Roles and mechanisms of human immunodeficiency virus protease inhibitor ritonavir and other anti-human immunodeficiency virus drugs in endothelial dysfunction of porcine pulmonary arteries and human pulmonary artery endothelial cells

Abstract

The objective of this study was to determine the effects of highly active antiretroviral therapy (HAART) drugs on pulmonary endothelial function. Porcine pulmonary arteries or human pulmonary arterial endothelial cells (HPAECs) were incubated with eight HAART drugs [ritonavir, indinavir, lopinavir, zidovudine (AZT), abacavir, stavudine, didanosine (ddI), and lamivudine] individually or in combination [three HAART drugs (3-plex; indinavir, stavudine, and ddI)] at their clinical plasma concentrations for 24 hours. Endothelium-dependent vasorelaxation in response to bradykinin was reduced significantly by the ritonavir in a concentration-dependent manner. Five other HAART drugs (indinavir, lamivudine, abacavir, AZT, and ddI) and the 3-plex significantly also impaired endothelium-dependent vasorelaxation in response to bradykinin. Five HAART drugs (ritonavir, indinavir, lamivudine, abacavir, and AZT) significantly decreased endothelial nitric oxide synthase (eNOS) expression and increased superoxide anion levels in both vessels and HPAECs. Furthermore, both ritonavir and AZT substantially activated ERK2 in HPAECs. Additionally, the antioxidants ginsenoside Rb1 and ginkgolide A effectively reversed HAART drug-induced vasomotor dysfunction and eNOS down-regulation. Inhibition of ERK1/2 also partially blocked ritonavir- and AZT-induced down-regulation of eNOS and vasomotor dysfunction. Thus, HAART drugs significantly impair endothelial functions of porcine pulmonary arteries and HPAECs, which may be mediated by eNOS down-regulation, oxidative stress, and ERK1/2 activation. These findings suggest that HAART drugs may contribute to the high incidence of pulmonary artery hypertension in human immunodeficiency virus-infected patients.

Figure 1

Effects of ritonavir, Rb1, and GA on vasomotor function in porcine pulmonary arteries. Porcine pulmonary artery rings were cultured with DMSO as a control or treated with anti-HIV drugs or antioxidants (Rb1 or GA) for 24 hours. A: Maximal contraction of the vessel rings in response to thromboxane A₂ analogue U46619 (3 × 10⁻⁷ mol/L). B: Precontracted vessels were tested for endothelium-dependent relaxation by adding a series of concentrations of bradykinin (10⁻¹¹ to 10⁻⁶ mol/L). C: Endothelium-independent relaxation in response to SNP (10⁻⁶ mol/L). D: Fifteen μmol/L ritonavir showed significant impairments in vasorelaxation compared with the control, whereas both Rb1 and GA significantly improved vasorelaxation compared with ritonavir alone group. *P < 0.05 versus controls (DMSO only). n = 3. RVT, ritonavir; Rb1, ginsenoside Rb1.

Figure 2

Effects of other HAART drugs on vasomotor function in porcine pulmonary arteries. Porcine pulmonary artery rings were treated with HAART drugs or with DMSO as a control for 24 hours. A: In response to 10⁻⁸ bradykinin, endothelium-dependent relaxation of vessels was significantly reduced by indinavir, lamivudine, abacavir, and AZT at their clinical plasma concentrations, respectively, compared with controls. B: Endothelium-dependent relaxation in response to bradykinin (10⁻¹¹ to 10⁻⁶ mol/L) was significantly reduced by AZT in a concentration-dependent manner compared with controls. C: Endothelium-dependent relaxation in response to bradykinin (10⁻⁸ mol/L) was significantly impaired by indinavir, ddI, or 3-plex, as compared with controls, whereas the Rb1 plus 3-plex group demonstrated a significant improvement in vasorelaxation compared with the 3-plex group (**P < 0.05, n = 3). *P < 0.05 versus controls (DMSO only). n = 3. Rb1, ginsenoside Rb1; IDV, indinavir.

Figure 3

Effects of HIV protease inhibitor ritonavir and other HAART drugs on eNOS expression in both porcine pulmonary arteries and HPAECs. Porcine pulmonary artery rings were cultured with or without HAART drugs for 24 hours, and the total mRNA was purified from endothelial layers of the rings. The eNOS mRNA was measured with real-time PCR. The eNOS mRNA level in each sample was normalized to CD31. Relative mRNA level was presented as 2^{−[Ct(CD31) − Ct(eNOS)]}. A: Ritonavir significantly decreased the porcine eNOS mRNA expression in a concentration-dependent manner compared with controls. B: The porcine eNOS mRNA was significantly reduced by indinavir, lamivudine, abacavir, and AZT at their clinical plasma concentrations, respectively, compared with controls. C: The eNOS protein levels in HPAECs were measured by flow cytometry. Consistent with mRNA levels, the treatments of the above drugs also induced decreases of eNOS protein levels in HPAECs (e, f, i, j, k, l). Rb1 or GA blocked the decrease of eNOS protein levels induced by RTV and AZT (g, h, m, n). x axis: eNOS expression fluorescence intensity; y axis: cell number. Individual percentage value represents the percentage of positive staining cells in the counted HPAECs from the different treatments. Figure 3C. a. No staining (control); b. Ab staining only; c. Rb1 (10 μM) only; d. GA (10 μM) only; e. RTV (15 μM); f. RTV (30 μM); g. RTV (15 μM) plus Rbl; h. RTV (15 μM) plus GA; i. Indinavir; j. Lamivudine; k. Abacavir; l. AZT; m. plus Rbl; and n. AZT plus GA. D: eNOS protein levels in HPAECs were demonstrated by Western blot. Quantitation of the band density was performed. *P < 0.05 versus controls (DMSO only). **P < 0.01 versus RTV (15 μM) treatment alone. n = 3. Rb1, ginsenoside Rb1; RTV, ritonavir; IDV, indinavir.

Figure 4

Effects of the HIV protease inhibitor ritonavir and other anti-HIV drugs on superoxide anion production in both porcine pulmonary arteries and HPAECs. Superoxide anion levels in the endothelial layer of porcine pulmonary arteries were tested with lucigenin-enhanced chemiluminescence assay. A: Ritonavir significantly increased the superoxide anion levels of the vessel rings in a concentration-dependent manner compared with controls. B: Superoxide anion levels of the vessel rings were significantly reduced by indinavir, lamivudine, abacavir, and AZT at their clinical plasma concentrations, respectively, compared with controls. C: Superoxide anion levels in HPAECs were stained with DHE and analyzed by FACS Calibur flow cytometry. Similarly, superoxide anion levels in HPAECs were reduced by ritonavir (c, d), indinavir (e), lamivudine (f), abacavir (g), and AZT (h). Individual percentage value represents the percentage of positively staining cells in the counted HPAECs from different treatments. Figure 4C. a. No DHE (negative control); b. DHE only (staining control); c. RTV (15 μM); d. RTV (30 μM); e. Indinavir; f. Laminudine; g. Abacavir; and h. AZT. D: Superoxide anion levels of the vessel rings were significantly impaired by 3-plex, whereas Rb1 significantly inhibited the increased levels of superoxide anion induced by 3-plex (**P < 0.05). *P < 0.05 versus controls (DMSO only). n = 3. Rb1, ginsenoside Rb1; RTV, ritonavir; IDV, indinavir.

Figure 5

Effects of HIV protease inhibitor ritonavir and other anti-HIV drugs on mitochondrial membrane potential in HPAECs. HPAECs were treated with ritonavir, indinavir, lamivudine, abacavi, and AZT at their clinical plasma concentrations, respectively, for 24 hours. A: Mitochondrial membrane potential was assessed with JC-1 staining and flow cytometry analysis. DMSO treatment was used as a control (a). Normal potential (red) in HPAECs was decreased in the treatment with ritonavir (b), indinavir (c), lamivudine (d), abacavir (e), and AZT (f). Individual percentage value represents the percentage of positive cells with mitochondrial JC-1 staining in HPAECs. B: Quantitative analysis of mitochondrial membrane potential in HPAECs. *P < 0.05 versus controls (DMSO only). n = 3. RTV, ritonavir.

Figure 6

Roles of MARK in HAART drug-induced endothelial dysfunction in HPAECs and in porcine pulmonary artery rings. HPAECs were treated with ritonavir (15 μmol/L) or AZT (8 μmol/L). The phosphorylated and total ERK2, JNK, and p38 proteins were detected by Bio-Plex immunoassay. A: Ritonavir treatment increased the ratios of phosphorylated and total ERK2, JNK, and p38 proteins at 20 (a), 90 (b), and 60 (c) minutes, respectively. B: AZT treatment increased the ratios of phosphorylated and total ERK2 proteins at 30 (a) minutes. However, there were no substantial changes of phosphorylation of JNK (b) and p38 (c) in response to AZT treatment. C: The blockage effect of MEK/ERK inhibitor (ERKi) on reduced expression of eNOS in HPAECs induced by AZT and RTV was demonstrated by Western blot. D: Superoxide anion levels in HPAECs were stained with DHE and analyzed by FACSCalibur flow cytometry. MEK/ERK inhibitor (ERKi) did not show any significant effect on the increased superoxide anion in HPAECs induced by AZT and RTV. E: MEK/ERK inhibitor (ERKi) partially blocked AZT- or RTV-induced decrease in endothelium-dependent vasorelaxation in response to bradykinin (10⁻⁶ mol/L) in porcine pulmonary artery rings. RTV, ritonavir; ERKi, PD98059.