Suppression of vascular endothelial growth factor expression at the transcriptional and post-transcriptional levels

Abstract

Gene expression is regulated at the transcriptional and post-transcriptional levels. Therefore, in order to achieve a high level of silencing, which includes minimizing any residual expression of a target gene, suppression at both the transcriptional and post-transcriptional levels is required. In this study, we describe a new method for highly efficient gene silencing that combines zinc finger protein-mediated transcriptional repression and small interfering RNA (siRNA)-mediated inhibition of post-transcriptional events. To measure the amount of gene expression under various conditions, we used a luciferase reporter gene that was driven by a variety of promoters, including that of the human vascular endothelial growth factor-A (VEGF-A) gene. We also measured expression of the endogenous VEGF-A gene. Inhibition of gene expression by each of the two individual technologies was effective, but in-depth analyses revealed residual expression of the target gene. The combination of specific zinc finger transcription factors and siRNAs greatly enhanced the silencing of the human VEGF-A gene, not only when cells were grown in the presence of normal amounts of oxygen but also under conditions of hypoxic stimulation. These results suggest that a bi-level approach to the silencing of VEGF-A expression may be clinically beneficial as part of a cancer treatment protocol.

Figure 1

Constructs used in the experiments described herein. (A) Expression vector constructs for ZFP-KRAB artificial transcriptional repressors. Zinc finger domains were designed and assembled on a p3 vector that included the CMV promoter, HA-epitope and NLS. The resulting zinc finger expression vectors pF268, pF121 and pF435 consist of three zinc finger domains, and pF109 consists of four zinc finger domains. DNA binding domains F268, F121, F109 and F435 were subcloned into pLFD-KRAB, yielding pF268-KRAB, pF121-KRAB, pF109-KRAB and pF435-KRAB. The amino acid sequence of the KRAB transcriptional repression domain is described in Materials and Methods. FD, functional domain. (B) Firefly luciferase reporter constructs with the TATA-minimal promoter (pGL-TA) and five copies of the Gal4-binding sequence (5 × Gal4BS) (pGL-TA-ZFPBS). A single copy of the ZFP binding sequences was introduced at the +18 position relative to the transcriptional initiation site (Inr). The binding sequences for each ZFP are described in Table 1. pGL-SV40-F268BS or pGL-CMV-F268BS are firefly luciferase reporter constructs that contain an SV40 promoter or a CMV promoter, respectively. At the 5′ end of the promoter, three copies of a zinc finger binding sequence for F268 were introduced. The binding sequence for F268 is described in Table 1. (C) pGL-VEGF is a luciferase reporter construct that contains the native human VEGF-A promoter sequences (from −950 to +450 relative to transcriptional initiation site). Binding sequences for F121 and F109 are present at positions +434 and −536, respectively.

Figure 2

Regulation of luciferase with a combination of siRNAs and F268-KRAB. HEK293 cells were transfected with 50 ng (+) of siRNA, 15 ng (+) of pF268-KRAB, neither effector (−) or both effectors together with (A) the pGL-TA-F268BS reporter (15 ng) plus transcriptional activator Gal4-VP16 (68.5 ng), (B) the pGL-SV40-F268BS reporter (15 ng) or (C) the pGL-CMV-F268BS reporter (15 ng). The Renilla luciferase expression plasmid (1.5 ng pRL-SV40) was added in equal amounts to the nucleic acid mixtures. The total amounts of DNA and RNA were brought to 150 ng by the addition of empty vector. Luciferase activities were measured 48 h post-transfection. Fold repression values were obtained after normalization of firefly luciferase activity to Renilla luciferase activity then compared with values obtained from control cultures transfected with empty vectors and reporters [pF268-KRAB (−), siRNA (−)]. Results are the mean values and standard errors of three independent experiments. (D) Simultaneous representation of the dose effects of the ZFP-TF and siRNA in the repression of luciferase activity measured in 293 cell cultures co-transfected with varying amounts of plasmids that encode a ZFP-TF (0–15 ng) or siRNA (0–50 ng) together with the reporter pGL-CMV-F268BS. Fold repression values were obtained after normalization of firefly luciferase activity to Renilla luciferase activity then compared with values obtained from control cultures [0 ng pF268-KRAB + 0 ng siRNA]. The total amount of nucleic acid was calibrated to a final amount of 100 ng with the empty vector pLFD-KRAB. Results are the mean values of three independent experiments.

Figure 3

Synergistic inhibitory effect of ZFP-TFs and siRNAs on luciferase production driven by the native human VEGF promoter. Two hundred and ninety-three cells were transfected with a luciferase reporter construct that contained the native human VEGF-A promoter together with either (A) 30 ng of pF121-KRAB (+), 50 ng of siRNA (+) or both, or (B) 30 ng pF109-KRAB (+), 50 ng siRNA (+) or both. Fold repression values were obtained by comparing repressed amounts of luciferase activity with that measured in cultures that had been transfected with empty vector (−/−) and reporter (pGL-VEGF) only [0 ng pF121-KRAB or pF109-KRAB, 0 ng siRNA]. Results are the mean values and standard errors of three independent experiments.

Figure 4

Synergistic repression of endogenous VEGF-A at the mRNA and protein level. 293F cells (10⁵ cells/24-well plate) were co-transfected with increasing amounts of a plasmid that encodes a ZFP-TF [pF435-KRAB; 0, 200, 400 ng] and an shRNA-encoding plasmid [shRNA+64; 0(−), 100 ng(+)]. Total transfected DNA was calibrated to a final amount of 500 ng with the addition of the pLFD-KRAB empty vector. VEGF-A mRNA was analyzed by RT–PCR using specific primers (see Materials and Methods). The amounts of VEGF mRNA were normalized to the amounts of GAPDH mRNA from the same reverse transcription product, then fold repression values were calculated by dividing normalized values from control cultures by that of effector-treated cultures (A). Results are the mean values and standard errors of three independent experiments. Also shown are simultaneous representations of the fold repression of secreted VEGF-A protein obtained from 293F cell cultures that had been co-transfected with various combinations of ZFP-encoding plasmids and either shRNA+64 (B) or shRNA+338 (C) and incubated for 72 h. Results are the mean values of three independent experiments.

Figure 5

Inhibition of hypoxia-induced VEGF activation by ZFPs and shRNAs. 293F cells were transfected with a plasmid encoding a ZFP-TF (pF435-KRAB), an shRNA expression vector (shVEGF+64) or both, and then cultivated for 48 h. During the last 7 h some cells were exposed to CoCl₂ (800 μM) to induce hypoxia (+). Secreted VEGF-A protein was analyzed by ELISA. Non-hypoxic (−) or hypoxic (+) cell cultures transfected with empty vectors (−shRNA+64, −F435-KRAB) served as controls. Results are the mean values and standard errors of two independent experiments.