Epigenetic Regulation of Lentiviral Transgene Vectors in a Large Animal Model

ARTICLE

doi:10.1016/j.ymthe.2005.07.685

Epigenetic Regulation of Lentiviral Transgene Vectors in a Large Animal Model Andreas Hofmann,1 Barbara Kessler,2 Sonja Ewerling,2 Andrea Kabermann,1 Gottfried Brem,3 Eckhard Wolf,2 and Alexander Pfeifer1,* 1

Department of Pharmacy, Molecular Pharmacology, Center for Drug Research, Butenandtstrasse 5 (C), and 2Institute of Molecular Animal Breeding/Gene Center, Feodor-Lynen-Strasse 25, Ludwig-Maximilians University, 81377 Munich, Germany 3 Agrobiogen GmbH, 86567 Hilgertshausen, Germany *To whom correspondence and reprint requests should be addressed. Fax: +49 89 2180 77326. E-mail: [email protected].

Available online 2 September 2005

Transgenic animals are of outstanding relevance for genetic studies and the development of novel therapies for human diseases. A recent development is the generation of transgenic animals by lentiviral gene transfer. So far, studies on lentiviral transgenesis focused on first-generation (founder or F0) animals—most of which carry multiple integrants. Here, we analyze transgene expression and epigenetic regulation of individual integrants in lentiviral transgenic pigs after segregation to the F1 generation. Unexpectedly, one-third of lentiviral integrants exhibited low expression levels and were hypermethylated, as demonstrated by methylation-sensitive Southern blotting and bisulfite sequencing. Proviral methylation density correlated inversely with expression levels. In addition, treatment of isolated transgenic fibroblasts with the DNA methylase inhibitor 5azacytidine induced a threefold increase in mean fluorescence intensity (MFI) from 8 to 26.1. Treatment with the histone deacetylase inhibitor trichostatin A enhanced MFI to only 11.1. Taken together, expression of lentiviral integrants in higher mammals is regulated by epigenetic modifications. In contrast to previous expectations, DNA methylation plays an important role in lentiviral expression. Key Words: lentiviral vectors, DNA methylation, epigenetic regulation, lentiviral transgenesis, transgenic pig

INTRODUCTION Lentiviral vectors are a promising tool for gene therapy as well as molecular biology. Like all retroviruses, lentiviruses stably integrate their genome into host chromosomes and form a provirus, which is the basis for viral transgenesis [1–3]. Although infection of early embryos with vectors derived from prototypic retroviruses like murine leukemia virus (MuLV) results in the generation of transgenic founder animals (F0 generation) that transmit the retroviral genomes to progeny (F1 generation), the retroviral genes are expressed neither in the F0 nor in the F1 generation because of transcriptional silencing [4]. Analysis of silenced retroviral genomes revealed de novo methylation of CpG nucleotides within the provirus [4,5]. Methylated DNA binds a variety of methyl-binding proteins, including methyl-CpG binding protein 2, which interacts with histone deacetylases (HDACs), thereby linking cytosine methylation with chromatin remodeling [6].

MOLECULAR THERAPY Vol. 13, No. 1, January 2006 Copyright C The American Society of Gene Therapy 1525-0016/$30.00

An alternative approach to retroviral transgenesis is the use of vectors derived from lentiviruses [7–9]. The first lentiviral transgenic animals were generated by infection of murine preimplantation embryos [10,11] or embryonic stem (ES) cells [11] with vectors derived from HIV-1. Compared to the classical DNA microinjection (DNA-MI) technique, lentiviral gene transfer results in a four- to eightfold higher rate of transgenic animals per embryo treated [1], and more than 90% of F0-generation animals express the transgene [10,11]. Apart from mouse, lentiviral transgenesis proved to be highly efficient also in rat [10], chicken [12,13], and large farm animals (pigs and cattle) [14–16]. Transgenic pigs are of high medical relevance: given the close physiologic and anatomic similarities to humans, pig models of human diseases would be much more clinically relevant than most rodent models. In addition, transgenic pigs could be used for cell-based therapies and organ transplantation (xenotransplantation) if rejection could be overcome [2,3].

59

ARTICLE

Due to the presence of multiple integrants in most of the transgenic founders [10,14,16] potential silencing of a minority of the proviruses might not have been detected. The long terminal repeat (LTR) of HIV-1 has been shown to be methylated in mammalian cells, suggesting that methylation might play a role in silencing of vectors derived not only from prototypic retroviruses but also from lentiviruses [17]. However, newer generations of lentiviral vectors contain a transcriptionally inactive LTR with self-inactivating (SIN) deletions [18,19], and therefore, it was postulated that these lentivectors might escape epigenetic mechanisms [10,11]. Nevertheless, the specific genomic locus into which an individual lentivector integrates could affect its transcriptional activity and several studies indicated that a repressive chromatin environment might lead to silencing of lentiviral integrants [20–22]. A recent study by Yao et al. demonstrated that lentiviral SIN vectors can be silenced in murine ES cells and that integrants that exhibit low levels of transgene expression are marked by hypoacetylated histone H3 and bound H1 [23]. Although a similar silencing mechanism has been observed in mice after delivery of a SIN lentivector by classical DNA-MI [24], nonviral DNA integration differs from viral integration in fundamental aspects like dependence on viral enzymes, target site selection, and concatemerization [25,26]. To address the question whether silencing of lentivectors occurs in transgenic animals generated by lentiviral infection, we analyzed transgene expression and epigenetic regulation in transgenic pigs after segregation of lentiviral integrants to the F1 generation. One-third of lentiviral integrants exhibited low to undetectable transgene expression levels. Using methylation-sensitive Southern blotting and sodium bisulfite sequencing, we found methylation of the internal promoter and coding region in lentivector proviruses with low levels of expression.

RESULTS Variation of Lentiviral Transgene Expression Levels We used lentiviral transgenic animals to investigate epigenetic regulation of integrated lentiviral vectors in higher mammals in vivo. For this purpose, we generated transgenic pigs by infection of preimplantation embryos [14] with the HIV-1-based lentiviral vector LV-PGK, which carries the green fluorescent protein (eGFP) under the control of the human phosphoglycerate kinase 1 (PGK) promoter [27]. The vast majority (78%) of the F0generation animals born alive carried more than one integrant ([14] and data not shown). To analyze individual lentiviral integrants, we chose to quantify expression in the F1 generation after germ-line transmission and segregation of the proviruses. Three eGFP-expressing transgenic founders (Nos. 8507, 8514, and 8506) carrying one to three integrants (Figs. 1A, 1B, and 1C) were mated

60

doi:10.1016/j.ymthe.2005.07.685

with wild-type pigs. To identify individual integrants, we digested skin DNA with BamHI, which cuts only once in the lentiviral genome and thus generates junction fragments of different sizes depending on the site of integration into the porcine genome. Southern blot analysis using an eGFP probe demonstrated six individual lentivirus-specific bands (a–f in Fig. 1), which segregated independently and were found alone (12 animals, underlined in Fig. 1) or in combination (2 animals). Western blot analysis (Figs. 1D, 1E, and 1F) revealed significant variations between different proviruses. The majority of the animals exhibited intermediate (integrants c, d, and f) to high (integrant b) expression levels, whereas animals carrying integrants a and e expressed only low to undetectable levels of the transgene (Fig. 1G). Nevertheless, RT-PCR analysis clearly demonstrated transcription from these integrants (see Supplementary Fig. 1s). Summation of transgene expression levels of all individual F1 integrants yields approximately the expression level of the respective founder integrant(s) (Fig. 1G), indicating that transmission through the germ line does not alter lentiviral transgene expression. Analysis of Proviral Integrity Previous studies had shown significant deletions in integrated retroviruses present in the germ line of animals generated by infection of preimplantation embryos with MuLV [28]. Therefore, we first analyzed the integrity of the integrated lentiviral genomes by Southern blotting of skin DNA digested with the restriction enzyme SstI, which cuts in the 5V and 3V LTRs of the integrated lentiviral vector (Fig. 1H, top). All animals analyzed exhibited the 3.4-kb band of full-length LV-PGK (Fig. 1H, bottom). In addition, PCR analyses with five primer pairs covering the whole provirus yielded correct amplicons (see Supplementary Fig. 2s). Thus, major rearrangements in the lentiviral genome can be excluded as a reason for proviral silencing. Lentiviral Expression Depends on Methylation Density We next analyzed methylation of CpG dinucleotides in the lentiviral insertions by methylation-sensitive Southern blot analysis. Several CpG sites are present in the PGK promoter [29] and the methylation status of two of these sites can be determined by digestion of genomic DNA with the methylation-sensitive restriction enzyme AscI (Fig. 2A). Combination of EcoRI with AscI results in a 1.5kb band when the AscI site is unmethylated, whereas a 1.8-kb band is seen in the Southern blot when the site is methylated (Fig. 2B). The ratio of methylated to unmethylated signals ranged from 0.05 (integrant b) to 2.9 (integrant e) (Fig. 2C). We assayed methylation in the coding region of the lentiviral genome by digestion of skin DNA with the methylation-sensitive restriction enzyme HpaII and its methylation-insensitive iso-

MOLECULAR THERAPY Vol. 13, No. 1, January 2006 Copyright C The American Society of Gene Therapy

doi:10.1016/j.ymthe.2005.07.685

ARTICLE

FIG. 1. Transgene expression levels of different lentiviral integrants. (A–C) Identification of individual proviruses from three founder (F0) animals—8507 (male, two integrants), 8514 (female, three integrants), and 8506 (male, one integrant)—and their offspring (8774–8783, 8784–8793, and 9031–9036, respectively) by Southern blot analysis using BamHI, which cuts only once in the proviral genome. Six individual proviruses (a–f) were segregated by mating of the three transgenic founders. (D–F) Western blot analysis of transgene expression in extracts of skin biopsies (top). h-Actin antibodies were used for loading control (bottom). (G) Mean eGFP fluorescence of animals carrying integrants a–f (black columns) compared to their founders (white columns). Expression levels were calculated relative to the actin loading control. (H) Southern blot analysis using SstI, which cuts in the LTRs of the provirus (top). All animals exhibit the 3.4-kb full-length proviral insertion, demonstrating the intactness of the integrated provirus (bottom). Triangle, self inactivating mutation; PGK, phosphoglycerate kinase promoter; eGFP, enhanced green fluorescent protein; LTR, long terminal repeat; cPPT, polypurine tract; WPRE, woodchuck hepatitis responsive element. Animals carrying single integrants are underlined.

schizomer MspI, followed by Southern blotting and hybridization with a full-length eGFP probe (Figs. 2D and 2E). The schematic diagram (Fig. 2D) shows the possible fragments if digestion in the coding region with HpaII is not inhibited (0.5 kb) or is inhibited by DNA methylation (N0.5 kb). The methylation ratio—as calculated by the quotient (M H)/H (M, MspI band intensity;

MOLECULAR THERAPY Vol. 13, No. 1, January 2006 Copyright C The American Society of Gene Therapy

H, HpaII band intensity)—ranged from 0.01 (no significant DNA methylation, integrant b) to 0.92 (highest degree of methylation, integrant a) (Fig. 2F). We observed no signals for the other HpaII sites within eGFP gene, indicating that these sites are also methylated. We performed additional Southern blot analysis using the methylation-sensitive restriction enzyme EagI (see Sup-

61

ARTICLE

doi:10.1016/j.ymthe.2005.07.685

FIG. 2. Correlation between methylation status of lentiviral integrants and transgene expression levels. (A–C) Analysis of the methylation status of the promoter region using the methylation-sensitive restriction enzyme AscI. (A, B) Unmethylated DNA yields a 1.5-kb band after EcoRI/AscI digestion. Methylation of the CpG dinucleotide present in the AscI site results in a 1.8-kb band in the Southern blot. (C) The methylation status of integrants a–f was quantified by calculating the ratio of methylated to unmethylated bands. (D–F) Methylation of the coding region was analyzed by digesting genomic DNA with the methylation-sensitive HpaII (H) and its methylation-insensitive isoschizomer MspI (M). (F) Methylation status of integrants a–f as quantified by calculating the ratio of MspI to HpaII bands: (M H)/H (M, MspI band intensity; H, HpaII band intensity). (G) Analysis of eGFP expression by Western blotting of skin biopsies. Mean transgene expression levels in F1 animals, as calculated by comparing individual bands with the band intensity of a recombinant eGFP standard, relative to the actin loading control. (H) Correlation between mean eGFP expression and methylation levels in the promoter region (filled circles) and the coding region (open circles) of individual integrants (a–f).

plementary Fig. 3s) that confirmed the results of the AscI and HpaII digestion. Western blotting demonstrated that highly methylated integrants (a and e) express the lowest concentration of eGFP (Fig. 2G). Comparison of the level of gene expression with the degree of CpG methylation revealed that both the methylation of the PGK promoter and the coding region correlate inversely with expression of the transgene (Fig. 2H). Analysis of the Methylation Status by Bisulfite Sequencing Next, we used sodium bisulfite sequencing to examine the methylation status of the lentiviral proviruses at high

62

resolution. In the PGK promoter, we analyzed a 234-bp fragment with 24 CpGs (Fig. 3A). Two of these CpGs lie within the AscI site analyzed by Southern blotting (Figs. 2A and 3A). In animal 8778, which carries integrant b with high-level expression, we detected no methylated CpGs (mCpGs) (n = 6, Fig. 3A). In contrast, 87.5% (21 F 1.4 mCpGs; n = 8, FSEM) of the CpGs in the same region were methylated in animal 8779, which carries integrant a and exhibits low-level expression (Fig. 3A). The lowest expressing provirus (integrant e) demonstrated the highest degree of methylation, with 23 F 0.5 of 24 CpGs being methylated (95%; Fig. 3A). We also characterized the CpG methylation pattern of the eGFP cassette (Fig. 3B). A total

MOLECULAR THERAPY Vol. 13, No. 1, January 2006 Copyright C The American Society of Gene Therapy

ARTICLE

doi:10.1016/j.ymthe.2005.07.685

FIG. 3. High-resolution DNA methylation analysis of CpG dinucleotides of the proviral PGK promoter and the eGFP gene by bisulfite genomic sequencing. (A) Methylation pattern of a 234-bp region (24 CpG islands; top) in the PGK promoter of animal 8778 carrying integrant b (upper middle), animal 8779 carrying integrant a (lower middle), and animal 8791 carrying integrant e (bottom). The AscI recognition site, which was used for Southern blot analysis, is indicated. (B) Mapping of the methylation status of the eGFP cassette (255 bases, 18 CpG sites) of animals 8778 (upper middle), 8779 (lower middle), and 8791 (bottom). Closed circles indicate methylated CpGs, open circles indicate unmethylated CpGs.

of 18 CpGs (255-bp fragment) encompassing a part of the 0.5-kb HpaII/MspI fragment (Figs. 2D and 3B) were analyzed. Only 1.9% (0.3 F 0.2; n = 6) of these CpGs were methylated in animal 8778 (integrant b) (Fig. 3B). Again, genomic DNA of animal 8779 (integrant a) and animal 8791 (integrant e) exhibited a high degree of methylation in the coding region. Ninety-four percent (17 F 0.5; n = 8) and 97% (17.5 F 0.2; n = 8) of the CpGs analyzed were methylated in provirus a and e, respectively (Fig. 3B). Effects of 5-Azacytidine (5-AzaC) and Trichostatin A (TSA) To investigate the role of cytosine methylation in the regulation of lentiviral transgene expression on a cellular level, we isolated skin fibroblasts from animals 8779 and 8778. Treatment of fibroblasts from animal 8779 (highly methylated integrant a) with the potent DNA methylase inhibitor 5-azaC for 48 h (Fig. 4A) significantly increased mean fluorescence intensity (MFI) after 5 days in culture from 8 F 5.4 (DMSO-treated cells) to 26.1 F 10.5 (n = 7, FSEM) (Figs. 4B, 4C, and 4F). The number of eGFP+ cells increased after 5-acaC treatment from 8.5 F 3.8% to 34 F 6.2% (Figs. 4B, 4C, and 4F). Incubation with the histone deacetylase inhibitor TSA only slightly increased MFI and the percentage of eGFP+ cells to 11.1 F 6.5 and to 15 F 6.4% (n = 7, FSEM), respectively (Figs. 4A, 4D, and 4F). Consecutive 5-azaC/TSA treatment further enhanced the 5-azaC effect and increased MFI to 31.3 F 15 and eGFP+ cells to 27.5 F 6.6% (n = 7, FSEM) (Figs. 4A, 4E, and 4F). In contrast, the same experimental conditions did not significantly affect eGFP expression of fibroblasts from animal 8778, which contains the integrant b (data not shown). In parallel, we also analyzed the effects of 5-azaC and TSA on DNA methylation. We performed methylationsensitive Southern blot analysis using EcoRI and AscI to

MOLECULAR THERAPY Vol. 13, No. 1, January 2006 Copyright C The American Society of Gene Therapy

determine promoter methylation (Fig. 4G, see also Fig. 2A). We quantified methylation status again by calculating the ratio of methylated (1.8-kb) to unmethylated (1.5kb) bands (Fig. 4H). 5-AzaC-treated cells exhibited a 45% reduction in methylation density compared to DMSOtreated control cells (Fig. 4H). We observed no changes in the methylation status after TSA treatment (Fig. 4H). Taken together, inhibition of cellular methylases reactivated silenced transgene expression of highly methylated lentiviral integrants. TSA treatment had only a marginal influence on lentiviral transgene expression in cells isolated from a low-expresser transgenic animal.

DISCUSSION In this study, we analyzed epigenetic regulation of individual integrants in lentiviral transgenic animals in vivo. Unexpectedly, one-third of the proviruses analyzed exhibited only low or undetectable levels of expression and a high degree of methylation of CpG dinucleotides, reminiscent of gene silencing. Previous studies on lentiviral transgenesis had demonstrated unprecedented high frequencies of transgenic animals expressing the transgene and suggested the lack of silencing of lentiviral vectors [10,11,14–16]. However, these studies were focused on F0-generation animals, most of which carry multiple integrants, and variations in transgene expression of different proviruses might not have been detected. We found an inverse correlation between proviral methylation and expression levels. In addition, treatment of isolated transgenic fibroblasts with 5-azaC resulted in a reactivation of eGFP expression. Thus, cytosine methylation is an important regulator of the efficacy of transgene expression from lentiviral vectors. De novo methylation also plays a major role in gene silencing of prototypic retroviruses. An important difference between lentiviral and retroviral vectors appears to

63

ARTICLE

doi:10.1016/j.ymthe.2005.07.685

FIG. 4. Analysis of transgene expression in skin fibroblasts isolated from lentiviral transgenic pig 8779. (A) Cells were treated for 2 days with DMSO, 5-azaC (Aza), or TSA (TSA). Thereafter, the cells were incubated for 3 days in medium (F), in the presence of DMSO (DMSO), or in the presence of TSA (Aza/TSA). After 5 days, cells were analyzed by FACScan and methylationsensitive Southern blot analysis. (B–E) Representative FACScan analyses of transgenic fibroblasts. Cells were treated with (B) DMSO, (C) 5azaC, (D) TSA, or (E) 5-azaC and TSA. Overall mean and geometric mean fluorescence values are given. Numbers under the gate refer to % eGFP+ cells. (F) Statistical analysis of the effect of 5-azaC, TSA, or consecutive 5-azaC/TSA treatment on mean fluorescence. DMSO-treated cells (as control) were set as 100%. Geometric means F SEM (n = 7 experiments, *P b 0.05 versus control) are given. (G) Analysis of proviral methylation status using the methylation-sensitive restriction enzyme AscI. Unmethylated DNA yields a 1.5-kb band, methylation of the restriction site will result in a 1.8-kb signal. (H) The methylation level of skin fibroblasts was quantified by calculating the ratio of methylated to unmethylated bands.

be the degree of silencing: while only a minority of lentiviral integrants exhibit low expression levels in transgenic animals, expression of MuLV-derived proviruses is almost completely shut off during differentiation. Support for this notion comes from a previous study that compared directly efficacies and transgene expression of non-SIN vectors derived from a lentivirus (HIV-1) and a simple retrovirus (MuLV), which carried the same internal promoter and transgene, in murine ES cells [30]. The lentiviral transgene was efficiently expressed in ES cells and continued to be expressed during differentiation, whereas the MuLV-based vector showed a dramatic suppression of expression during or after differentiation [30]. However, high lentivector copy numbers (N1) might have masked silencing in this study [30], emphasizing the

64

need for in vivo analyses of single integrants. Different susceptibilities to gene silencing might be explained by differences in the basic biology and in the vector design: (i) Genomic integration is significantly different between MuLV- and HIV-derived vectors [31,32]. While 17% of MuLV integration sites are found within 1 kb of CpG islands, only 2% of HIV integrations land in the vicinity of CpG islands [32]. In addition, HIV favors integration into transcriptionally active regions of the host genome, which should facilitate high-level transcription of HIVderived transgene vectors [31,32]. (ii) In lentiviral vectors, SIN mutations that delete viral promoter and enhancer sequences have been successfully incorporated [18,19]. The SIN configuration not only minimizes the possibility of insertional activation of cellular oncogenes, but should

MOLECULAR THERAPY Vol. 13, No. 1, January 2006 Copyright C The American Society of Gene Therapy

doi:10.1016/j.ymthe.2005.07.685

also reduce the number of potential elements that are able to recruit the silencing machinery to the integrated vector. In contrast, the development of SIN vectors derived from prototypic retroviruses like MuLV has been hampered by the fact that extensive deletions within the MuLV LTRs result in low retroviral titers [33]. Previously, it was suggested that silencing of retroviral and lentiviral vectors does not require DNA methylation, but depends primarily on histone deacetylation [24]. Furthermore, it was shown that silenced lentiviral transgenes in ES cells are marked by repressive chromatin composed of hypoacetylated histones H3 and H1 [23]. Unfortunately, from this study no data on the methylation status of lentiviral vectors are available [23]. Our data indicate that DNA hypermethylation plays a major role in lentivirus silencing in transgenic animals. Although we cannot exclude a role for chromatin modifications in the establishment of silencing, TSA-dependent HDACs do not play a major role in maintenance of lentivirus silencing. Taken together, our data clearly show that lentiviral transgenesis is affected by varying degrees of epigenetic modification. Since the only difference between the vector proviruses analyzed is the site of integration into the host genome, these variations could be explained by heterogeneity of the genetic environment surrounding the proviruses. The methylation status of the adjacent genomic DNA as well as the chromatin environment may have permissive or repressive effects on lentivector proviruses. Our findings have implications for the design of studies based on lentiviral transgenesis: F0 animals, which mostly carry multiple integrants, are a fast and efficient way to generate transgenic animals that express high levels of the transgene even if a subset of integrants is silenced. However, F0 animals are derived directly from lentiviral infection, and the degree of mosaicism depends on the time point of infection. Transmission of the proviruses through the germ line to the F1 generation ensures the presence of the provirus in all cells of the organism, but it also leads to segregation of the integrants. Given the many biomedical applications of lentiviral vectors, understanding their epigenetic regulation is of high relevance not only for viral transgenesis, but also for basic science and stem cell-based gene therapy. Understanding the factors that regulate methylation of integrated lentiviruses not only would help to improve lentiviral expression in transgenic animals, but also could have a major impact on gene therapy approaches based on lentiviral vectors. Similarly, the factors that recruit the cellular silencing machinery to some but not all lentiviral proviruses could be a potential target for HIV therapy.

MATERIALS

AND

METHODS

Lentiviral vector production. The SIN lentiviral vector (LV-PGK) was recently described [27]. Recombinant lentivirus was produced using standard lentivirus production methods [11].

MOLECULAR THERAPY Vol. 13, No. 1, January 2006 Copyright C The American Society of Gene Therapy

ARTICLE

Transgenic pigs. Founder animals carrying one or multiple copies of the provirus were generated by infection of preimplantation embryos with lentiviral SIN vectors as previously described [14]. F1 animals with single lentiviral integrants were obtained by outcrossing of multitransgenic founder animals with wild-type pigs. Genomic DNA extraction and Southern blot hybridization. Small pieces of skin or pellets of skin fibroblasts were digested overnight in 500 Al extraction buffer (100 mM Tris, 5 mM EDTA, 0.2% SDS, 200 mM NaCl, and 0.1 mg/ml proteinase K; Roche, Mannheim, Germany) at 558C. Genomic DNA was extracted using standard phenol–chloroform alcohol extraction. Southern blot analysis was carried out as previously described [14] using a full-length eGFP probe. Western blot. Skin samples were lysed in protease-inhibiting solution (0.5% Triton X-100, 150 mM NaCl, 2 mM CaCl2, and protease inhibitors) and used directly for Western blot. After separation by 15% SDS–PAGE, proteins were transferred to polyvinylidene difluoride membrane (Immobilon-P transfer membrane; Millipore Corp., Bedford, MA, USA). The presence of eGFP was shown using monoclonal antibodies against eGFP (Clontech Laboratories, Palo Alto, CA, USA), secondary peroxidaseconjugated antibodies (Dianova, Hamburg, Germany), and an enhanced chemiluminescence detection kit (Amersham Pharmacia Biotech, Buckinghamshire, UK). Quantification of eGFP expression was carried out as previously described [14]. Cell culture. Porcine skin fibroblasts were obtained from skin biopsies of transgenic F1 piglets. Cells were isolated by lysis of the tissue with 0.05% trypsin (Gibco–Invitrogen, Karlsruhe, Germany) and seeded on six-well tissue culture dishes. Fibroblasts were cultured in DulbeccoTs modified EagleTs medium (Gibco) supplemented with 10% fetal calf serum (Biochrom AG, Berlin, Germany) and antibiotics (100 U/ml penicillin G and 100 Ag/ml streptomycin; Biochrom AG). Cells were immortalized by infection with viral vectors containing the large tumor antigen of simian virus 40 [34]. At 30–40% confluence cells were attached for 24 h and treated with 8 AM methylase inhibitor [35] 5-azacytidine (Sigma; solved in water), 100 nM deacetylase inhibitor [36] trichostatin A (Sigma; solved in 62% DMSO), or DMSO (control) for 2 days. Since 5-azaC is not stable during this incubation period [37], the medium was changed every 24 h. After 2 days of incubation with 5-azaC or TSA cells were cultured for 3 more days with and without TSA. Fibroblasts were fixed in 4% paraformaldehyde for 15 min at 48C and resuspended in phosphatebuffered saline, and fluorescence was measured by FACScan (Becton– Dickinson, Rutherford, NJ, USA). Bisulfite genomic sequencing. Genomic DNA was digested with EcoRI overnight to obtain small DNA fragments [38]. The DNA was then extracted with phenol, precipitated with ethanol, vacuum dried, and resolved in TE buffer (10 mM Tris, 0.1 mM EDTA, pH 8.0). DNA was frozen and thawed twice ( 808C and room temperature for 8 min each) and then heated at 1008C for 10 min. Two micrograms of DNA was denatured in a final volume of 20 Al containing 0.3 M freshly prepared NaOH and incubated at 378C for 15 min. Immediately before use, a bisulfite– hydrochinone solution (BSH), containing 3.6 M sodium bisulfite (Sigma), 0.6 mM hydrochinone (Sigma), adjusted to pH 5.0 with NaOH, was prepared. Cytosines were sulfonated by adding 120 Al of BSH to the denatured DNA and incubated in a thermocycler by the following cycle protocol: 30 s at 958C and 15 min at 508C for 15 cycles. Then, DNA was desalted by using the Wizard purification system (Promega, Mannheim, Germany) according to the manufacturerTs instructions and resolved in 103.6 Al of TE buffer. Then 6.4 Al of freshly prepared 5 M NaOH was added and incubated at 378C for 15 min. The solution was neutralized by addition of 47 Al of 10 M NH4OAc (pH 7.0), precipitated with 500 Al of 96% ethanol, vacuum dried, and resolved in 50 Al of TE buffer [39]. Aliquots were stored at 208C. DNA sequences were amplified by PCR. Three to 10 Al of bisulfite-converted DNA was mixed with 100 pM each PCR primer in 50 Al of reaction buffer containing 200 AM dNTPs and 2.5 U of Taq polymerase (Promega). The mixture was incubated in a thermocycler for 30 s at 958C, 1 min at 408C, and 1 min at 728C for a total of 42 cycles. For analysis of the PGK promoter ATGGGTTGTGGTTAATAG and

65

ARTICLE

CCCTTACTAACCATAATAA were used as forward primer and reverse primer, respectively. For the coding region GGGTATAAGTTGGAGTATAA (forward) and CTCCAACAAAACCATATAAT (reverse) were used. The PCR products were separated by gel electrophoresis, purified using a gel extraction kit (Amersham), and subcloned into the pCA2.1 vector (Topo TA cloning kit; Invitrogen). Six to eight clones were sequenced using an ABI Prism DNA sequencer (Perkin–Elmer).

ACKNOWLEDGMENTS We are grateful to Heidi Sebald for expert technical assistance and Theo Rein (Max-Planck-Institute for Psychiatry, Munich) for helpful advice on bisulfite sequencing. We thank Didier Trono (School of Life Sciences, Swiss Institute of Technology, Lausanne) and Luigi Naldini (San Raffaele Telethon Institute for Gene Therapy, San Raffaele Scientific Institute, Milan) for providing lentiviral plasmids. This study was supported by the Bayerische Forschungsstiftung (BFS 492/02) and the Deutsche Forschungsgemeinschaft. RECEIVED FOR PUBLICATION MAY 10, 2005; REVISED JULY 17, 2005; ACCEPTED JULY 18, 2005.

APPENDIX A. SUPPLEMENTARY DATA Supplementary data associated for this article can be found, in the online version, at 10.1016/j.ymthe.2005. 07.685. REFERENCES 1. Pfeifer, A. (2004). Lentiviral transgenesis. Transgenic Res. 13: 513 – 522. 2. Fassler, R. (2004). Lentiviral transgene vectors. EMBO Rep. 5: 28 – 29. 3. Whitelaw, C. B. (2004). Transgenic livestock made easy. Trends Biotechnol. 22: 157 – 159. 4. Jahner‘, D., et al. (1982). De novo methylation and expression of retroviral genomes during mouse embryogenesis. Nature 298: 623 – 628. 5. Laker, C., et al. (1998). Host cis-mediated extinction of a retrovirus permissive for expression in embryonal stem cells during differentiation. J. Virol. 72: 339 – 348. 6. Nan, X., et al. (1998). Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex. Nature 393: 386 – 389. 7. Wiznerowicz, M., and Trono, D. (2005). Harnessing HIV for therapy, basic research and biotechnology. Trends Biotechnol. 23: 42 – 47. 8. Ailles, L. E., and Naldini, L. (2002). HIV-1-derived lentiviral vectors. Curr. Top. Microbiol. Immunol. 261: 31 – 52. 9. Pfeifer, A., and Verma, I. M. (2001). Gene therapy: promises and problems. Annu. Rev. Genom. Hum. Genet. 2: 177 – 211. 10. Lois, C., Hong, E. J., Pease, S., Brown, E. J., and Baltimore, D. (2002). Germline transmission and tissue-specific expression of transgenes delivered by lentiviral vectors. Science 295: 868 – 872. 11. Pfeifer, A., Ikawa, M., Dayn, Y., and Verma, I. M. (2002). Transgenesis by lentiviral vectors: lack of gene silencing in mammalian embryonic stem cells and preimplantation embryos. Proc. Natl. Acad. Sci. USA 99: 2140 – 2145. 12. McGrew, M. J., et al. (2004). Efficient production of germline transgenic chickens using lentiviral vectors. EMBO Rep. 5: 728 – 733. 13. Chapman, S. C., et al. (2004). Ubiquitous GFP expression in transgenic chickens using a lentiviral vector. Development 132: 935 – 940. 14. Hofmann, A., et al. (2003). Efficient transgenesis in farm animals by lentiviral vectors. EMBO Rep. 4: 1054 – 1060. 15. Hofmann, A., et al. (2004). Generation of transgenic cattle by lentiviral gene transfer into oocytes. Biol. Reprod. 71: 405 – 409.

66

doi:10.1016/j.ymthe.2005.07.685

16. Whitelaw, C. B., et al. (2004). Efficient generation of transgenic pigs using equine infectious anaemia virus (EIAV) derived vector. FEBS Lett. 571: 233 – 236. 17. Bednarik, D. P., Cook, J. A., and Pitha, P. M. (1990). Inactivation of the HIV LTR by DNA CpG methylation: evidence for a role in latency. EMBO J. 9: 1157 – 1164. 18. Miyoshi, H., Blomer, U., Takahashi, M., Gage, F. H., and Verma, I. M. (1998). Development of a self-inactivating lentivirus vector. J. Virol. 72: 8150 – 8157. 19. Zufferey, R., et al. (1998). Self-inactivating lentivirus vector for safe and efficient in vivo gene delivery. J. Virol. 72: 9873 – 9880. 20. Hino, S., Fan, J., Taguwa, S., Akasaka, K., and Matsuoka, M. (2004). Sea urchin insulator protects lentiviral vector from silencing by maintaining active chromatin structure. Gene Ther. 11: 819 – 828. 21. Mohamedali, A., et al. (2004). Self-inactivating lentiviral vectors resist proviral methylation but do not confer position-independent expression in hematopoietic stem cells. Mol. Ther. 10: 249 – 259. 22. Ramezani, A., Hawley, T. S., and Hawley, R. G. (2003). Performance- and safetyenhanced lentiviral vectors containing the human interferon-beta scaffold attachment region and the chicken beta-globin insulator. Blood 101: 4717 – 4724. 23. Yao, S., et al. (2004). Retrovirus silencing, variegation, extinction, and memory are controlled by a dynamic interplay of multiple epigenetic modifications. Mol. Ther. 10: 27 – 36. 24. Pannell, D., et al. (2000). Retrovirus vector silencing is de novo methylase independent and marked by a repressive histone code. EMBO J. 19: 5884 – 5894. 25. Goff, S. P. (2001). Retroviridae: the retroviruses and their replication. In Fields Virology (P. M. Howley, D. M. Knipe, D. Griffin, R. A. Lamb, A. Martin, B. Roizman, S. E. Straus Eds.), 4th ed. Vol. 2, pp. 1871 – 1939. Lippincott–Raven, Philadelphia. 26. Robertson, G., Garrick, D., Wilson, M., Martin, D. I., and Whitelaw, E. (1996). Agedependent silencing of globin transgenes in the mouse. Nucleic Acids Res. 24: 1465 – 1471. 27. Follenzi, A., Ailles, L. E., Bakovic, S., Geuna, M., and Naldini, L. (2000). Gene transfer by lentiviral vectors is limited by nuclear translocation and rescued by HIV-1 pol sequences. Nat. Genet. 25: 217 – 222. 28. Jahner, D., and Jaenisch, R. (1980). Integration of Moloney leukaemia virus into the germ line of mice: correlation between site of integration and virus activation. Nature 287: 456 – 458. 29. Singer-Sam, J., et al. (1983). Isolation of a cDNA clone for human X-linked 3phosphoglycerate kinase by use of a mixture of synthetic oligodeoxyribonucleotides as a detection probe. Proc. Natl. Acad. Sci. USA 80: 802 – 806. 30. Hamaguchi, I., et al. (2000). Lentivirus vector gene expression during ES cell-derived hematopoietic development in vitro. J. Virol. 74: 10778 – 10784. 31. Schroder, A. R., et al. (2002). HIV-1 integration in the human genome favors active genes and local hotspots. Cell 110: 521 – 529. 32. Wu, X., Li, Y., Crise, B., and Burgess, S. M. (2003). Transcription start regions in the human genome are favored targets for MLV integration. Science 300: 1749 – 1751. 33. Yee, J. K., et al. (1987). Gene expression from transcriptionally disabled retroviral vectors. Proc. Natl. Acad. Sci. USA 84: 5197 – 5201. 34. Salmon, P., et al. (2000). Reversible immortalization of human primary cells by lentivector-mediated transfer of specific genes. Mol. Ther. 2: 404 – 414. 35. Groudine, M., Eisenman, R., and Weintraub, H. (1981). Chromatin structure of endogenous retroviral genes and activation by an inhibitor of DNA methylation. Nature 292: 311 – 317. 36. Hoshikawa, Y., Kwon, H. J., Yoshida, M., Horinouchi, S., and Beppu, T. (1994). Trichostatin A induces morphological changes and gelsolin expression by inhibiting histone deacetylase in human carcinoma cell lines. Exp. Cell Res. 214: 189 – 197. 37. Notari, R. E., and DeYoung, J. L. (1975). Kinetics and mechanisms of degradation of the antileukemic agent 5-azacytidine in aqueous solutions. J. Pharm. Sci. 64: 1148 – 1157. 38. Rein, T., Zorbas, H., and DePamphilis, M. L. (1997). Active mammalian replication origins are associated with a high-density cluster of mCpG dinucleotides. Mol. Cell Biol. 17: 416 – 426. 39. Clark, S. J., Harrison, J., Paul, C. L., and Frommer, M. (1994). High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 22: 2990 – 2997.

MOLECULAR THERAPY Vol. 13, No. 1, January 2006 Copyright C The American Society of Gene Therapy