|Home | About | Journals | Submit | Contact Us | Français|
Tailored approaches to restore defective transcription responsible for severe diseases have been poorly explored. We tested transcription activator-like effectors fused to an activation domain (TALE-TFs) in a coagulation factor VII (FVII) deficiency model. In this model, the deficiency is caused by the −94C>G or −61T>G mutation, which abrogate the binding of Sp1 or HNF-4 transcription factors. Reporter assays in hepatoma HepG2 cells naturally expressing FVII identified a single TALE-TF (TF4) that, by targeting the region between mutations, specifically trans-activated both the variant (>100-fold) and wild-type (20–40-fold) F7 promoters. Importantly, in the genomic context of transfected HepG2 and transduced primary hepatocytes, TF4 increased F7 mRNA and protein levels (2- to 3-fold) without detectable off-target effects, even for the homologous F10 gene. The ectopic F7 expression in renal HEK293 cells was modestly affected by TF4 or by TALE-TF combinations. These results provide experimental evidence for TALE-TFs as gene-specific tools useful to counteract disease-causing promoter mutations.
Little has been done to counteract the effect of promoter mutations, which represent a relevant cause of human genetic disease (http://www.hgmd.cf.ac.uk), via engineered transcription factors (eTFs).
One versatile tool is the transcription activator-like effectors (TALEs) that exhibit simple DNA-binding properties with a conserved central region composed of a series of 33–35 amino acid repetitive elements1,2, each containing a highly variable di-residue (RVD) at the 12th–13th positions that dictates the binding preference3,4. Therefore, the tailored mutagenesis of the RVDs and fusion to transcriptional activators (i.e., VP16, VP64, p300, p65, TET1)5 generates TALE-TFs that can target a selected promoter region and increase gene expression by stimulating transcription6 (Fig. 1a). Thus far, TALE-TFs have been mainly exploited to study stem-cell maintenance and differentiation7,8,9,10,11,12,13,14 or to increase transcription of the frataxin gene in the presence of a trinucleotide (GAA) repeat expansion15. Single nucleotide changes severely impairing the transcription of disease genes have not yet been addressed.
Factor VII (FVII) is the serine-protease that triggers blood coagulation, and it offers a unique model of finely regulated gene expression. Therefore, FVII can be taken as model system to study the positive or negative modulation of gene transcription. Here, we chose two mutations known to cause severe FVII deficiency16 as paradigmatic examples of mutations that impair transcriptional activity by affecting key transcription factor binding sites, which is a common pathogenic mechanism induced by promoter defects. Through reporter gene assays, we identified a single TALE-TF that was able to rescue the defective F7 promoter, and it was also active in the context of chromatin in cultured cells and primary hepatocytes. Thus, this work provides evidence of the efficacy of TALE-TFs on rescuing disease-causing promoter mutations.
The −94C>G and the −61T>G homozygous promoter mutations are known to cause severe FVII deficiency by impairing the binding of the Sp1 and HNF-4 transcription factors17,18, respectively (Fig. 1b). Using computational analyses19, we designed four TALE-TFs (TF1-4) targeting sequences 18–26bp in length within the 70bp region between positions −122 to −68 of the F7 proximal promoter (Fig. 1b), which are predicted to be unique in the human genome (Table S1). The TALE-derived DNA binding domains were expressed as a fusion protein with the VP64 effector domain (Fig. 1a). VP64 is a powerful and versatile trans-activator because its activity can be modulated by altering the number of VP domains5,8,9,13.
Screening conducted using reporter gene assays in FVII-expressing hepatoma cells (HepG2) identified a single TALE-TF (TF4), which was designed to target the F7 promoter sequence 5′TCCCTCTGTCACCCTTGGAGGC3′ between the Sp1 and HNF-4 binding sites, that showed trans-activation features. Specifically, TF4 remarkably increased the transcriptional activity of the severely impaired F7 promoter variants to an extent significantly higher than that of the wild-type construct (pFVII-luc−94g, 5.9±0.6-fold; pFVII-luc−61g, 25±4.9-fold). The pFVII-lucwt luciferase activity was also significantly enhanced (pFVII-lucwt, 36.6±17.3-fold) (Fig. 1c). Taking into account that tumour cells show different transcriptional mechanisms and profiles20, TF4 activity was also explored in HuH7 hepatoma cells. In this context, TF4 exhibited robust trans-activation activity on F7 promoter variants (pFVII-luc−94g, 23±6.1-fold; pFVII-luc−61g, 76±10.2-fold) and wild-type (17.1±1.5-fold) (Fig. 1d). Conversely, TF1, TF2 and TF3, which were designed to target upstream regions (Fig. 1b), were ineffective in rescuing F7 promoter variants and seemed to exert an inhibitory effect (Fig. 1c). These results may be due to an underlying competition and/or steric hindrance with the endogenous transcription factors on their binding sites.
As hepatocytes from the severely affected FVII-deficient patients with promoter mutations were unavailable, we tested TF4 on the endogenous F7 gene expression in HepG2 cells. RT-PCR analysis conducted on the total RNA extracted from pTF4-transfected cells revealed an increased amount of F7 mRNA compared to control cells, which corresponded to a 3.6±0.3-fold increase upon qPCR (Fig. 2a). Intrigued by this observation, we investigated the impact of TF4 on FVII protein expression. For untransfected HepG2 cells, FVII expression is directly related to the cell number (~1 and 2ng/ml from 0.5×106 and 1.0×106 cells, respectively) (Fig. 2b, blue bars). In medium with 0.5×106 cells transfected with pTF4, we observed approximatively a two-fold increase of FVII antigen levels (Fig. 2b, red bar), which was consistent with the extent of the increase in the F7 mRNA levels. Most importantly, the expressed FVII was enzymatically active, as measured by a very sensitive functional fluorogenic assay in plasma that evaluates the activation of factor X, the natural substrate of FVII21. In this assay, the activity of FVII responsible for triggering the coagulation cascade process inversely correlates with lag time22. As shown in Fig. 2c, in medium with 0.5×106 cells transfected with pTF4, we detected an increase of secreted protein levels with a concurrent shortening of lag times relative to untransfected cells (42 vs 50minutes). These results were comparable to those from 1.0×106 control cells (Fig. 2c).
This coherent increase in F7 mRNA, secreted protein and activity levels observed upon pTF4 treatment appears to be remarkable in the light of the very poor transfection efficiency (approximately 5%) of HepG2 cells, as estimated by FACS analysis (GFP tracer encoded by the pTF4 vector; data not shown).
Certainly, the efficacy of TALE-TFs strongly depends on the chromatin status and the accessibility of the target regions23, and this issue is one that cannot be properly assessed in hepatoma cells and through reporter gene assays exploiting episomal vectors. Thus, we transduced Hep10 human primary hepatocytes using adeno-associated viral vectors expressing the entire TF4 (AAV8-TF4) or a variant lacking the DNA binding domain (AAV8-ΔDBD) as a control (Fig. 2d).
The RNA analysis showed a 3.7±0.33-fold induction of F7 gene expression in AAV8-TF4 versus AAV8-ΔDBD transduced Hep10 cells (Fig. 2e). FXa generation assays in medium revealed a significant shortening of lag times between AAV8-TF4 and AAV8-ΔDBD transduced cells, indicating an increase in the levels of secreted FVII protein able to promote the generation of FXa in plasma systems (Fig. 2f).
Similar to HepG2 cells, the transactivation of the endogenous F7 transcription in Hep10 cells may be limited by competition with the HNF4 transcription factor. Additionally, transactivation could be further limited by the rate of both infection and transgene expression of AAV vectors in vitro24,25. However, from a translational therapeutic perspective, the use of AAV is very promising26,27,28.
To investigate differences in the target region accessibility caused by different chromatin statuses, the effect of TFs was also explored in human embryonic kidney 293 cells (HEK293). Because HEK293 cells are of kidney origin, they lack the liver-specific HNF4 transcription factor and do not significantly express FVII. This system enabled us to provide insights into both the TF4 mechanism of transcription activation and the steric hindrance hypothesis in HepG2 cells for the TFs 1–3. Even in this experimental setting, only TF4 was able to significantly increase the transcriptional levels of the endogenous F7 gene (Fig. 3). These findings point to a mechanism in which TF4 activity is not mediated by HNF-4, in accordance with the known ability of VP64 to recruit the basal transcriptional machinery1,29.
Moreover, these data support the hypothesis that TF4 activity depends on its position and the accessibility to the target region. In fact, in the absence of the HNF-4 transcription factor, the single TFs 1–3, which previously inhibited the activity of the FVII-lucwt construct potentially through competition with endogenous transcription factors, were unable to trans-activate F7 gene expression in HEK293 cells (Fig. 3). Noticeably, by transfecting pTF4 in combination with the other TALE-TFs, we detected an additionally increased expression of the endogenous F7 (>20-fold), which was modest compared to that in untreated HepG2 cells. Consistently, the combination of TF1-3 was ineffective (Fig. 3).
Taken together, the observed synergistic effects note that TF1-3 could bind to their target, but they may bind too far from the transcription start site to enhance gene expression.
Target specificity is a crucial issue when proposing a new therapeutic approach. Thus, we assessed our designed TALE-TFs in multiple control experiments for off-target effects. The key role of the TALE-DNA binding domain was to guarantee specific targeting. This specificity was demonstrated in control experiments where the VP64 activation domain was expressed alone. Negligible effects on the expression of the pHD plasmid were observed in these experiments (Fig. 1c). In complementary experiments, TF4 was unable to increase the transcription of neither the pBasic construct, which does not have a promoter, or the pSlug vector, which possesses a ubiquitous and F7-unrelated Slug promoter30 (Fig. 4a). Most importantly, the TF4 efficacy was abolished upon the partial mutagenesis (five out of the 22 nucleotides) of the target sequence (Fig. 4a, inset), either in the wild-type (pFVII-lucwtΔ5) or mutated (pFVII-luc−61Δ5) contexts, a finding that supports high specificity and a reduced probability of off-target effects (Fig. 4a).
The F7 model also offered us the opportunity to test, in the context of chromatin, the specificity of TF4 for activation of F7 over the neighbouring and homologous F10 gene, which is located approximately 2.1kb downstream of the F7 locus (Fig. 1a) and exhibits a 15-fold higher expression level than FVII. Our results with RT-PCR and qPCR analyses indicated that TF4 exerted no effect on F10 transcription in both HepG2 (Fig. 4b) and in HEK293 cell lines (Fig. 4c).
To investigate in detail the possibility of TF4 off-targeting, we tested the expression of the four genes showing the highest score of off-target prediction (supplementary Table S1). qPCR performed in pTF4-treated versus untreated HepG2 indicated no significant influence on these targets (Fig. 4d), which rules out unspecific effects on the most probable candidates.
Taken together, these results obtained with complementary approaches demonstrate the ability of TF4 to specifically enhance the F7 gene promoter function, which leads to a concurrent increase of secreted FVII protein and activity levels. Taking into account the very low therapeutic threshold for FVII deficiency31, this enhancement could rescue coagulation and reduce haemorrhagic diathesis in severe patients. It is worth noting that none of the promoter mutations identified so far in FVII deficient patients (http://www.hgmd.cf.ac.uk) fall within the region targeted by the active TF4. This observation implies that TF4 is applicable for all F7 promoter mutations; thus, cohorts of patients could benefit without any modification of the TALE DNA binding domain.
We believe that these novel data lay a robust foundation for the further manipulation of the TALE-TF construct to control the strength or mechanism of the activation domain (i.e., number of VP repeats13,32 or different domains5) as well as its expression levels (promoter strength) and tissue specificity (i.e., liver specific promoters33). Although we provided evidence for the successful use of TF4 delivered by an AAV vector with liver tropism only in a cellular model34, this result could lead to optimized TF4-derivatives that could be used to increase specificity and prevent possible FVII over-expression and consequent thrombotic risks.
In conclusion, to the best of our knowledge, this study provides the first experimental insights into the ability of a single engineered TALE-TF to rescue, with negligible off-target effects, gene transcription impaired by two promoter mutations. If delivered in vivo, these results could open new therapeutic strategies for coagulation factors as well as other genetic diseases sharing similar pathogenic mechanisms.
To create the pFVII-lucwt vector, the F7 proximal promoter region (520bp) was amplified with primers 5′AAAAAGCTTCTGTGGCTCACCTAAGAAACCAG3′ and 5′AAAAAGCTTGATGAA ATCTCTGCAGTGCTGC3′ and cloned upstream of the firefly luciferase sequence into the pGL3 Basic Vector (Promega, Madison, USA) using the Hind III restriction sites (underlined).
The pFVII-luc variants were created by mutagenesis (QuickChange II Site-Directed Mutagenesis Kit, Agilent Technologies, Santa Clara, USA) of pFVII-lucwt using forward primers 5′TCCTCCCCTCCGCCATCCCTCTG3 (pFVII-luc−94g), 5′GCAGAGAACGTTGCCCGTCAGTC3′ (pFVII-luc−61g) and 5′CCCCCATCCCTCTGTTGGAGGCAGAGAA3′ (pFVII-lucwtΔ5 and pFVII-luc−61Δ5). The reverse primers were complementary to the forward primers.
TALE-TFs (named pTF1-4) were designed using the TAL Effector Targeter webtool19 entering the 150bp of the minimal promoter region of the F7 gene, scanning for repeat arrays of 18–26 TALE and T as an upstream base and unchecking all the other options.
Four TALEs binding to the minimal promoter sequence and selected upon the minimal number of off-targets were created as described32.
The FVII studied in this work is of liver origin, so the reporter gene assays were conducted in hepatoma HepG2 and HuH7 cells. Cells were seeded into 12-well plates (0.2×106 cells/well) and transfected using Lipofectamine 3000 (Life Technologies, Carlsbad, USA) with 1μg of each pFVII-luc variant and/or 1μg of each pTF or pHD, expressing the VP64 alone, as a control. To normalize for the transfection efficiency, the cells were co-transfected with 100ng of pRluc (Promega) expressing Renilla luciferase. Forty-eight hours post-transfection, cells were lysed to evaluate the luciferase activity by the Dual-Luciferase assay (Promega).
To evaluate transfection efficiency, FACS analysis was conducted in HepG2 cells transfected with TALE-TFs vectors, which express GFP fused to the 3′ of the TALE-TF genes by a self-cleaving P2A linker.
TF4 or the ΔDBD (TF without the DNA binding domains) backbones were cloned in the AAV vector plasmid under the liver specific chimeric promoter of human alpha-1-antitrypsin and cer/hepatic locus control region25 and packaged by ViGene Biosciences, Inc., as AAV8 strain. The titers reuslted 3.61×1014GC/ml for the AAV8-TF4 and 3.15×61×1014GC/ml for the AAV8-ΔDBD.
Hep10 cells (human hepatocytes from ten donors) were obtained from Thermo Fisher Scientific (Waltham, USA) and were cultured according to the manufacturer’s instructions. Briefly, one vial of Hep10 cells was thawed, diluted in Cryopreserved Hepatocytes Recovery Medium (CHRM), centrifuged at 100 rcf for 10minutes at 4°C and resuspended in Williams’ Medium E (1x, no phenol red) supplemented with the Hepatocyte Maintenance Supplement Pack. Twenty-four hours later, cells were counted and seeded in 12-well plates at a density of 0.3×106 cells/ml in Williams’ Medium E and the Supplement pack as described above, with the addition of Vitamin K 5μg/ml and the AAV8-TF4 or pAAV8-ΔDBD at 1000 MOI. Forty-height hours later, the cells and medium were collected for the analysis of RNA and FVII activity levels.
HepG2, Huh7, HEK293 and Hep10 cells were transfected as described above. Forty-eight hours later, the cells were collected and the total RNA was isolated using TRIzol reagent (Life Technologies) to generate cDNA through the iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, USA).
F7 and F10 transcripts were detected by RT-PCR and quantitative PCR (qPCR), with primers Ex5F 5′GAGAACGGCGGCTGTGAG3′, Ex7R 5′GTTCCTCCAGTTCTTGATTTTGTCG3′ and primers Ex3F 5′AATGAATTCTGGAATAAATACA3′, and Ex5R 5′CTGTGGGAATGCAGGCCTTGC3′, respectively. RT-PCR was performed with AmpliTaq Gold 360 (Thermo Fisher Scientific) for 35 cycles, whereas qPCR employed the SsoFast EvaGreen Supermix (Bio-Rad) for 40 cycles. GAPDH and 18S were considered as housekeeping genes. The calculation of the fold changes was based on the 2−ΔΔCT method (Applied Biosystems User Bulletin #2).
The protein levels for FVII were evaluated by ELISA (Cedarlane, Burlington, Ontario, Canada) in medium collected 48hours post-transfection. Known concentrations of plasma-derived FVII (Haematologic Technologies Inc., Essex Junction, VT, USA) were used to create the reference curve.
Activated factor X (FXa) generation assays were performed as described21. Medium from transfected cells was added 1:1 (vol/vol) to commercial FVII-deficient plasma (BioMedical Inc., George King, USA), adjusted to a final dilution of 1:20 in a dilution buffer (20mM HEPES, 150mM NaCl, 0.1% PEG-8000, pH 7.4) and sub-sampled in a 96-well microplate (Costar, Corning, NY, USA). Coagulation was triggered by the addition of a mixture containing a specific fluorogenic substrate for FXa (150μM, American Diagnostica Inc., Greenwich, USA) and 1/100 Innovin (Siemens Healthcare, Marburg, Germany) prepared in a reaction buffer (20mM HEPES, 150mM NaCl, 5mM CaCl2, 0.1% PEG-8000, pH 7.4). The generation of FXa was measured as fluorescence emission (Relative Fluorescence Units, RFU; 360nm excitation, 465nm emission) over time on a microplate fluorimeter (Fluoroskan Ascent FL, Thermo Fisher Scientific, Helsinki, Finland).
FXa generation assays were analysed by the statistical software GraphPad Prism 5 (GraphPad Software, San Diego, USA). The specific parameter lag time (expressed in minutes) in FXa generation assays was extrapolated from the first derivative of relative fluorescence units (RFU) as a function of time (minutes).
TALE-TFs off-target analysis was performed by entering the target region for each TALE-TF in the Target Finder software19 with default options. The first ten off-target regions for each TF are reported with the prediction score in the supplementary Table S1. The off-target sites for the TF4 were analysed by qPCR as described above, with the following primers:
ABR_F, CTGCTGAGGAAACACCACAG; ABR_R, CCCCACAACTTGCTCCTAAG; LRRC3B_F, GCCCAAGCAAGGAAAGAAAT; LRRC3B_R, GGTAAAGGTTCCAGCCCTGT; ARID1B_F, GCGTGTGCAGGAGTTCAATA; ARID1B_R, GGAATTTCCATCTTGCTCTCA; and MYO5C_F, TCCTAAATAGCCGGGAGGAT; MYO5C_R, CTGCTGGGGATCTGAATCAT.
How to cite this article: Barbon, E. et al. An engineered tale-transcription factor rescues transcription of factor VII impaired by promoter mutations and enhances its endogenous expression in hepatocytes. Sci. Rep. 6, 28304; doi: 10.1038/srep28304 (2016).
This work was funded through the Early Career Bayer Hemophilia Awards Program (BHAP 2014; E.B., M.P. and M.B.) and Telethon (GGP14190; E.B., A.B. and M.P.). The authors would like to thank Prof. F. Zhang (Broad Institute, MIT, Cambridge, Massachusetts, USA) for kindly providing the backbone vectors and for the helpful suggestions in the preparation of TALE-TF plasmids. We also wish to thank Prof. R. Piva for kindly providing the pSlug vector. The AAV production was provided by Zairen Sun and the team at ViGene Biosciences (via Science Exchange). The flow cytometry acquisitions were carried out by the Flow Cytometry and Cell Sorting Facility of LTTA Center - University of Ferrara, Italy. Dr. F. Brugnoli and Dr. F. Casciano are also acknowledged for their technical support.
Author Contributions E.B. performed the TALE-TF assembly experiments with the reporter genes and the experiments in cultured cells and compiled the figures; S.P. created the pFVII-luc constructs and performed the luciferase assay experiments; A.B. performed the FXa generation activity assays; M.B. performed the experiments with hepatocytes, the off-target analysis, and drafted the paper; E.B., F.B., M.P. and M.B. conceived the study, designed the experiments, analysed and interpreted the data, and wrote the manuscript.