Doxycycline modulates VEGF-A expression: Failure of doxycycline-inducible lentivirus shRNA vector to knockdown VEGF-A expression in transgenic mice

Abstract

Vascular endothelial growth factor-A (VEGF-A) is the master regulator of angiogenesis, vascular permeability and growth. However, its role in mature blood vessels is still not well understood. To better understand the role of VEGF-A in the adult vasculature, we generated a VEGF-A knockdown mouse model carrying a doxycycline (dox)-regulatable short hairpin RNA (shRNA) transgene, which silences VEGF-A. The aim was to find the critical level of VEGF-A reduction for vascular well-being in vivo. In vitro, the dox-inducible lentiviral shRNA vector decreased VEGF-A expression efficiently and dose-dependently in mouse endothelial cells and cardiomyocytes. In the generated transgenic mice plasma VEGF-A levels decreased shortly after the dox treatment but returned back to normal after two weeks. VEGF-A expression decreased shortly after the dox treatment only in some tissues. Surprisingly, increasing the dox exposure time and dose led to elevated VEGF-A expression in some tissues of both wildtype and knockdown mice, suggesting that dox itself has an effect on VEGF-A expression. When the effect of dox on VEGF-A levels was further tested in naïve/non-transduced cells, the dox administration led to a decreased VEGF-A expression in endothelial cells but to an increased expression in cardiomyocytes. In conclusion, the VEGF-A knockdown was achieved in a dox-regulatable fashion with a VEGF-A shRNA vector in vitro, but not in the knockdown mouse model in vivo. Dox itself was found to regulate VEGF-A expression explaining the unexpected results in mice. The effect of dox on VEGF-A levels might at least partly explain its previously reported beneficial effects on myocardial and brain ischemia. Also, this effect on VEGF-A should be taken into account in all studies using dox-regulated vectors.

Fig 1. Schematic drawing of the dox-regulatable VEGF-A shRNA lentivirus vector.

Lentivirus vector for inducible knockdown permits constitutive expression of a Tet-transactivating component (rtTA3) with a Venus selection marker (green fluorescent protein -like protein) and tetracycline-regulated expression of VEGF-A-shRNAs. The shRNA transcripts were designed as primary microRNA mimics i.e. they were embedded in the primary transcript of human miRNA30. The lentiviruses were self-inactivating (SIN) third generation vectors, in which part of the viral 3´LTR has been deleted preventing the viral replication. The vectors contain a central polypurine tract (indicated as FLAP) for enhancement of viral titers and a woodchuck hepatitis virus posttranscriptional regulatory element (WRE) for better transgene expression [24].

Fig 2. VEGF-A knockdown via RNAi in mouse endothelial cells and cardiomyocytes in a doxycycline-regulatable fashion.

Secreted VEGF-A protein amount was most efficiently decreased with sh1 -knockdown vector (T-1 and TT-1) in mouse endothelial cells in comparison to sh2 (T-2 and TT-2) and sh3 (T-3 and TT-3) vectors. With the sh1 construct the normal TRE promoter (T-1) seemed to be slightly more efficient and less leaky than the tight TRE promoter (TT-1) (a). The magnitude of VEGF-A knockdown with the selected T-1 vector increased with increasing dox doses in endothelial cells (b). With the increasing dox doses also the amount of Venus expressing cells (%) was increased as follows: 40 ± 7% (–dox), 46 ± 8% (dox 10 ng/ml), 85 ± 3% (dox 100 ng/ml) and 78 ± 2% (dox 1000 ng/ml) (c). In cardiomyocytes the knockdown of VEGF-A with the T-1 vector was shown to be dependent of the amount of virus (MOI) and dox exposure time and the decrease was larger at the cellular level in comparison to the secreted protein (d-e). Results are shown as mean ± S.D., n = 3/group. The percentage of VEGF-A protein concentration compared to control +dox group (a) or non-transduced (NT) group (b, d, e) are shown above bars. *p<0.05, **p<0.01 and ***p<0.001 compared to control +dox group (a) or non-transduced (NT) group (b, d, e), 1-way ANOVA with Dunnett´s post hoc test. NT = non-transduced cells, MOI = multiplicity of infection, Contr = Control lentivirus vector targeting Luciferase. Scale bar 200 μm (c).

Fig 3

Plasma VEGF-A concentration in response to dox treatment (a). Tissue VEGF-A mRNA and protein levels after two weeks of 1 mg/ml dox treatment and five weeks of 2 mg/ml dox dose treatment (b-k). Dox treatment decreased plasma VEGF-A levels in TG mice after 1 week, after which the plasma VEGF-A increased. In WT mice a decreasing effect on plasma VEGF-A was seen (a). *p<0.05, **p<0.01 and ***p<0.001 compared to baseline within each group, 1-way ANOVA with Dunnett´s post hoc test, n = 9-10/group. In the selected tissues, aorta, heart and kidney, the dox treatment with the 1 mg/ml dox dose for 2 weeks showed decreasing trend in VEGF-A mRNA expression in TG mice in comparison to WT mice (b, d, h), which was associated with increased cardiac VEGF-A protein (e) and decreased kidney VEGF-A protein levels (i). When the dox dose was doubled and the treatment time increased to 5 weeks, a trend towards increasing VEGF-A expression was seen in all three tissues in both WT and TG mice (c, f, j). However, no changes were detected in protein levels (g, k). **p<0.01 and ***p<0.001 compared to WT+dox group in 2 weeks experiment (b, d, e, h, i) or to no dox group (-dox) in 5 weeks experiment (c, f, g, j, k), t-test, n = 6-24/group.

Fig 4

Tissue Venus mRNA expression after 4 and 10 weeks of dox treatment in VEGF-A knockdown TG mice (a-d). Dox treatment increased the Venus expression in the heart (a), kidney (b), lungs (c) and skeletal muscle (d) in TG mice after 10 weeks of dox treatment in comparison to TG mice without dox treatment. The Venus mRNA expression was generally higher after 10 weeks dox treatment when compared to 4 weeks dox treatment in TG mice (a-d). No Venus expression was detected in the WT mice with 4 weeks dox treatment (data not shown). n = 5-10/group.

Fig 5

VEGF mRNA and protein levels in naïve/non-transduced mouse endothelial cells (a-b) and cardiomyocytes (c-d) in vitro after dox treatment of 10–1000 ng/ml for 24–72 h. VEGF-A mRNA levels were significantly decreased in endothelial cells 48–72 h after dox treatment with all used dox doses and already at 24 h with the highest dose of 1000 ng/ml (a). The cellular VEGF-A protein levels remained unchanged (b). In cardiomyocytes the VEGF-A mRNA levels increased (c) and cellular VEGF-A protein amount decreased after dox treatment of 48–72 h (d). Results are shown as mean ± S.D., one-way ANOVA with Dunnett’s post hoc test, *p<0.05, ***p<0.001 compared with the group of no dox treatment (0 ng/ml), n = 3/group.