from GM979 ALCRhprRluc AgprFluc cells at early (2 months) and late (11 months) times of bulk culture A
Fig. 2. Ag-F/(Ag-F + h-R) ratio in bulk cultures of GM979ALCRhprRluc AgprgprFluc cells over time. The Ag-F/(Ag-F + h-R) Luciferase activity is plotted against time from transfection. The exponential decay curve of first grade that significantly (Chi2 = 0.0002) fits the distribution of the experimental points is indicated as dotted line.
Early time Clones Bulk culture Late time Clones Bulk culture
g-Firefly
h-Renilla
A
g-F/(Ag-F + hR)
0.15 F 0.11 0.06 F 0.02
2.99 F 1.07 6.54 F 2.6
0.06 F 0.04 0.03 F 0.01
0.02 F 0.01 0.27 F 0.03
12.9 F 4.6 55.4 F 4.0
0.007 F 0.001 0.005 F 0.001
The levels of activity expressed by the corresponding cells in bulk cultures are presented as well.
178
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
between the relative activity expressed from the two promoters by each single clone (Fig. 3). Although the amount of Ag-F and h-R Luciferase expressed by cells from the bulk cultures is higher than the average amount of Luciferase expressed by single clones (Table 1), there is no significant difference in the Ag-F/(AgF + h-R) ratio measured in cells in bulk culture and those expressed by the corresponding single clones (Table 1). These results indicate that during the time in bulk culture the cells coordinately increase the expression of both reporter genes. Differential effects of butyric acid on Ac-and b-reporter activity in early and late GM979 lLCRbprRluc AcprFluc cells The modifications observed in the steady-state level of hand Ag-globin promoter activity in GM979 ALCRhprprRluc AgprFluc cells over time prompted us to investigate whether compounds known to induce g-globin gene expression would activate the Ag-globin promoter in these cells at any stage of the culture. Since previous studies had shown that short chain fatty acids specifically induce g-globin transcription, we treated GM979ALCRhprRluc AgprFluc cells obtained at early and late stage of the bulk culture for 4 days with increasing doses of butyric acid and compared the level of Ag-F and h-R activities expressed in the different conditions (Table 2). Incubation with butyric acid of cells at the early stage of the culture (b2 months) strongly increased A g-F activity. The maximal effects were exerted at the concentration of 1 mM. Incubation with butyric acid scarcely affected the h-R activity expressed by these cells (Table 2). As a result, butyric acid induced an evident, concentrationdependent increase of the Ag-F/(Ag-F+h-R) ratio in early GM979ALCRhprRluc AgprFluc cells (Table 2, Ag-F/(Ag-F + h-R)). On the other hand, GM979 ALCRhprRluc AgprFluc cells at 11 months of bulk culture had a clearly different response to butyrate. At this point in time, butyric acid modestly increased Ag-F activity, while the drug had a strong inhibitory effect on the h-R activity that, however, remained still much higher than that expressed by cells at early time points. As a
result, although butyric acid increased the Ag-F/(Ag-F + h-R) ratio also in GM979 ALCRhprRluc AgprFluc cells at 11 months of bulk culture, the effect was lower than that induced in cells which had been in culture only for 2 months. The reduced response to butyric acid of the Ag globin promoter in late (11 months) bulk culture is not due to silencing of the promoter. In fact, the amount of Firefly Luciferase produced in untreated late (11 months) cells is higher than that produced by early (2 months) bulk culture cells (Figs. 1 and 3). Furthermore, late bulk culture cells increased by 40-fold (0.09 vs. 3.00, P b 0.05) Firefly Luciferase activity upon treatment with propionic acid (5 mM), another compound known to induce HbF [10] (Table 3). Both propionic and butyric acid have an inhibitory effect on cell proliferation as demonstrated by the reduced number of cell doublings occurring during the 4 days of incubation with the inducer (Tables 2 and 3).
Discussion To provide for a fast and reliable assay for chemical inducers of human HbF, the murine GM979 cell line was transfected with a dual Luciferase construct (ALCRhprRluc AgprFluc) [8] that allows direct comparison of the reporter activity driven by the human Ag- and h-globin promoters [7,11]. Here we show that the resulting GM979 ALCRhprRluc AgprFluc cell line maintains the ability to produce both Firefly and Renilla Luciferases over a period of more than 1 year of continuous bulk culture (Figs. 1 and 2). Over the same time period, it maintains its ability to respond to butyric acid by increasing the Luciferase driven by the Ag-promoter (Table 2). Changes were, however, observed in the pattern of reporter activity expressed by GM979 ALCRhprRluc AgprFluc cells over time. The Ag-F/(Ag-F + h-R) ratio decreased shortly after transfection, switching from a fetal-neonatal-like ratio (8–2%), that had been observed in the first 2 weeks of culture, to an adult-like ratio (0.1–0.01%) after the first 2 months of bulk culture. Thereafter, it remained stable for more than 1 year in culture (Fig. 1 and data not shown).
Table 2 Effects of increasing concentrations of butyric acid on Ag-driven (Ag-Firefly) and h-driven (h-Renilla) Luciferase activity, as well as on the Ag-F/(Ag-F + h-R) ratio, in GM979 ALCRhprRluc AgprFluc cells at 2 (Early phase) and 11 (Late Phase) months of bulk culture Early phase
Late phase
Butyric acid
A
Controls 0.5 mM 1 mM 1.5 mM Controls 0.5 mM 1 mM 1.5 mM
0.06 0.68 1.61 1.66 0.17 0.30 0.33 0.52
g-Firefly F F F F F F F F
0.01 0.10* 0.13* 0.16* 0.04 0.09* 0.07* 0.13*
h-Renilla 3.71 F 11.24 F 5.12 F 4.91 F 58.19 F 42.95 F 21.20 F 16.53 F
0.74 1.94* 2.1* 2.5* 19.27 12.39* 5.0* 4.2*
A
g-F/(Ag-F + h-R)
0.02 0.07 0.24 0.25 0.002 0.003 0.009 0.017
F F F F F F F F
0.01 0.02 0.02* 0.02* 0.0001 0.001 0.001* 0.003*
Cell doublings 3.77 1.95 0.89 0.46 3.66 1.45 0.29 1.36
F F F F F F F F
0.04 0.22* 0.13* 0.14* 0.09 0.11* 0.12* 0.32*
The effects of butyric acid on cellular proliferation are expressed as number of cell doublings during the 4 days of incubation (a negative number indicate cell death). Data are presented as mean F SEM of at least 4 separate experiments performed in triplicate. * P b 0.05.
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
179
Table 3 Effects of propionic acid (5 mM) on Ag-driven (Ag-Firefly) and h-driven (h-Renilla) Luciferase activity, as well as on the Ag-F/(Ag-F + h-R) ratio, in GM979 ALCRhprRluc AgprFluc cells at 11 (Late Phase) months of bulk culture Controls Propionic acid
Ag-Firefly
h-Renilla
0.09 F 0.01 3.00 F 0.09*
21.5 F 5.40 79.8 F 49.1
A
g-F/(Ag-F + h-R)
0.003 F 0.001 0.027 F 0.006*
Cell number (106) 1.37 F 0.07 0.087 F 0.01
The cell number recovered after the 4 days of incubation is also presented, as a measure of toxicity. Data are presented as mean F SEM of at least 3 separate experiments. * P b 0.05.
Unlike most of the MEL cell lines, the murine GM979 cell line co-expresses both murine embryonic (qy) and adult (hmaj) globin genes and allows expression of human fetal globin genes in hetero-specific hybrids constructed with nonerythroid human cells [12]. These observations have suggested that these cells contain trans-activating factors that allow activation of silenced human g-globin genes in humanmouse hybrids [12] and that are able to keep the human gglobin genes in an inducible state. However, the relatively high expression of the human g-globin genes in the heterospecific hybrids was transient and disappeared after a few weeks. The g-globin expression in these clones, however, remained inducible by chemical agents, such as butyric acid, indicating that the endogenous regulatory elements, although favoring an adult globin expression phenotype, do not limit gene expression driven by the g-globin promoter [13]. In normal primary erythroid cells, as well as in murine cell lines carrying the complete human h-globin locus, the switch observed over time from a fetal to an adult Hb pattern implies an autonomous silencing of the g-promoter and competitive transcription from the h-globin one. In GM979 ALCRhprRluc AgprFluc cells, the decrease in the Ag-F/(Ag-F + h-R) ratio observed in late bulk culture is not mediated by silencing of the Ag-promoter. In fact, the amount of both Ag-F and h-R Luciferases per cell increases steadily during the 10 months of bulk culture (Figs. 1A and B). However, while the average amount of h-R per cell increases over time by 10-fold, that of the Ag-F increases only 3-fold. It is, therefore, a different rate in the increase of promoter activity over time, rather than gene silencing, that confers an adult-like globin ratio to the ratio between the Luciferase activities driven by the two promoters at the late bulk cultures. A possible explanation of this phenomenon could be progressive selection over time in culture of those clones that preferentially express an adult-type of globin gene regulation. However, the amount of both Ag-F and h-R Luciferases expressed per cell by single clones isolated from the bulk culture at early and late time point is highly heterogeneous and spans a range of four AFU logs (Fig. 3). No decrease in the overall variability of the Luciferase activity expressed by the single clones was observed over time. The fact that the average values of the Ag-F and h-R Luciferases per cell and of the Ag-F/(Ag-F + h-R) ratio expressed by the single clones are not significantly different from those found in cells from the corresponding bulk culture indicates that the clones are representative of the total population (Table 1).
The continuous treatment of GM979ALCRhprRlucgprFluc cells with the selecting agent G418 allows for selection of cells in which the transgene is inserted in stable regions of the DNA and prevents that the reporter is lost by deletion [14]. Since the reporter and the NeoR vector are physically separated, treatment with G418 does not select for high level of transcription from the reporter constructs [8]. Mechanism endogenous to GM979 cells, possibly similar to those that affected the gene expressed by the hetero-specific hybrids, might underlay the increased expression of the h-driven activity in these cells over time. There is no correlation between the relative amount of A g-F and h-R Luciferase activities expressed by single GM979ALCRhprRluc AgprFluc clones isolated from early bulk cultures (Fig. 3). Instead, clones derived from late bulk culture show a consistent and highly significant ( P b 0.004) positive correlation between the Ag-F and h-R Luciferase activity they express (Fig. 3). The heterogeneous Ag-F/(AgF + h-R) ratios observed in clones obtained at early time suggest that the transcription driven from the Ag- and hglobin promoter is regulated in an independent manner in GM979 ALCRhprRluc AgprFluc cells at early points of time. This variability also suggests that early on after transfection, the Ag-globin promoter is regulated in a fetal fashion and does not obey the endogenous regulatory mechanism(s) present in GM979 cells. A similar delay in enforcement of the epigenetic mechanisms has been found with the hybrids between fetal human erythroblasts and MEL cells. In these experiments, the synthesis of the human globins would switch to the adult phenotype in a few weeks of culture [15]. The progressive switching of the Ag-F/(Ag-F + h-R) to an adult-like range of values could reflect the impact of epigenetic mechanisms, which would normally silence the A g-F and increase the h-R promoter activity. The minimal correspondence to the normal structure of the globin locus in the artificial construct arrangement could explain why there is only a partial reduction of the Ag-F promoter activity over time. The response to the HbF inducer butyric acid was stronger in the early stage of the bulk culture, where it was mediated mainly by a selective increase of the Ag-F production (Table 2). However, in primary cells, the reactivation of a silenced human Ag-globin gene by a chemical inducer is contemporary with a reduction of transcription from the h-globin gene [16]. This balance is correctly reproduced in late bulk cultures, when after induction with butyric acid the lower A
180
G. Migliaccio et al. / Blood Cells, Molecules, and Diseases 34 (2005) 174–180
increase in Ag-F is accompanied by a decrease of the h-R transcription. It should be remembered that butyric acid is not only an inhibitor of histone-deacetylases but also a posttranslational regulator of human g-globin mRNA [17], and thus is only partially active in an artificial model of HbF reactivation based on reporter genes. In effect, GM979 cells from late bulk culture strongly (by 40-fold) increased Ag-F activity in response to propionic acid, another known inducer of HbF expression [10]. The ALCR construct has been shown to be strictly required for the continuous activity of both adult and fetal globin gene promoters, either in vivo and in vitro, and to confer a partial position-independent and copy number-dependent expression to the transgene [18]. The change in regulation of the activity of the Ag- and h-globin promoters in GM979 ALCRhprRluc AgprFluc cells in early and late bulk culture indicates the presence of other mechanisms, not linked to the insertion site and to the LCR, which might regulate the activity of these promoters and their response to histonedeacetylase inhibitors, like butyric and propionic acid. In conclusion, the GM979 ALCRhprRluc AgprFluc cell line switches soon after transfection from a fetal to an adultlike Ag-F/(Ag-F + h-R) ratio. Such a switch is linked to the acquisition of a differential response to histone-deacetylases inhibitors, like butyric acid. These data confirm GM979 ALCRhprRluc AgprFluc cells from late bulk culture as a model for the identification of new HbF inducers. Moreover, the rapid change observed in the first weeks after transfection could represent a new model to assay LCR independent mechanism(s) of regulation of the Ag- and hglobin promoters.
Acknowledgments The authors gratefully acknowledge Prof. Francesco Antonio Manzoli for continuous support. This study was supported by, Progetti di ricerca di Interesse Nazionale 2001 and 2002 from the Ministry of Health, MIUR 60% Grant 2002, Grant E.1172 from Telethon Foundation, Progetti FIRB 2002 and 2003, National Project on Stem Cells, Institutional Funds from Istituto Superiore di Sanita`, and grant NIH-HL20899.
References [1] M. Bradai, M.T. Abad, S. Pissard, F. Lamraoui, L. Skopinski, M. de Montalembert, Hydroxyurea can eliminate transfusion requirements in children with severe beta-thalassemia, Blood 102 (2003) 1529 – 1530.
[2] M. Maier-Redelsperger, D. Labie, J. Elion, Long-term hydroxyurea treatment in young sickle cell patients, Curr. Opin. Hematol. 6 (1999) 115 – 120. [3] R.E. Ware, B. Eggleston, R. Redding-Lallinger, W.C. Wang, K. Smith-Whitley, C. Daeschner, B. Gee, L.A. Styles, R.W. Helms, T.R. Kinney, K. Ohene-Frempong, Predictors of fetal hemoglobin response in children with sickle cell anemia receiving hydroxyurea therapy, Blood 99 (2002) 10 – 14. [4] W.C. Wang, R.W. Helms, H.S. Lynn, R. Redding-Lallinger, B.E. Gee, K. Ohene-Frempong, K. Smith-Whitley, M.A. Waclawiw, E.P. Vichinsky, L.A. Styles, R.E. Ware, T.R. Kinney, Effect of hydroxyurea on growth in children with sickle cell anemia: results of the HUG-KIDS Study, J. Pediatr. 140 (2002) 225 – 229. [5] G.F. Atweh, A.N. Schechter, Pharmacologic induction of fetal hemoglobin: raising the therapeutic bar in sickle cell disease, Curr. Opin. Hematol. 8 (2001) 123 – 130. [6] J.A. Ho, C.V. Pickens, M.P. Gamcsik, O.M. Colvin, R.E. Ware, M.P. Gamscik, In vitro induction of fetal hemoglobin in human erythroid progenitor cells, Exp. Hematol. 31 (2003) 586 – 591. [7] H. Cao, G. Stamatoyannopoulos, M. Jung, Induction of human {gamma} globin gene expression by histone deacetylase inhibitors, Blood 103 (2004) 701 – 709. [8] E. Skarpidi, G. Vassilopoulos, Q.L. Li, G. Stamatoyannopoulos, Novel in vitro assay for the detection of pharmacologic inducers of fetal hemoglobin, Blood 96 (2000) 321 – 326. [9] Y. Ikawa, M. Aida, Y. Inoue, Isolation and characterization of high and low differentiation-inducible Friend leukemia lines, Gann 67 (1976) 767 – 770. [10] E. Liakopoulou, Q. Li, G. Stamatoyannopoulos, Induction of fetal hemoglobin by propionic and butyric acid derivatives: correlations between chemical structure and potency of Hb F induction, Blood Cells Mol. Diseases 29 (2002) 48 – 56. [11] E. Skarpidi, H. Cao, B. Heltweg, B.F. White, R.L. Marhenke, M. Jung, G. Stamatoyannopoulos, Hydroxamide derivatives of shortchain fatty acids are potent inducers of human fetal globin gene expression, Exp. Hematol. 31 (2003) 197 – 203. [12] G. Zitnik, P. Hines, G. Stamatoyannopoulos, T. Papayannopoulou, Murine erythroleukemia cell line GM979 contains factors that can activate silent chromosomal human gamma-globin genes, Proc. Natl. Acad. Sci. U. S. A. 88 (1991) 2530 – 2534. [13] G. Zitnik, K. Peterson, G. Stamatoyannopoulos, T. Papayannopoulou, Effects of butyrate and glucocorticoids on gamma- to beta-globin gene switching in somatic cell hybrids, Mol. Cell. Biol. 15 (1995) 790 – 795. [14] A.R. Migliaccio, C. Bengra, J. Ling, W. Pi, C. Li, S. Zeng, M. Keskintepe, B. Whitney, M. Sanchez, G. Migliaccio, D. Tuan, Stable and unstable transgene integration sites in the human genome: extinction of the Green Fluorescent Protein transgene in K562 cells, Gene 256 (2000) 197 – 214. [15] T. Papayannopoulou, M. Brice, G. Stamatoyannopoulos, Analysis of human hemoglobin switching in MEL x human fetal erythroid cell hybrids, Cell 46 (1986) 469 – 476. [16] R.D. Smith, J. Li, C.T. Noguchi, A.N. Schechter, Quantitative PCR analysis of HbF inducers in primary human adult erythroid cells, Blood 95 (2000) 863 – 869. [17] R.S. Weinberg, X. Ji, M. Sutton, S. Perrine, Y. Galperin, Q. Li, S.A. Liebhaber, G. Stamatoyannopoulos, G.F. Atweh, Butyrate increases the efficiency of translation of {gamma}-Globin mRNA, Blood. [18] Q. Li, K.R. Peterson, X. Fang, G. Stamatoyannopoulos, Locus control regions, Blood 100 (2002) 3077 – 3086.