Carvedilol inhibits EGF-mediated JB6 P+ colony formation through a mechanism independent of adrenoceptors

Abstract

Carvedilol is reported to prevent cancers in humans and animal models. However, a molecular mechanism has yet to be established, and the extent to which other β-blockers are chemopreventive remains relatively unknown. A comparative pharmacological approach was utilized with the expectation that a mechanism of action could be devised. JB6 Cl 41-5a (JB6 P+) murine epidermal cells were used to elucidate the chemopreventative properties of β-blockers, as JB6 P+ cells recapitulate in vivo tumor promotion and chemoprevention. The initial hypothesis was that β-blockers that are GRK/β-arrestin biased agonists, like carvedilol, are chemopreventive. Sixteen β-blockers of different classes, isoproterenol, and HEAT HCl were individually co-administered with epidermal growth factor (EGF) to JB6 P+ cells to examine the chemopreventative properties of each ligand. Cytotoxicity was examined to ensure that the anti-transformation effects of each ligand were not due to cellular growth inhibition. Many of the examined β-blockers suppressed EGF-induced JB6 P+ cell transformation in a non-cytotoxic and concentration-dependent manner. However, the IC50 values are high for the most potent inhibitors (243, 326, and 431 nM for carvedilol, labetalol, and alprenolol, respectively) and there is no correlation between pharmacological properties and inhibition of transformation. Therefore, the role of α1- and β2-adrenergic receptors (AR) was examined by standard competition assays and shRNA targeting β2-ARs, the only β-AR expressed in JB6 P+ cells. The results reveal that pharmacological inhibition of α1- and β2-ARs and genetic knockdown of β2-ARs did not abrogate carvedilol-mediated inhibition of EGF-induced JB6 P+ cell transformation. Furthermore, topical administration of carvedilol protected mice from UV-induced skin damage, while genetic ablation of β2-ARs increased carvedilol-mediated effects. Therefore, the prevailing hypothesis that the chemopreventive property of carvedilol is mediated through β-ARs is not supported by this data.

Fig 1. Carvedilol prevented EGF-mediated transformation of JB6 P+ cells independently of its cytotoxic effects.

Examination of carvedilol in three assays: 10 μM EGF-mediated JB6 P+ cell colony formation in soft agar (black) n = 20 to 22, 72-hour SRB cell viability assay (blue) n = 6, and 2-week MTS cell viability assay (pink) n = 6. There is no statistical difference between SRB and MTS assays. Data represented as mean ± SD after normalization to control (EGF minus DMSO control for soft agar, and DMSO for SRB and MTS).

Fig 2. Select β-AR ligands prevent EGF-mediated neoplastic transformation of JB6 P+ cells.

Exposure of JB6 P+ cells to 10 ng/mL EGF and the indicated concentration of β-AR ligands for 2-weeks in soft agar (black); n = 5 to 25. For cell viability experiments, JB6 P+ cells were exposed to the indicated concentration of β-AR ligands for 72 hrs for an SRB assay (blue); n = 6. Data represented as mean ± SD after normalization to control (EGF minus DMSO control for soft agar, and DMSO for SRB).

Fig 3. β-blocker-mediated inhibition of EGF-induced neoplastic transformation of JB6 P+ cells does not correlate with β2-AR affinity or lipophilicity.

The x-axis is the reported logK_d values (A and B), and XLogP3 obtained from PubChem (C). The y-axis is (A) logIC₅₀ and (B and C) logIΔ_2*PSD values for transformation. Pearson’s correlation shows that there is no correlation between log K_d values and either measurement, but there is a correlation between XLogP3 and logIΔ_2*PSD. The dotted gray line represents the 95% confidence interval of the linear regression (gray line). For clarity, labels only appear in panel A. In panel B green and black symbols represent effective and non-effective β-blockers as shown in Table 1, respectively, and an orange circle represents β-blockers that are reported to have biased characteristics at the β2-AR. Panel C is similarly color-coded with open circles representing ineffective β-blockers, green circles represent effective β-blockers, and the orange halo represents β-blockers that are reported to have biased characteristics at the β2-AR.

Fig 4. Expression of α1-ARs and role of α1-ARs in EGF-induced neoplastic transformation of JB6 P+ cells.

(A) Universal mouse (positive control) and JB6 P+ cDNA was amplified with primers specific for each α1- and β-AR. Only α1D-AR and β2-AR are expressed in JB6 P+ cells. (B) 10 ng/mL EGF-mediated JB6 P+ cell colony formation in soft agar (black) n = 8 and 72-hour SRB cell viability assay (blue) n = 6. Data represented as mean ± SD after normalization to control (EGF minus DMSO control for soft agar, and DMSO for SRB and MTS). A red asterisk (*) indicate that the HEAT HCl data are statistically lower than control (p < 0.05 as per an ANOVA with Tukey-Kramer post hoc test), which is not shown for the colony formation assay, indicating toxicity.

Fig 5. Antagonism of adrenergic receptors fails to prevent carvedilol-mediated inhibition of EGF-induced neoplastic transformation of JB6 P+ cells.

JB6 P+ cells were exposed to EGF (10 ng/ml) and increasing concentrations of carvedilol (Car) in the absence and presence of (A) 10 μM nadolol, (B) 10 μM CGP 12177, or (C) 100 nM HEAT HCl. Cells were cultured for 14 days and the colonies counted under a microscope, n = 8. Data represented as mean ± SD after normalization to control (EGF alone minus DMSO control).

Fig 6. Knockdown of the β2-AR fails to prevent carvedilol-mediated inhibition of EGF-induced neoplastic transformation of JB6 P+ cells.

(A) qPCR indicates that transduction of JB6 P+ cells with lentiviruses carrying a short hairpin sequence for Adrb2 decreases β2-AR expression, n = 3. Statistical analyses were determined using a 2-Factor ANOVA followed by Tukey Kramer post hoc test; different Greek letters signify statistical differences (P < 0.05). Data represented as mean ± SD. (B) After 3-days of infection, the JB6 P+ cells were exposed to 10 ng/ml EGF for seven days and increasing concentrations of carvedilol; n = 3 independent experiments with eight internal replicates detailed in Table 2. Data represented as mean ± SD after normalization to control (EGF alone minus DMSO control).

Fig 7. Short term UV-induced epidermal thickening in wild-type and β2-adrenergic receptor knockout mice.

Hairless mice were exposed to 200 mJ/cm² UV on day 1 and subsequently exposed to radiation every other day. On opposing days bi-fold epidermal thickness was measured with calipers, then the mice were treated with vehicle, 5 μM 4-OHC, or 5 μM carvedilol. Closed symbols represent wild type and open symbols knockout. Statistical analysis was conducted by a 3-Way RM-ANOVA and a Tukey-Kramer post hoc test. For clarity, statistical differences are noted in the text. Data represented as mean ± SD; n = 3, except for 4-OHC which had two mice in each group.