Mg supplementation attenuates ritonavir-induced hyperlipidemia, oxidative stress, and cardiac dysfunction in rats

Abstract

Use of protease inhibitors (PI) in HIV patients is associated with hyperlipidemia and increased risk of coronary heart disease. Chronic systemic and cardiac effects of ritonavir (RTV), a universal PI booster, and Mg supplementation were examined. RTV was administered (75 mg·kg⁻¹·day⁻¹ po) to Lewis×Brown-Norway hybrid (LBNF1) rats for up to 8 wk; significant increases in plasma triglyceride and cholesterol occurred from 8 days to 8 wk. At 5 wk, the expression of selected hepatic genes (CYP7A1, CITED2, G6PC, and ME-1), which are key to lipid catabolism/synthesis, were altered toward lipogenesis. Dietary Mg supplementation (six-fold higher) completely reversed the altered expression of these genes and attenuated both hypertriglyceridemia and hypercholesterolemia. Neutrophils isolated from the RTV-treated rats displayed a three-fold higher basal and a twofold higher stimulated superoxide production; plasma isoprostane and red blood cell (RBC) GSSG levels were elevated two- to three-fold. All oxidative indices were normalized by Mg supplementation. After 5 wk, RTV caused significant decreases in cardiac left ventricular (LV) shortening fraction and LV ejection fraction; mitral valve early/late atrial ventricular filling (E/A) ratio was reduced accompanied by LV posterior wall thinning. Immunohistochemical staining revealed significant white blood cell (WBC) infiltration (5 wk) and prominent fibrosis (8 wk) in the RTV hearts. Mg supplementation attenuated RTV-induced declines in systolic and diastolic (improved mitral valve E/A ratio) function (>70%), lessened LV posterior wall thinning (by 75%), and substantially decreased the pathological markers. The known clinical hyperlipidemia effects of RTV can be mimicked in the LBNF1 rats; in association, systemic oxidative stress and progressive cardiac dysfunction occurred. Remarkably, Mg supplementation alone suppressed RTV-mediated hyperlipidemia, oxidative stress, and cardiac dysfunction.

according to a recent report (47), there are 34 million people worldwide infected with HIV-1, and of these, 1.3 million are in the United States (7). Nucleoside/nucleotide analogs, with Zidovudine (AZT) as the prototype compound, were first introduced in the late 1980s and continue to play a significant role in HIV therapy (13). With the introduction of protein inhibitors (PIs) in the mid 1990s, HIV-1 replication in the patients was shown to be dramatically reduced, and currently, to the extent that HIV-1 infection has become a more manageable disease (20, 23). However, along with the chronic use of PIs, significant metabolic side effects of hyperlipidemia, tissue lipodystrophy, and insulin resistance may result (18, 42). Hyperlipidemia may contribute to the increased atherosclerotic vascular disease, resulting in endothelial and cardiac dysfunction (18, 19). Prior studies strongly suggested that ritonavir (RTV) alone may elicit endothelial dysfunction through oxidative mechanisms (11, 27). Nevertheless, the relationship between RTV-mediated oxidative stress and lipogenesis remains unclear. RTV is an early-generation prototypical PI; since introduction, RTV has been continuously used as a key component of the first line highly active antiretroviral treatment (HAART) regimens (13). Treatment with RTV, as a monotherapy, resulted in substantial elevations in plasma cholesterol and TG levels (14). As a result, RTV usage at its full dose has been discontinued; nevertheless, it continues to be used at a lower dose to “boost” serum concentrations of other PIs (amprenavir, atazanavir, lopinavir, and saquinavir), by inhibiting their metabolism via the P-450 3A4 isozyme (51).

The role of Mg and hypertension on lipid metabolism remains unclear. We, as well as others (5, 29, 41), have reported that Mg deficiency may promote hyperlipidemia. Moreover, an earlier study indicated that Mg supplementation could attenuate diet-induced triglyceridemia in a rabbit model (3). Recently, we reported that Mg supplementation provided effective antiperoxidative effects against AZT-induced systemic oxidative stress and cardiac inflammation in normomagnesemic rats (31). The Lewis × Brown Norway F1 hybrid rat (LBNF1) was recently shown to be very sensitive to RTV-induced hyperlipidemia in a manner similar to that exhibited by humans (50). In the current study, we used this rat model to examine the effects of dietary Mg supplementation on RTV-mediated hyperlipidemia, and changes in hepatic lipid metabolism genes. We also assessed whether RTV promoted systemic oxidative stress and enhanced cardiac dysfunction, and whether Mg supplementation effectively attenuates these deleterious side effects.

MATERIALS AND METHODS

All animal experiments were guided by the principles for the care and use of laboratory animals as recommended by the U.S. Department of Health and Human Services and were approved by The George Washington University Institutional Animal Care and Use Committee.

Animal model, RTV treatment, and diets.

Male LBNF1 rats (160 g, 4–5 wk old) were custom-ordered from Harlan Laboratories (Indianapolis, IN). After 1 wk of quarantine, the rats were divided into five groups (8 animals per group): 1) None Control, 2) Vehicle Control; 3) RTV alone, 4) RTV+Mg supplement, and 5) Mg supplement alone. Groups 1, 2, and 3 were fed a Mg-normal (0.1% MgO/kg food) diet; groups 4 and 5 received a Mg-supplemented (0.6% MgO) diet (31) containing extracted casein as the diet base and essential vitamins and nutrients (Harlan Teklad, Madison, WI) for up to 8 wk. RTV administration (75 mg·kg⁻¹·day⁻¹) was given to groups 3 and 4 by gavage; groups 2 and 5 received identical volumes of drug vehicle solution (10% ethanol, 5% DMSO, 50% propylene glycol, and 35% water). Blood samples were collected periodically by tail bleeding or by cardiac puncture during euthanasia from nonfasting animals. Plasma samples were obtained immediately by centrifugation at 200 g × 10 min and stored at −80°C until needed. At the end of week 5, five animals from each group received noninvasive echocardiography and then were euthanized for tissue pathology and blood biochemistry analyses. The experimental duration for the remaining three animals per group was extended to 8 wk to allow cardiac functional and pathological assessment after this prolonged period. Liver and cardiac samples were rapidly excised, processed, frozen, and stored at −80°C. Plasma Mg levels were determined by flame emission atomic absorption spectroscopy (AA) (31, 34).

Selected hepatic lipogenic gene expression determined by real-time quantitative PCR.

For the LBNF1 rat liver gene expression, 30 mg of liver from all groups were homogenized with and extracted for total RNA using TRIzol reagent (Invitrogen, Carlsbad, CA), according to the manufacturer's procedures. cDNA was synthesized and amplified from total RNA using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA), and quantitative real-time PCR was performed using SYBR GreenER qPCR Super Mix (Invitrogen) with the Applied Biosystems 7300 real-time PCR system. The hepatic expression levels of CYP7A1, G6PC, CITED2, and ME1 genes were quantified with expression values normalized to GAPDH mRNA. Rat GAPDH primers were forward 5′-GGGGCTCTCTGCTCCTCCCTG-3′ and reverse 5′-GAGACGAGGCTGGCACTGCAC-3′. For gene expression of the selected lipogenic genes, the following primers were used: CYP7A1, forward: GCCGTGTTGGTGAGCTGTTGC and reverse: GGAGGTTCACCAGCTTTCCTTCTCC; CITED2, forward: CGCAAAGACGGAAGGACTGGAA and reverse: CTCGGGAACTGCCCCATGCC; G6P-catalytic, forward: GCGCCCGTATTGGTGGGTCC and reverse: GGGACTCCCTGGTCCAGTCTCA; and malic enzyme-1, forward: CTGCCTTCAACGAGCGGCCC and reverse: GATCGCACGGCCCTTGGTCA.

Determination of plasma lipids, systemic oxidative indices, and neutrophil ROS activity.

Plasma total triglyceride (TG) content was determined quantitatively by a colorimetric (at 570 nm) EnzyChrom triglyceride assay kit (BioAssay Systems, Hayward, CA), according to instructions. Plasma total cholesterol levels were determined by a fluorometric method using a Cayman's cholesterol assay kit (4), and total plasma cholesterol content was estimated on the basis of relative fluorescence obtained from standards using an excitation of 550 nm and emission of 590 nm. Plasma 8-isoprostane levels were determined by an enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI) (31, 32, 35), and RBC reduced (GSH) and oxidized (GSSG) glutathione levels were assessed enzymatically by the DTNB-GSSG reductase method (31, 33–35). Neutrophils from whole blood were isolated by a step-gradient centrifugation within 20 min after death, and both basal (no stimulus) and PMA-stimulated (by 125 ng/ml phorbol 12-myristate 13-acetate) superoxide productions were assayed as described previously (31, 32, 35). Superoxide anions released were estimated by SOD-inhibitable reduction of cytochrome c using the extinction coefficient: E550= 2.1 × 10.4 M⁻¹ cm⁻¹.

Histochemical analysis.

Cardiac tissue was rapidly excised, rinsed in saline, embedded in OCT compound, quickly frozen and kept at −80°C until used. Cryosections, 5 μm thick, were stained by an indirect, immunohistochemical staining method (31) using mouse anti-rat CD11b antibody (1:500; Millipore, Billerica, MA) and the Vecta-Stain Elite ABC kit immunoperoxidase system (Vector Laboratories, Burlingame, CA). Samples were examined under an Olympus BX60 microscope, and multiple images were taken with a digital camera (Evolution Color MP; Media Cybernetics, Silver Spring, MD). CD11b-positive cells were counted in the micrographs, and quantification was normalized on the basis of standardized area obtained with Adobe Photoshop. Routine H&E and Masson Trichrome [for collagen fibers by Accustain trichrome stains (Masson) kit, Sigma Aldrich, St. Louis, MO] staining were performed on 5-μm-thick cryosections for overall morphology and collagen content (31, 33). The Accustain protocol included treatment with Bouin's solution, which provided fixing of the tissue sections with formaldehyde and intensified the final coloration; therefore, more consistent and reliable images for fibrosis were obtained. ImageJ software (Image processing and analysis in Java, http://imagej.nih.gov) was used to quantify the area of the blue-stained collagen in the images of ventricle tissue sections stained with Trichrome Masson Blue, according to the procedure of Rangan and Tesch (40). ImageJ is a reliable program for computer-assisted analysis in bright-field microscopy, provided that the micrographs can be easily converted into binary black and white images, which were readily achieved in our case. A 50-μm bar was used by the software to set the scale to micrometers so that area measurements on selected images were expressed in square micrometers. Briefly, 24-bit RGB images were converted into three 8-bit grayscale images by using the red-green-blue stack function and analyzed through the red channel, where contrast between the collagen areas and the background was the greatest. Then by selecting the “Image/Adjust/Threshold” function, the threshold was set and manually adjusted for collagen area staining detection. The threshold tool allows a lower and upper value to segment the grayscale image into areas of interest and to exclude background. When satisfied with the threshold levels, the function “Apply”, converted the image to binary (black and white) staining. The area included within the threshold was then calculated through ImageJ's “Measure” function and reported in square micrometers.

Echocardiography.

Cardiac functional/anatomical parameters were assessed during anesthesia (2% inhaled isoflurane mixed with 100% oxygen) periodically using the GE VingMed System Five (GE Healthcare Canada) with a 10-MHz probe (28, 32). Aortic and pulmonary artery diameters were measured to calculate stroke volume. The left ventricular wall thickness and internal diameters in systole and diastole were used to calculate the shortening fraction. Spectral Doppler velocities of the pulmonic and aortic outflows were measured to calculate cardiac output, and the tricuspid and mitral valve inflows were measured to assess ventricular diastolic function.

Statistical analysis.

Data were checked by F-test for equality of groups' variation and paired Student's t-test for each subgroup vs. reference subgroup. Data were shown using means ± SE of four to eight animals per group for up to 5 wk and three animals for the 8-wk determinations. Significance was considered at P < 0.05; selected data were analyzed using two-way ANOVA followed by Tukey's test.

RESULTS

RTV treatment and Mg diets effect on weight gain (5 wk) and plasma Mg levels.

Treatment of LBNF1 rats with RTV for 5 wk resulted in a moderate but significant weight loss (Table 1, −14.5%, P < 0.05) compared with untreated (none) controls; Mg supplementation attenuated part of the weight loss (−9.9%, NS vs. vehicle control or vs. RTV alone) in RTV-treated rats, whereas Mg supplementation alone did not alter weight gain. Using AA, we found that RTV alone caused a modest, but insignificant, hypomagnesemia (−12%, P = 0.1); however, Mg supplementation with or without RTV coadministration significantly (P < 0.01) raised the plasma Mg levels to 30–33% above controls (Table 1). At the end of 8 wk, RTV alone caused a −5% change (not significant, NS) in plasma Mg levels, whereas that for RTV+ Mg supplement and Mg-supplemented groups maintained +28% and +30% (both P < 0.05), respectively, above none controls.

Table 1.

RTV treatment (5 wk) and Mg supplementation on rat body

Weight and Plasma Mg Concentration
	Body Weight (g)
Rat Groups	Day 0	5 Weeks	Plasma Mg [mM]
None Ctl	187 ± 6	332 ± 8	0.790 ± 0.052
Veh. Ctl	189 ± 5	329 ± 6	0.797 ± 0.045
RTV	183 ± 5	284 ± 10*	0.706 ± 0.056
RTV+Mg-supp.	187 ± 4	295 ± 11+	1.053 ± 0.063**
Mg-supp. Alone	179 ± 7	326 ± 8	1.040 ± 0.070**

Data are expressed as means ± SE; n = 5–8.

^*P < 0.05,

^**P < 0.01 vs. None Ctl or Veh. Ctl; +P < 0.05 vs. None Ctl, but NS vs. Veh. Ctl or RTV alone.

Selected hepatic gene expressions related to lipid metabolism.

Sprague-Dawley (SD) rats were found to be less sensitive to PI-induced lipogenesis than LBNF1 rats. Using microarray assessment, Yang et al. (50) reported that a subset of four genes important to lipid and carbohydrate metabolism were differentially regulated by RTV in SD and LBNF1 rats after only 5 days of RTV treatment. Using a RT quantitative-PCR approach, we assessed the effects of RTV treatment at a comparable dose (75 mg·day⁻¹·kg⁻¹) on this subset of genes. Indeed, 5 wk of RTV treatment significantly downregulated the expressions of CYP7A1, CITED2, and G6PC by 47% (P < 0.01), 53% (P < 0.025), and 60% (P < 0.01), respectively, but upregulated ME-1 robustly by 3.2-fold (P < 0.025). When Mg supplementation was introduced in addition to RTV, the decreased expressions of CYP7A1, CITED2, and G6PC, were completely prevented (P < 0.025; Fig. 1A) and achieved levels comparable to (for CYP7A1 and G6PC) or above (for CITED2) control values. Concomitantly, the more than three-fold upregulation of ME-1 induced by RTV was lowered to only 0.8-fold of control levels by Mg supplementation (Fig. 1B). Of note, the mRNA levels of CITED2 in animals with Mg supplementation alone were above 200% of controls.

Fig. 1.

Real-time PCR results representing the effects of ritonavir (RTV) treatment (75 mg·kg⁻¹·day⁻¹) for 5 wk with or without Mg supplementation (6-fold) on LBN-F1 rat hepatic mRNA levels of selected hyperlipidemic genes: CYP7A1, CITED2, G6PC (A), and ME-1 (B). Values are expressed as means of 4–5 ± SE. #P < 0.025 vs. Veh Ctl and vs. RTV+Mg-supp; +P < 0.05; ++P < 0.01 vs. RTV alone.

Changes in plasma triglyceride and cholesterol levels.

The temporal effects of RTV administration on changes in the blood lipids for the entire experimental period were assessed. As shown in Fig. 2, administration of RTV for only 8 days resulted in dramatic and significant elevations in both plasma TG (A: 75% increase) and cholesterol (B: 80% increase) levels. The time course monitoring revealed that both TG and cholesterol levels remained elevated throughout most of the 8-wk period of RTV administration. Impressively, Mg supplementation almost completely (>90%) and significantly (P < 0.01 or <0.05) suppressed the elevations of both TG and cholesterol caused by RTV. Vehicle alone or Mg supplementation alone did not alter the plasma TG and cholesterol levels compared with none controls.

Fig. 2.

Time course changes in plasma total triglyceride (A) and cholesterol (B) levels in RTV-treated rats ± Mg supplementation. Plasma samples were obtained from LBN-F1 rats receiving RTV orally (75mg·kg⁻¹·day⁻¹) for 8 days, 3, 5, and 8 wk, while on normal or six-fold supplemented Mg diets. Controls received vehicle only. Means of 5–8 rats ± SE; n = 3 at 8 wk; #P < 0.001, **P < 0.01, *P < 0.05 vs. Veh. Ctl or Mg-supp alone; +P < 0.05, ++P < 0.01 vs. RTV alone but NS vs. Mg-supp.

Oxidative stress indices.

RTV was reported to promote elevated superoxide anion production in cultured cells and tissues in vitro (11). Neutrophils (PMNs) can be a major cellular source of reactive oxygen species (ROS) production. Larger volume of blood samples obtained from euthanized animals (5 wk) allowed us to examine both basal and stimulated ROS generation activity in the isolated PMNs. As represented in Fig. 3A, PMNs from control rats displayed a low level of basal superoxide-generating activity; however, this activity was significantly elevated 3.2-fold in the RTV-treated group, and cotreatment with the dietary Mg supplement substantially suppressed (∼80%) this elevated basal superoxide-generating activity. When PMA (125 ng/ml) was included in the assay, all samples were stimulated to produce higher levels of superoxide anion (Fig. 3A). However, PMNs from the RTV-treated rats also exhibited a twofold higher activity compared with stimulated PMNs from the control group. Mg supplementation lowered the PMA-stimulated activity, approaching levels in control samples.

Fig. 3.

A: RTV promotes basal and stimulated superoxide generation, which were attenuated by Mg supplementation. Neutrophils were isolated by a step gradient centrifugation method from the whole blood of 5-wk rats receiving RTV with or without Mg supplementation. The ability of the neutrophils to generate superoxide anions ± 125 ng PMA was determined by SOD-inhibitable cytochrome-c reduction. Values are expressed as means ± SE from 5 rats. #P < 0.025 vs. Veh. Ctl. +P < 0.05 vs. RTV. RTV enhances circulating 8-isoprostane (B) and RBC GSSG (C) levels in normal Mg rats and the attenuating effects of Mg supplementation. Plasma 8-isoprostane was determined by an EIA immunoassay kit, and RBC glutathione was determined enzymatically by the DTNB-GSSG reductase method; other conditions are as described in Fig. 2. *P < 0.05, **P < 0.01 vs. Veh. Ctl; +P < 0.05 vs. RTV alone.

As a key index of systemic oxidative stress, plasma levels of F₂-like isoprostanes, which are derived from nonenzymatic peroxidation of polyunsaturated fatty acids, were determined (31, 35); this index was often used in previous studies assessing HAART effects in HIV patients (22, 25). As shown in Fig. 3B, RTV treatment alone resulted in a significant (P < 0.01) increase (50% higher than Veh. Ctl) in plasma 8-isoprostane level. Concomitantly, the RBC GSSG level obtained from the 5-wk RTV-treated animals was elevated three-fold. When combined with Mg supplementation, both the elevated plasma isoprostane and RBC GSSG levels were attenuated to levels comparable to controls (Fig. 3, B and C). In data not shown, Veh. alone, or Mg supplementation alone caused no significant differences compared with None Ctls.

Cardiac WBC infiltration and fibrosis.

Immunohistochemical assessment revealed that 5 wk of RTV treatment resulted in substantial perivascular WBC infiltration, as represented by the enhanced staining for CD11b in ventricles from the RTV-treated rats (Fig. 4A1). Quantitative morphometric analysis based on the density of CD11b-positive cells per standardized area revealed that RTV treatment resulted in a 2.4-fold increase in ventricular WBC infiltration (Fig. 4A2). Cotreatment with Mg supplementation led to a significant reduction in WBC infiltrates; Mg supplementation alone did not have any effect on CD11b infiltrates and resembled vehicle control samples. Using the Masson Trichrome method, we found only mild fibrosis in ventricular tissue from two of the RTV-treated rats at 5 wk; however, by 8 wk, prominent fibrosis occurred in both the left and right ventricles of all examined RTV-treated rats, comparing to None Ctl (Fig. 4B1). Mg supplementation largely attenuated the collagen formation in the RTV-treated rats (Fig. 4, B1 and B2). A computer-assisted analysis for the fibrosis areas (4B2) in the different groups were None Ctl: 1.26 ± 0.12%; Veh. Ctl: 1.82 ± 0.14% (P = 0.055); RTV-treated: 3.46 ± 0.36% (P < 0.0025); and RTV+Mg-supp: 2.16 ± 0.19% (P < 0.01 vs. RTV alone). In data not shown, Mg supplementation alone caused no change in fibrosis area compared with None Ctl.

Fig. 4.

A1: micrographs (magnification: ×20) showing immunohistochemical staining (brown) for CD11b in cardiac ventricular sections from None Ctl, + RTV, and RTV + Mg-supplemented rats. All tissues were obtained from 5 wk groups. A2: numerical density of CD11b positive cells per square millimeter in ventricles from rats treated with RTV alone or RTV + Mg-suppl diets for 5 wk. Data were derived from 5–6 different fields per heart for 4 animals each group; ** P < 0.01 vs None Ctl, ⁺ P < 0.025 vs RTV alone. B1: Masson's Trichrome staining: fibrotic development in RTV treated ventricles of the 8 wk rats and the attenuation effect of Mg supplementation. B2: data were derived from 6–8 different fields per heart for three animals per group. Fibrotic area was expressed as [μm²] per constant area of 392,220 [μm²]; ^P = 0.055, #P < 0.0025 vs. Veh. Ctl; ++P < 0.01 vs. RTV alone.

Changes in cardiac function using noninvasive echocardiography.

The RTV administration studies were extended to 8 wk to assess the progressive development of cardiac dysfunction and the potential benefits of dietary Mg supplementation using echocardiography. As represented in Fig. 5, no significant alterations in echo parameters were noted after 3 wk of RTV treatment; however, at 5 wk, RTV caused a small, yet significant, decrease in left ventricular systolic function, as indicated by a 3.5% decrease in left ventricular ejection fraction (LVEF; Fig. 5A), and a 6.5% lowering of % fractional shorterning (FS; Fig. 5B). When extended to 8 wk, the progressive worsening of systolic function was prominent with RTV treatment: LVEF decreased further to 7%, and %FS fell further to 13.5%. By using two-dimensional, M-mode and pulsed Doppler images, we obtained measurements of left ventricular (LV) posterior wall thickness and internal diameter; substantial decreases were revealed in LV posterior wall thickness in diastole (LVPWd: 10%) and systole (LVPWs: 11%; P < 0.05) after 5 wk of RTV treatment (Fig. 6A), with further declines after 8 wk (LVPWd: 13%; and LVPWs: 22%; P < 0.025) indicative of progression toward a dilated cardiomyopathy. At 5 wk, RTV also resulted in LV diastolic dysfunction as evidenced by a 17% decrease (P < 0.05) in mitral valve early/late atrial ventricular filling (E/A) ratio (Fig. 6B). Moreover, cardiac output (CO) was significantly reduced by 12% (Fig. 6B). In data not shown, mitral valve E/A ratio or CO remained similarly depressed in the 8 wk RTV-treated rats. Cotreatment with Mg supplementation completely prevented RTV-mediated declines in LV systolic function (Fig. 5, A and B), substantially attenuated diastolic dysfunction (Fig. 6B: E/A ratio by 70%), restored CO to normal vehicle control levels (Fig. 6B), and lessened thinning of LVPWs by 75% at 8 wk (Fig. 6A).

Fig. 5.

Echocardiographic changes in left ventricular (LV) systolic function during prolonged RTV treatment with or without Mg supplementation in rats. RTV caused progressive declines in left ventricular ejection fraction (LVEF; A) and % fractional shortening of LV wall [LV % fractional shortening (FS); B] vs. time-matched vehicle controls for up to 8 wk, and Mg supplement was completely protective. Values are expressed as means ± SE; n = 4–6 rats; n = 3 at 8 wk. *P < 0.05, #P < 0.025 vs. Veh Ctl.

Fig. 6.

A: time-dependent changes in echocardiographic anatomical parameter: left ventricular posterior wall dimension in systole (LVPWs) during RTV treatment with or without Mg supplement in rats. Values are expressed as means ± SE; n = 4–6 rats; n = 3 at 8 wk. B: effect of RTV treatment with or without Mg supplement in rats after 5 wk on mitral valve E/A ratio, and cardiac output (CO) vs. time-matched vehicle control. Values are expressed as means ± SE; n = 4–6 rats. *P < 0.05 and #P < 0.025 vs. Veh Ctl.

DISCUSSION

We are encouraged by our findings that Mg supplementation can attenuate RTV-induced hyperlipidemia, oxidative stress, and cardiac dysfunction in the rat model. However, we do recognize that there are some limitations that exist in our study, including the fact that 1) RTV is normally administered to patients with another PI and that 2) the HIV-1 infection is not present in our animal model. Among all clinically used PIs, RTV is known to cause the most significant hyperlipidemia side effects in patients (39). Although no longer used as the primary PI for therapy, RTV is currently used as a universal “booster” to improve the availability of other prescribed PIs, due to its ability to inhibit the drug-metabolizing P-450 3A4 (51). Thus, studying effects of RTV alone would yield valuable information concerning its independent merits and side effects. Regarding the second limitation, it is likely that HIV infection alone may also cause separate lipid, metabolic, and cardiovascular consequences (27); however, results from this study may serve as useful references for the more complex clinical setting involving HIV-infected patients who receive different HAART combination regimens. A clinical study has demonstrated that even at a reduced dosage, RTV still caused elevated plasma fatty acids and cholesterol in healthy volunteers (43). Therefore, studies addressing this side effect of RTV alone remain relevant to the currently used HAART therapies. We also recognize that part of our study might be somewhat weakened by the relative low sample size, especially for the results obtained from week 8 groups. However, the observed cardiac dysfunction and changes in plasma lipid profile in the 8-wk experimental groups were consistent with the trend progression seen at 5 wk and earlier. In addition, it may be argued that the dose of RTV used in the current study is higher than (5–11-fold above) the recommended human dose. Nevertheless, the dose used in the current study is appropriate for rats, and this is recognized in a guidance issued by the U.S. Food and Drug Administration (16). Rats have much higher drug metabolic rates compared with humans and require at least 6-fold higher dose levels to achieve clinically comparable circulating HAART levels. In the present study, the selected dose of RTV did cause hyperlipidemia in LBNF1 rats, which are one of the few strains reported to exhibit this PI-induced sensitivity observed in humans (50). RTV induced sustained elevations in plasma cholesterol and TG levels throughout the entire 8-wk treatment period. Most importantly, to our knowledge, this is the first report showing that RTV-induced elevations in both cholesterol and TG levels were attenuated >90% by cotreating rats with a Mg-supplemented diet. However, the reduced body weight (−14.5%) likely reflects some degree of drug toxicity unrelated to hyperlipidemia; therefore, this loss of body weight was not corrected entirely by Mg supplementation.

The precise mechanisms underlying the antihyperlipidemic and cardioprotective effects of Mg supplementation during RTV treatment remain unclear. Mg may play a key role in the regulation of carbohydrate and lipid metabolism (44, 45). Severe magnesium deficiency can promote lipogenesis (5, 29, 41). In the present study, RTV treatment alone decreased circulating Mg levels, but this decrease was mild (−12%) and may not be the major cause leading to hyperlipidemia. In a cultured cell system, increased lipogenic gene transcription was promoted by uncontrolled ROS generation (26). In our magnesium-deficient hamster model (5), systemic hyperlipidemia (elevated cholesterol and triglyceride levels) was promoted by enhanced free radical generation associated with catecholamine administration. In the present study, chronic administration of RTV resulted in persistent oxidative stress, as evidenced by increased systemic isoprostane levels and elevated RBC GSSG content, along with activation of circulating neutrophil ROS generation. Quite possibly, the increased ROS stress caused by RTV directly or indirectly participated in altered expression of the metabolic genes leading to hyperlipidemia. Indeed, by assessing the effects of RTV treatment and the Mg effects on the selected hepatic gene expression, we found that RTV treatment resulted in substantial downregulation of CYP7A1, CITED2, and G6PC, but upregulation of ME-1. All four genes directly or indirectly promote hyperlipidemia either through increased cholesterol and fatty acid biosynthesis or decreased hepatic lipid degradation. CYP7A1 is also known as cholesterol 7-α-monooxygenase or cytochrome P-450 7A1, which is the rate-limiting enzyme in the synthesis of bile acid from cholesterol (10). The official designation for CITED2 is Cbp/p300-interacting transactivator with a Glu/Asp-rich carboxy-terminal domain; the gene product of CITED2 is a coactivator of the peroxisome proliferator-activated receptor-α pathway which promotes fatty acid beta-oxidation (15). Thus, downregulation of CYP7A1 and CITED2 would lead to decreased catabolism of cholesterol and fatty acids, respectively. Presumably, the Mg-mediated upregulation of CYP7A1 is to increase the production of bile acids and reduce the level of cholesterol in hepatocytes. The gene product of G-6-P-catalytic subunit (G6PC) is essential to the glucose-6-phosphatase activity catalyzing the conversion of glucose-6-phosphate to glucose; this is the rate-limiting step for gluconeogenesis derived from glycogen breakdown. The effect of G6PC downregulation may promote glycogen storage and development of fatty liver; thus, its effects on plasma lipid changes are likely indirect (9). On the other hand, malic enzyme-1 (ME-1) catalyzes the production of NADPH from malate and NADP⁺; NADPH is essential for both fatty acid and cholesterol synthesis (1, 21). Therefore, ME-1 is considered to be a bona fide lipogenic enzyme in the liver. While RTV treatment led to downregulation of CYP7A1, CITED2, and G6PC, and dramatic upregulation of ME-1, leading to either decreased lipid catabolism or increased lipid biosynthesis, Mg supplementation reversed the downregulation of the first 3 genes, and completely prevented the upregulation of ME-1. In doing so, Mg supplementation completely prevented the RTV-mediated elevations of triglyceride and cholesterol levels in the blood. It may be plausible that magnesium's opposing regulatory effects on lipogenic gene expression are a consequence of its systemic antioxidant actions, as suggested by its attenuating effects on elevated circulating isoprostane and RBC GSSG levels. In agreement with this hypothesis, we observed that catecholamine-induced hyperlipidemia in a hamster model was attenuated by dietary administration of probucol, which is also a potent antioxidant (5). In in vitro studies (30), we found that RTV alone (5–15 μM) caused increased superoxide production in cultured endothelial cells along with increased cellular GSSG levels; both oxidative indices were suppressed by higher extracellular Mg (2 mM), suggesting RTV alone can induce oxidative stress, which can be suppressed by Mg supplementation. However, hypercholesterolemia/hyperlipidemia per se may also contribute to oxidative/nitrosative stress, as suggested by a recent study that indicates that hypercholesterolemia promoted upregulation of NADPH oxidase 4, and this, in turn, led to cardiac dysfunction (48). Further studies are required to provide definitive evidence of whether NADPH oxidase 4 activation/inhibition may be involved in our RTV/Mg supplementation model.

RTV treatment also caused enhanced basal and PMA-stimulated free radical generation from the circulating neutrophils. Enhanced basal activity can be interpreted as partial endogenous activation of the neutrophils, whereas the twofold higher stimulated activity may represent significant upregulation of the NADPH oxidase system during RTV treatment. Since neutrophils are a major cellular source of free radical production, both its basal activation and NADPH oxidase upregulation might contribute significantly to systemic oxidative stress caused by RTV. Previous studies also reported that RTV administration resulted in endothelial toxicity associated with increased superoxide production in cultured cells and in isolated endothelium from RTV-treated animals (11, 49). Consistent with elevated ROS generation from circulating PMNs, and likely from the endothelium, we observed significant elevations in plasma 8-isoprostane levels along with elevated RBC GSSG ratio; RBCs are constantly subjected to oxidative stress derived from the activated endothelium, especially when passing through the microvascular beds. Mg supplementation of RTV-treated rats suppressed both the basal and stimulated activities of the neutrophils. A previous report suggested that Ca²⁺ influx is required for both priming and activation of NADPH oxidase activity in neutrophils (6). In the present study, high dietary Mg resulted in significantly higher (+32%) than normal circulating Mg levels. Since Mg can be a natural “calcium antagonist” (2), we suggest that the anticalcium effect of high circulating Mg may contribute to the suppressed activation of neutrophils from RTV-treated rats. Attenuation of RTV-induced elevations of circulating isoprostane and RBC GSSG levels by Mg supplementation may be the consequence of Mg's systemic antiperoxidative effects, not only on neutrophil ROS generation, but likely on oxidative activation of other ROS-producing tissues (e.g., endothelium).

At the cardiac tissue level, histochemical results revealed that RTV treatment of the rats for 5 wk led to significant elevations in WBC infiltration. These observations are consistent with in vitro findings that several protease inhibitors, including RTV, increased endothelial permeability and leukocyte adhesion (8, 37). Longer RTV treatment (8 wk) induced substantial collagen accumulation consistent with increased perivascular fibrosis. Clinical use of RTV was reported to be associated with increased risk of cardiovascular disease, generally thought to be secondary to lipid-mediated premature atherosclerosis (24, 46). However, the effects of RTV use and the role of hyperlipidemia on cardiac function and structure are unclear. In a clinical report, Meng et al. (36) described that use of HIV protease inhibitors, including RTV, in 98 HIV patients was associated with left ventricular morphological changes and diastolic dysfunction. Two experimental reports demonstrated that hyperlipidemia can exert direct oxidative and nitrosative stress on myocardium, leading to cardiac dysfunction in rats (38) and mice (12). In the current study, substantial hyperlipidemia occurred early (by 1–2 wk), but significant decreases in LV systolic and diastolic function were revealed only at or after 5 wk of RTV treatment, and the progression of dysfunction appeared to coincide with the appearance of WBC infiltration. By 8 wk, systolic function deteriorated much further, and this was associated with prominent ventricular fibrosis. It remains to be determined whether the development of fibrosis might be secondary to delayed hypertension during the later stage of RTV treatment. Similar declining trends were observed for cardiac output, dimensions of the LVPWs, and mitral valve E/A ratio, suggestive of early diastolic dysfunction. RTV-treated rats receiving Mg supplementation displayed significant preservation of LV systolic and diastolic function, normalized cardiac output, and lessened the thinning of LVPWs during the 5–8-wk treatment period. Concomitantly, Mg supplementation completely inhibited the sustained hyperlipidemia; WBC infiltration was largely averted at 5 wk, and the increased collagen buildup was diminished in the hearts of RTV-treated rats at 8 wk. The observed increases in cardiac WBC infiltration along with enhanced systemic oxidative stress during RTV treatment and the inhibitory effects of Mg supplementation permit the speculation that persistent exposure to inflammatory, oxidative, and nitrosative activity during chronic RTV treatment may contribute to the progressive development of LV systolic and diastolic dysfunction.

Perspectives and Significance

Our studies demonstrated the impact of chronic RTV treatment on the development of systemic oxidative stress [ROS/reactive nitrogen species (RNS) increase] and hyperlipidemia; both may lead to cardiac inflammation and dysfunction in the rat (Fig. 7). The beneficial effects of concurrent Mg supplementation were associated with its opposite (to RTV) regulatory effects on key hepatic lipogenic genes. Its ability to downregulate ROS production from inflammatory cells may contribute to its systemic antioxidant and hypolipidemic effects, and may underlie its contribution to the preservation of cardiac function during long-term RTV exposure. Alternatively, Mg supplementation may suppress the induced hyperlipidemia directly and the associated oxidative stress and cardiac dysfunction. Because the use of most lipid-lowering drugs is complicated by drug interactions, as many statins are metabolized by CYP3A4, which is inhibited by most protease inhibitors (17), dietary Mg supplementation might prove to be a useful and economical adjunct therapy to counter the potential lipid metabolic side effects and cardiovascular toxicity associated with long-term use of PIs in HIV-1 therapy.

Fig. 7.

Scheme of antioxidant effects of Mg on hyperlipidemia, ROS/reactive nitrogen species (RNS) stress, and cardiac injury during RTV therapy.

GRANTS

This study was supported by National Institutes of Health Grant R21-NR-012649 to I. T. Mak.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: I.T.M. conception and design of research; I.T.M., J.H.K., X.C., J.J.C., and C.F.S. performed experiments; I.T.M., J.H.K., X.C., and J.J.C. analyzed data; I.T.M., X.C., J.J.C., C.F.S., and W.B.W. interpreted results of experiments; I.T.M., J.H.K., X.C., and J.J.C. prepared figures; I.T.M. drafted manuscript; I.T.M., J.H.K., X.C., J.J.C., C.F.S., and W.B.W. edited and revised manuscript; I.T.M., J.H.K., X.C., J.J.C., C.F.S., and W.B.W. approved final version of manuscript.