Phosphodiesterase 4 Inhibition as a Therapeutic Target for Alcoholic Liver Disease: From Bedside to Bench

Abstract

Alcoholic liver disease (ALD) is a major cause of liver-related mortality. There is still no US Food and Drug Administration-approved therapy for ALD, and therefore, identifying therapeutic targets is needed. Our previous work demonstrated that ethanol exposure leads to up-regulation of cAMP-degrading phosphodiesterase 4 (PDE4) expression, which compromises normal cAMP signaling in monocytes/macrophages and hepatocytes. This effect of ethanol on cAMP signaling contributes to dysregulated inflammatory response and altered lipid metabolism. It is unknown whether chronic alcohol consumption in humans alters hepatic PDE4 expression and cAMP signaling and whether inadequate cAMP signaling plays a pathogenic role in alcohol-induced liver injury. Our present work shows that expression of the PDE4 subfamily of enzymes is significantly up-regulated and cAMP levels are markedly decreased in hepatic tissues of patients with severe ALD. We also demonstrate the anti-inflammatory efficacy of roflumilast, a clinically available PDE4 inhibitor, on endotoxin-inducible proinflammatory cytokine production ex vivo in whole blood of patients with alcoholic hepatitis. Moreover, we demonstrate that ethanol-mediated changes in hepatic PDE4 and cAMP levels play a causal role in liver injury in in vivo and in vitro models of ALD. This study employs a drug delivery system that specifically delivers the PDE4 inhibitor rolipram to the liver to avoid central nervous system side effects associated with this drug. Our results show that PDE4 inhibition significantly attenuates ethanol-induced hepatic steatosis and injury through multiple mechanisms, including reduced oxidative and endoplasmic reticulum stress both in vivo and in vitro. Conclusion: Increased PDE4 plays a pathogenic role in the development of ALD; hence, directed interventions aimed at inhibiting PDE4 might be an effective treatment for ALD.

Figure 1.. Severe alcoholic hepatitis patients have low hepatic cAMP and high PDE4 levels.

(A) Representative images of livers stained with hematoxylin and eosin from five donors (no-AH) and five severe alcoholic hepatitis patients (AH). Arrows point to ballooning of hepatocytes, rectangles mark lobular activity and fibrotic septa are marked by ovals. (B) Hepatic cAMP levels were assessed by cAMP ELISA and normalized to gram of tissue. (C) Representative Western blot analysis of hepatic protein levels of PDE4A, PDE4B and PDE4D. (D) mRNA expression of PDE4A, PDE4B and PDE4D from livers was assessed using real time qRT-PCR. (E) Representative images of immunohistochemical analysis of PDE4A, B and D for one donor (no-AH-3) and one severe AH (AH-5) livers. Black circles represent donor livers obtained from Johns Hopkins University, open circles represent tissues from Brown Cancer Center at University of Louisville, black rectangles represent actively drinking AH patients, open rectangles show the levels from abstaining ASH patients who underwent liver transplantation. Values are mean ± SD, *<0.05, ****<0.001.

Figure 1.. Severe alcoholic hepatitis patients have low hepatic cAMP and high PDE4 levels.

(A) Representative images of livers stained with hematoxylin and eosin from five donors (no-AH) and five severe alcoholic hepatitis patients (AH). Arrows point to ballooning of hepatocytes, rectangles mark lobular activity and fibrotic septa are marked by ovals. (B) Hepatic cAMP levels were assessed by cAMP ELISA and normalized to gram of tissue. (C) Representative Western blot analysis of hepatic protein levels of PDE4A, PDE4B and PDE4D. (D) mRNA expression of PDE4A, PDE4B and PDE4D from livers was assessed using real time qRT-PCR. (E) Representative images of immunohistochemical analysis of PDE4A, B and D for one donor (no-AH-3) and one severe AH (AH-5) livers. Black circles represent donor livers obtained from Johns Hopkins University, open circles represent tissues from Brown Cancer Center at University of Louisville, black rectangles represent actively drinking AH patients, open rectangles show the levels from abstaining ASH patients who underwent liver transplantation. Values are mean ± SD, *<0.05, ****<0.001.

Figure 2.. Effect of PDE4-specific inhibitor, Roflumilast on LPS-inducible TNF and IL-1β production in AH patients’ whole blood ex vivo.

Whole blood was drawn after fasting from twenty AH patients before they started any treatment. Blood was diluted with RPMI media and plated in 24 well plates. Diluted blood was either left untreated (UT) or stimulated with 100 ng/ml LPS overnight. Additional wells were pretreated with 0.1 and 1 μM active metabolite of Roflumilast (Roflumilast N-oxide - RNO) 30 minutes before LPS stimulation. Cytokine levels in cell free supernatant were measured by using MSD platform. Values are mean ± SD, n=20, **<0.01.

Figure 3.. Chronic-binge ethanol feeding decreases hepatic cAMP levels in PDE4-dependent manner in mice.

(A) Distribution of labeled-FLVs overtime after i.p. infusion. Near-infrared dye is mostly localized to liver after 24h. There are negligible levels of signal present in the brain, lungs or heart. (B) Chronic-binge ethanol feeding increases hepatic Pde4b and Pde4d mRNA levels. (C) Representative images of hepatic immunohistochemical analysis of Pde4b and d in PF and AF mice. (D) PDE4 inhibitor prevents ethanol-mediated decrease in hepatic cAMP levels. Values are mean ± SD, n=6–8 in each group, *<0.05, **<0.01. PF-pair-fed, AF-chronic-binge ethanol fed, Rol -Rolipram. FLVs-Rol – Rolipram delivered using fusogenic lipid vesicles.

Figure 4.. PDE4 inhibition attenuates ethanol-induced hepatic injury and steatosis.

(A) Rolipram decreases ethanol induced liver injury as shown by plasma ALT and AST levels. (B) Rolipram attenuates the systemic inflammatory marker, sCD14, in ethanol fed mice. (C) H&E staining, arrow indicates neutrophil infiltration and (D) Oil Red O staining of liver sections. (E) Hepatic triglycerides (TAG). (F) Hepatic free fatty acids (NEFA). Values are mean ± SD, n=6–8 in each group, *<0.05, **<0.01, ***<0.001, ****<0.0001. PF-pair-fed, AF-chronic-binge ethanol fed, Rol -Rolipram. FLVs-Rol – Rolipram delivered using fusogenic lipid vesicles.

Figure 5.. Decreased hepatic endoplasmic reticulum (ER) stress by PDE4 inhibition in chronic-binge ethanol fed mice.

(A) Representative images of hepatic immunostaining with 4-hydroxynonenal (4-HNE) and acrolein-FDP-lysine antibodies show increased accumulation of lipid peroxidation product 4-HNE and acrolein-protein adducts in ethanol fed mice (AF), which is reduced in mice treated with Rolipram (Rol+AF and FLVs-Rol+AF). (B) Representative images of hepatic Dihydroethidium (DHE) staining to detect ROS formation demonstrating the effect of Rolipram on reducing ROS formation. (C) Hepatic CYP2E1 protein levels analyzed by Western blot, representative images of three mice per treatment group. (D) Real time qPCR analysis of hepatic Sod1 and Sod2 mRNA levels shows that Rolipram treatment alleviates the inhibitory effect of chronic-binge ethanol feeding on Sod1/2 mRNA levels. n=6–8 in each group. (E) Hepatic Atf3, Atf4, Chop and Gadd34 mRNA levels. (F) Representative Western blot analysis shows the effect of Rolipram attenuating hepatic CHOP and ATF3 protein levels in ethanol fed mice. Numbers on Western blot represent mean densitometry ratios and SD normalized to GAPDH. Values are mean ± SD, *<0.05, **<0.01, ***<0.001, ****<0.0001.

Figure 5.. Decreased hepatic endoplasmic reticulum (ER) stress by PDE4 inhibition in chronic-binge ethanol fed mice.

(A) Representative images of hepatic immunostaining with 4-hydroxynonenal (4-HNE) and acrolein-FDP-lysine antibodies show increased accumulation of lipid peroxidation product 4-HNE and acrolein-protein adducts in ethanol fed mice (AF), which is reduced in mice treated with Rolipram (Rol+AF and FLVs-Rol+AF). (B) Representative images of hepatic Dihydroethidium (DHE) staining to detect ROS formation demonstrating the effect of Rolipram on reducing ROS formation. (C) Hepatic CYP2E1 protein levels analyzed by Western blot, representative images of three mice per treatment group. (D) Real time qPCR analysis of hepatic Sod1 and Sod2 mRNA levels shows that Rolipram treatment alleviates the inhibitory effect of chronic-binge ethanol feeding on Sod1/2 mRNA levels. n=6–8 in each group. (E) Hepatic Atf3, Atf4, Chop and Gadd34 mRNA levels. (F) Representative Western blot analysis shows the effect of Rolipram attenuating hepatic CHOP and ATF3 protein levels in ethanol fed mice. Numbers on Western blot represent mean densitometry ratios and SD normalized to GAPDH. Values are mean ± SD, *<0.05, **<0.01, ***<0.001, ****<0.0001.

Figure 6.. PDE4 inhibition attenuates hepatocytes apoptosis in chronic-binge ethanol fed mice.

(A) Representative TUNEL staining of liver sections and (B) quantification of apoptosis. **<0.01, ***<0.001, ****<0.0001. Representative Western blot images of three mice per treatment group of (C) anti-apoptotic Bcl-xl and (D) Pro-caspase protein levels. *<0.05 compared to other two groups.

Figure 7.. Rolipram protects hepatocytes against ER stress and apoptosis in vitro.

(A) H4IIEC3 cells were left untreated (Ctrl+UT) or treated with ethanol (EtOH, 50mM, 24h) and TNFα (20ng/mL) for an additional 3 h followed by staining with Annexin V and propidium iodide to evaluate survival. Rolipram treatment reduced Annexin V-positive cell number from 65.1 to 35 treated with Ethanol and TNF. (B) Chop mRNA was quantified in H4IIEC3 cells treated as described in (A). (C) Western blot analysis of Bcl-xl in ethanol treated cells, R-Rolipram, B - PKA agonist Sp 5,6-DCl-cBIMPS (100 μM), 8C – EPAC agonist 8-pCPT-2’-O-Me-cAMP (100 μM), T-TNF. (D) Control and ethanol-exposed cells were pretreated with 10 μM Rolipram and further stimulated with TNF, 20 ng/ml for 90 minutes. JNK activation was assessed by Western blot analysis for pJNK (Thr183/Tyr185). Values are mean ± SD, *<0.05, **<0.01, ****<0.0001.