Effect of treatment with human apolipoprotein A-I on atherosclerosis in uremic apolipoprotein-E deficient mice

Atherosclerosis 202 (2009) 372–381

Effect of treatment with human apolipoprotein A-I on atherosclerosis in uremic apolipoprotein-E deficient mice Tanja X. Pedersen a,∗ , Susanne Bro a,b , Mikkel H. Andersen c , Michael Etzerodt c , Matti Jauhiainen d , Søren Moestrup e , Lars B. Nielsen a,f a

Department of Clinical Biochemistry, KB 3011, Rigshospitalet, Blegdamsvej 9, DK-2100, Copenhagen, Denmark b Department of Nephrology, Rigshospitalet, Copenhagen, Denmark c Borean Pharma, Aarhus, Denmark d National Public Health Institute, Helsinki, Finland e Institute of Medical Biochemistry, University of Aarhus, Denmark f Department of Biomedical Science, University of Copenhagen, Denmark

Received 12 October 2007; received in revised form 24 February 2008; accepted 4 April 2008 Available online 1 May 2008

Abstract Objective: Uremia markedly increases the risk of atherosclerosis. Thus, effective anti-atherogenic treatments are needed for uremic patients. This study examined effects of non-lipidated recombinant human apoA-I (h-apoA-I) and a recombinant trimeric apoA-I molecule (TripA-I) on lipid metabolism and atherosclerosis in uremic apoE−/− mice. Methods and results: Upon intraperitoneal injection, h-apoA-I and TripA-I rapidly associated with plasma HDL and reduced mouse apoA-I plasma levels without affecting total or HDL cholesterol concentrations. The plasma half-life was ∼36 h for TripA-I and ∼16 h for h-apoA-I. Injection of h-apoA-I (100 mg/kg) or TripA-I (100 mg/kg) twice weekly for 7 weeks did not affect the cross-sectional area of atherosclerotic lesions in the aortic root, or the en face lesion area and cholesterol content in the thoracic aorta in uremic apoE−/− mice. Also, the treatments did not affect expression of selected inflammatory genes in the thoracic aorta or plasma concentrations of soluble ICAM-1 and VCAM-1. However, h-apoA-I-treated mice had larger smooth muscle cell-staining areas in aortic root plaques than PBS-treated mice (4.8 ± 0.8% vs. 2.5 ± 0.6%, P < 0.05). Conclusions: The data suggest that long-term treatment with non-lipidated h-apoA-I or TripA-I might affect plaque composition but does not reduce atherosclerotic lesion size in uremic apoE−/− mice. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Uremic atherosclerosis; Apolipoprotein A-I; Lipid metabolism

1. Introduction Uremic patients have extremely increased risk of cardiovascular disease, i.e. the annual incidence of cardiovascular events is ∼20% in hemodialysis patients [1]. The underlying causes probably include accelerated development of atherosclerosis. Indeed, uremia in apolipoprotein-E defi∗

Corresponding author. Tel.: +45 35 45 29 34; fax: +45 35 45 25 24. E-mail address: [email protected] (T.X. Pedersen).

0021-9150/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2008.04.041

cient (apoE−/−) mice markedly accelerates formation of atherosclerosis in aorta [2–5]. Hence, aortic atherosclerotic lesions were 10-fold and 6-fold larger in uremic than in non-uremic apoE−/− mice at 12 and 22 weeks after induction of uremia, respectively [2,3]. Disappointingly, however, the recent 4D study examining the effect of statins in uremic diabetes patients only resulted in a marginal 8% reduction of cardiovascular events [6]. Thus, there is an urgent need to develop new treatment strategies to prevent cardiovascular disease in uremic patients.

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381

Since uremic patients often have pre-existing atherosclerosis, a rational therapy in this high-risk group could include therapies that specifically induce lesion regression. Epidemiological studies show an inverse relationship between risk of cardiovascular disease and plasma concentrations of high-density lipoprotein (HDL) cholesterol as well as HDL’s major protein component apolipoprotein A-I (apoA-I) [7–9]. The atheroprotective effect of HDL has been ascribed to the ability of apoA-I to promote cholesterol efflux from macrophages and attenuate inflammation. Thus, multiple studies have shown that apoA-I can mobilize cholesterol from foam cells in vitro [10,11]. Also, increasing the plasma apoA-I concentration in vivo increases reverse cholesterol transport from peripheral cells to the liver, increases neutral sterol excretion in feces in humans [12] and mice [13,14], and decreases expression of adhesion molecules in endothelial cells in rabbits [15,16]. Thus, apoA-I-increasing therapies, e.g. by repeated injection of apoA-I, seem well suited to target pre-existing atherosclerosis. Due to its small size (∼28.5 kDa) lipid-free apoAI is rapidly cleared from the circulation by filtration in the kidneys, where apoA-I is taken up by proximal tubule cells in a megalin and cubilin dependent process [17]. While apoA-I mainly circulates bound to large HDL particles that are not filtered in the kidney, plasma HDL undergoes continuous remodelling [18] and it is conceivable that rapid clearance of small apoA-I containing HDL particles or lipid-free apoA-I contributes to the overall clearance of apoA-I. Hence, displacement of apoA-I from HDL in apoA-II transgenic mice [19] and ABCA-1 deficiency where apoA-I lipidation is impaired result in rapid renal catabolism of apoAI [20]. Attempting to prolong the plasma half-life and thus potentially increase the therapeutic efficiency of injections, Graversen et al. have developed a trimerized version of human apoA-I (TripA-I) [21]. In vitro studies suggest that TripA-I maintains its ability to efflux cholesterol and TripA-I indeed appear to have a slower clearance than normal apoA-I in LDL-receptor-deficient mice [21]. In TripA-I, the trimerization domain from tetranectin has been fused N-terminally to the human apoA-I sequence. Like apoA-I, TripA-I binds ABCA-1 and mediates cholesterol efflux from foam cells. Importantly, TripA-I appear to be superior to apoA-I in mobilizing labelled cholesterol from J774 macrophages. Moreover, TripA-I, like apoA-I, acts as a co-factor in lecithin cholesterol acyltransferase-dependent esterification of cholesterol [21]. To explore whether repeated injections of apoAI protect against the pro-atherogenic effect of uremia, we treated uremic apoE−/− with non-lipidated human recombinant apoA-I or TripA-I and examined the effects on plasma lipoproteins and aortic lesion formation.

373

2. Materials and methods 2.1. Mice Female apoE−/− mice (C57BL/6Jbom-Apoetm1Unc , Taconic M&B laboratory Animals and Services for Research, Ry, Denmark) were 5/6 nephrectomized in two operations as previously described [3]. Briefly, at the age of 10 weeks, both poles of the right kidney were removed and 2 weeks later the entire left kidney was removed. For information regarding mouse care, anesthesia and analgesia please see Supplemental data. The experiments were performed according to the principles stated in the Danish law on animal experiments and were approved by the Animal Experiments Inspectorate, Ministry of Justice, Denmark. 2.2. Formulation of human recombinant apolipoprotein A-I and trimeric apoA-I Both human apoA-I (h-apoA-I) and trimeric apoA-I (TripA-I) were expressed in E. coli at Borean Pharma A/S, phenol extracted, purified on a Ni2+ -NTA-agarose column, digested with coagulation factor Xa, further purified on Q-sepharose or SP-sepharose columns, respectively, gel filtered into 25 mM (NH4 )2 CO3 , pH 8.8, lyophilized, washed with chloroform/methanol to remove endotoxins and E. coli lipids, resuspended in guanidinium-HCl, gel filtered into 25 mM (NH4 )2 CO3, lyophilized, and finally resuspended in PBS buffer, pH 7.4 (non-lipidated versions) or bound to dimyristoyl phosphatidylcholine (DMPC) (lipidated versions) by resuspension in PBS buffer, pH 7.4 at a 2.4:1 lipid to protein ratio (mol:mol) as described by Jonas [22]. Protein concentrations were determined by UV spectroscopy using calculated extinction coefficients (ProtParam) and by the BCA protein assay (Pierce; 23225; Rockford, IL, USA). Protein preparations were tested for endotoxins using the Limulus Amebocyte Lysate (LAL) QCL-1000 kit (Cambrex) and purity was assessed by RP-HPLC, nativeand SDS-PAGE, and size-exclusion chromatography-HPLC methods. 2.3. Experimental studies To determine the plasma clearance of h-apoA-I and TripAI, uremic apoE−/− mice were given a single intraperitoneal (i.p.) injection of lipidated (n = 2) or non-lipidated (n = 1) h-apoA-I (100 mg/kg) or lipidated (n = 2) or non-lipidated (n = 2) TripA-I (100 mg/kg). Blood samples were collected from the tail vein before, 4, 8, 24, 48, and 78 h after injection. To explore the effect of h-apoA-I and TripA-I injections on plasma lipids and lipoproteins, uremic apoE−/− mice were given a single i.p. injection of non-lipidated h-apoA-I (100 mg/kg, n = 6), non-lipidated TripA-I (100 mg/kg, n = 6), or phosphate buffered saline, pH 7.4 (PBS) (n = 5). Blood samples were taken prior to injections (from the retro-orbital

374

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381

venous plexus), after 1 h (from the tail vein), and after 4 h (from the retro-orbital venous plexus). To determine the impact of h-apoA-I and TripA-I injections on progression of atherosclerosis, uremic mice were stratified into three groups 17 weeks after the second operation. Each mouse then received two weekly i.p. injections for a total of 7 weeks of h-apoA-I (100 mg/kg, n = 10), TripA-I (100 mg/kg, n = 11), or PBS (n = 10). Two mice in the TripA-I group were sacrificed due to fights between the mice. At the termination of the study, the mice were anesthesized and perfused with 0.9% NaCl through the left ventricle until the eluate became clear [3] and the relevant organs were isolated as detailed in Supplemental data. 2.4. Western blotting Proteins were separated on 12% polyacrylamide gels (NuPAGE, Novex Bis-Tris, Invitrogen, Taastrup, Denmark) and transferred to Hybond-P 0.45-␮m PVDF membranes using a semi-dry electroblotter. After blocking, membranes were incubated with either biotinylated mouse-anti-human apoA-I IgG, followed by horse-radish-peroxidase (HRP) labelled anti-biotin antibody, goat-anti-human apoA-I followed by HRP-conjugated rabbit-anti-goat antibody, or rabbit-anti-mouse apoA-I antibody followed by HRPconjugated goat-anti-rabbit antibody. For further details, please see Supplemental data. 2.5. Non-denaturing polyacrylamide gel electrophoresis Five ␮g of non-lipidated h-apoA-I and TripA-I were separated in native 4–20% Tris-Glycine gels (EC6025, Invitrogen, Taastrup, Denmark) using Novex Native sample buffer (LC2673, Invitrogen) and native Tris-Glycine running buffer (LC2672, Invitrogen). 2.6. Size-exclusion chromatography To separate proteins according to size, 2 mg non-lipidated h-apoA-I or TripA-I or pooled plasma samples (200–400 ␮L) from 5 to 10 mice were subjected to fast-phase liquid chromatography on a Superose 6 10/300 GL FPLC column (Amersham Pharmacia Biotech, Hoersholm, Denmark) using PBS with Na2 EDTA (0.1 g/L) as running buffer. The column was calibrated with human plasma lipoproteins. The protein concentration in eluted fractions was determined with the BCA protein assay (Pierce; 23225; Rockford, IL, USA). 2.7. Ultracentrifugation The density of pooled plasma samples (20 ␮L) or pure hapoA-I and TripA-I (200 ␮g in 20 ␮L) was adjusted to 1.063 or 1.21 g/mL with KBr and ultracentrifuged at 100,000 rpm for 5 h at 15 ◦ C using a Beckman TLA-100 rotor in an Optima Max-E ultracentrifuge (Ramcon, Birkerod, Denmark). Sub-

sequently, the top and bottom fractions were isolated and fractions hereof were used for Western blotting. 2.8. Plasma biochemistry Blood was collected in heparinized microtubes. For details regarding measurements of specific plasma proteins and lipids, please see Supplemental data. 2.9. RNA extraction and real-time PCR Liver and aorta RNA and cDNA were prepared as described previously [2]. Real-time PCR with a light-cycler (Roche) was used to measure gene expression as described in Supplemental data. Briefly, each reaction mixture contained cDNA synthesized from 20 ng total RNA and standard curves were made by serial dilutions of a pool of liver or aorta samples. 2.10. Histological analyses of aortic root atherosclerosis The formaldehyde-fixed heart with 1–2 mm of the aortic root was embedded in Tissue-Tek (Sakura Finetek Inc., Vaerloese, Denmark) and serial 10-␮m sections of the aortic sinus were cut in a cryostate. With the appearance of the first aortic valve, the sections were collected on SuperFrost Plus slides (Menzel-Glaser, Germany) as detailed in Supplemental data. For determination of the lipid content in the aortic root, 2 microscope slides containing 8 tissue sections were stained with oil-red-O (ORO) (Sigma–Aldrich, O08625) and counterstained with Mayer’s hematoxylin (Sigma–Aldrich, MHS-16). The microscope slides were chosen based on three criteria as defined in Supplemental data. The ORO stained area was quantitated on 6 sections from each mouse using the image analysis software IM50 (Leica, Copenhagen, Denmark) and this value was divided with the total plaque area in the same 6 sections to obtain the relative ORO stained area. For determination of the collagen content and the cytoplasma/cell contents in the aortic root, the microscope slide just proximal to the ones used for ORO staining was stained with Massons Trichrome stain (Sigma–Aldrich, HT-15). Collagen was measured as the brightly blue-stained area and the cytoplasma/cell content as the red area in two tissue sections from each mouse using IM50. These values were divided with the total plaque area in the same two tissue sections to obtain the relative collagen and cytoplasma/cell content, respectively. The total plaque area in the aortic root (as presented in Fig. 4A) was determined as the average plaque area in ORO as well as trichrome stained tissue sections. 2.11. Immunohistochemistry on aortic root sections The DAKO-streptABComplex/HRP kit (Dakocytomation, Glostrup, Denmark) was used according to the

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381

375

manufacturers instruction with the modifications specified in Supplemental data. To stain smooth muscle cells, we applied a biotinylated primary antibody (Actin, smooth Ab (clone 1A4); MS-113B1; Neomarkers, Freemont, CA, USA) and IM50 was used to measure the ␣-smooth muscle actin staining area in 2–4 tissue sections per mouse. This value was divided with the total plaque area in the same 2–4 tissue sections, to obtain the relative smooth muscle cell content. To stain macrophages, we applied the F4/80 antibody (YSRTMCAP497; Accurate Chemical and Scientific Corporation, Westbury, NY, USA). F4/80 staining was scored as 1 (weak), 2 (moderate), or 3 (strong) by duplicate blinded microscopic evaluations of 2–4 tissue sections per mouse. 2.12. Analysis of aortic arch atherosclerosis The aortic arch was opened longitudinally and photographs were taken before the aorta was stored at −20 ◦ C for subsequent lipid extraction. The en face photographs were analyzed with IM50. The relative plaque area was determined by dividing the atherosclerotic area with the total area of the aortic arch. 2.13. Lipid extraction and thin layer chromatography (TLC) Lipids were extracted from the aortic arch as previously described [23] with the modifications described in Supplemental data. Levels of free and esterified cholesterol as well as triglycerides were determined using TLC [24,25] as detailed in Supplemental data.

Fig. 1. Plasma clearance of h-apoA-I and TripA-I. (A) Western blots of plasma with a biotinylated monoclonal antibody against human apoA-I at the indicated time points after i.p. injection of lipidated and non-lipidated (Nonlip.) h-apoA-I and TripA-I in uremic apoE−/− mice. (B) Plasma clearance of h-apoA-I (squares, n = 3) or TripA-I (triangles, n = 4). Plasma apoA-I concentrations were quantitated by Western blotting and data from lipidated and non-lipidated proteins were combined. Values are mean ± S.E.M.

2.14. Statistical analyses Statistical analyses were performed using GraphPad Prism 4 (GraphPad software Inc., San Diego, CA, USA). P < 0.05 was considered significant. 3. Results 3.1. Plasma metabolism of h-apoA-I and TripA-I in uremic apoE−/− mice Graversen et al. reported that the plasma half-life of lipidated TripA-I is longer than that of lipidated apoA-I when injected into LDL-receptor-deficient mice [21]. To assess the plasma clearance in uremic apoE−/− mice, we injected lipidated or non-lipidated preparations of h-apoA-I or TripA-I (100 mg protein/kg body weight intraperitoneally (i.p.)) in apoE−/− mice ∼2 weeks after induction of uremia. The plasma clearance of h-apoA-I and TripA-I was not affected by lipidation (Fig. 1A) and the plasma clearance of h-apoA-I was shorter than that of TripA-I (Fig. 1B). Non-lipidated hapoA-I and TripA-I preparations were used in all subsequent studies.

Due to the fusion of h-apoA-I to the tetranectin trimerization domain, TripA-I is a trimer in solution. Accordingly, TripA-I was larger (∼14 nm) than apoA-I (∼8 nm) as judged by non-denaturing polyacrylamide gel electrophoresis (Fig. 2A) and gel filtration chromatography (Fig. 2B). To determine whether h-apoA-I and TripA-I associated with lipoproteins in vivo, we injected h-apoA-I and TripA-I into uremic apoE−/− mice and separated the plasma lipoproteins by ultracentrifugation and gel filtration chromatography. Four hours after i.p. injection, both h-apoA-I and TripA-I were in the HDL (1.063 < d > 1.21 g/mL) density fraction, whereas the ‘injected material’ was in the d > 1.21 g/mL fraction (Fig. 2C). On gel filtration chromatography, h-apoA-I eluted with HDL-sized particles and TripA-I eluted with both HDL and LDL-sized particles (Fig. 3A). To assess the impact on plasma cholesterol, h-apoAI, TripA-I, or PBS were injected into 6, 6, or 5 uremic apoE−/− mice, respectively. The total plasma cholesterol concentration was not affected 1 and 4 h after i.p. injections (not shown). Also, the distribution of cholesterol between lipoprotein classes was unaffected by h-apoA-I or TripA-I (Fig. 3A). Upon injection of 100 mg/kg h-apoA-I into mice,

376

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381

Fig. 2. Characterization of h-apoA-I and TripA-I. (A) Coomassie-blue-stained gel after non-denaturing gel electrophoresis of h-apoA-I and TripA-I (5 ␮g each). (B) Gel filtration chromatography of non-lipidated h-apoA-I or TripA-I (2 mg each). (C) Western blots of plasma and injected material using biotinylated human apoA-I specific antibodies. Ultracentrifugation of non-lipidated h-apoA-I and TripA-I (injected material) and plasma taken 4 h after injection of h-apoA-I or TripA-I (plasma 4 h) was performed at ds 1.063 g/mL and ds 1.21 g/mL.

the expected human apoA-I plasma concentration at 4 h is similar to that of endogenous mouse apoA-I, i.e. ∼1 mg/mL. Overexpression of human apoA-I in transgenic mice lowers mouse apoA-I plasma concentrations [26–28,29,30]. To examine whether injections of h-apoA-I or TripA-I might have the same effect in uremic apoE−/− mice, we analyzed mouse apoA-I levels in plasma by semi-quantitative Western blotting. Four hours after injection, the mouse apoA-I levels had decreased substantially in mice receiving either h-apoA-

I or TripA-I injection (Fig. 3B). This result was reiterated by ELISA measurements of mouse apoA-I in plasma taken 24 h after injections of h-apoA-I or TripA-I, where plasma levels of mouse apoA-I were 1.25 ± 0.05 mg/mL (n = 10) in PBS-injected, 1.12 ± 0.04 mg/mL (n = 10) in h-apoA-Iinjected, and 1.14 ± 0.02 mg/mL (n = 9) in TripA-I injected mice (P < 0.05 for PBS vs. each of the two treatments). The lowering of plasma mouse apoA-I was not caused by reduced expression of mouse apoA-I mRNA in the livers, which was

Fig. 3. Effects of h-apoA-I and TripA-I on plasma lipoproteins. (A) Cholesterol gel filtration profiles of plasma from uremic apoE−/− mice taken prior to (0 h) or 4 h after (4 h) i.p. injection of PBS, h-apoA-I (100 mg/kg), or TripA-I (100 mg/kg). VLDL, LDL, HDL, and non-lipoprotein (non-LP) fractions were pooled and used for Western blots with antibodies specific for either human apoA-I (h-apoA-I) or mouse apoA-I (mapoA-I). The plasma used for gel filtration (plasma) was included in the Western blots. (B) Western blots showing mouse and human apoA-I in plasma prior to (0 h), 1 (1 h) or 4 h (4 h) after an i.p. injection of PBS, h-apoA-I, or TripA-I.

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381

377

similar in h-apoA-I, TripA-I, and PBS-injected mice (data not shown). 3.2. Effects of long-term treatment with h-apoA-I and TripA-I on plasma lipids in uremic apoE−/− mice In previous studies, the mean plaque area fraction in aortas of uremic apoE−/− mice rose from 4% to 27% between 12 and 22 weeks after induction of uremia [2,3]. To target rapidly expanding atherosclerotic lesions in uremic apoE−/− mice, we initiated treatment with h-apoA-I (100 mg/kg in 200 ␮L PBS i.p., twice weekly), TripA-I (100 mg/kg in 200 ␮L PBS i.p., twice weekly), or PBS (200 ␮L i.p., twice weekly) 17 weeks after induction of uremia and terminated the study 7 weeks later. Induction of uremia by subtotal nephrectomy (5/6 NX) increased plasma urea concentrations to ∼30 mmol/L, i.e., 2–3-fold above the average value in non-uremic apoE−/− mice [2]. Treatment with h-apoA-I and TripA-I did not affect plasma markers of uremia (i.e., creatinine, urea, calcium and phosphate), body weight, or liver enzymes (Supplemental Table 1). Long-term treatment with h-apoA-I and TripA-I did not significantly affect plasma concentrations of total, free or esterified cholesterol or total phospholipids when measured 24 h after the last injection (Table 1). Also, the plasma cholesterol distribution between the different lipoprotein fractions was not affected as judged by gel filtration chromatography profiles (data not shown). Nevertheless, the average plasma concentration of triglycerides was ∼100% higher in h-apoA-I and TripA-I-treated than in PBS-treated mice, although the differences between the mean values were only statistically significant for TripA-I (Table 1). In accordance with this observation, injection of apoA-I has previously been reported to increase plasma triglycerides in humans [31]. 3.3. Effects of long-term treatment with h-apoA-I and TripA-I on atherosclerosis in uremic apoE−/− mice To assess the putative impact of long-term treatment with h-apoA-I and TripA-I on atherosclerosis in uremic apoE−/− mice, we made cross-sections of the aorta at the level of the aortic root and evaluated plaque size and composition. The total plaque area in the aortic root was similar in mice treated with h-apoA-I, TripA-I, and PBS (Fig. 4A). However, the composition of the plaque appeared to be changed by treatment. Even though the relative areas staining for neutral lipids (oil-red-O) and collagen (Trichrome, blue areas) did not differ significantly (Fig. 4B), the combined neutral lipid and collagen staining areas was significantly smaller in TripA-I-treated mice than in the PBS-treated mice (66.8 ± 4.1 vs. 81.6 ± 5.0%; P < 0.05) and also tended to be reduced in the h-apoA-I-treated mice (73.2 ± 5.9%). Accordingly, the relative areas staining for cytoplasma (with Trichrome, red areas) tended to be larger in the h-apoA-I and TripA-I-treated

Fig. 4. Effect of long-term treatment with h-apoA-I and TripA-I on atherosclerotic lesions in the aortic root. (A) Total plaque area in the aortic root of uremic apoE−/− mice treated with PBS (open bars), h-apoA-I (grey bars), or TripA-I (black bars). (B) Relative plaque areas stained for neutral lipids, collagen, and cytoplasma/cells. (C) Relative plaque area staining with an ␣-smooth muscle cell actin specific antibody. P values for two-group comparisons are shown in brackets. n.s.: non-significant. (D) Semi-quantitative assessment of macrophage staining of atherosclerotic lesions. All values are mean ± S.E.M.

groups (Fig. 4B), but again this difference was only borderline statistically significant. To detect which cell types might account for the putative increase in cell content, we performed immunohistochemistry with smooth muscle cell and macrophage specific antibodies. The relative area of ␣-smooth muscle actin-staining cells was increased in hapoA-I-treated compared with PBS-treated mice and did not differ from that in TripA-I-treated mice (Fig. 4C). Scoring of macrophage-stained sections showed no difference between the three treatment groups (Fig. 4D). To further evaluate the putative effects on atherosclerosis, we measured the en face plaque area and lipids in the aortic arch and proximal portion of the thoracic aorta. In accordance with the lack of effect on total plaque areas in cross-sections from the aortic root, the en face lesion area and contents of free and esterified cholesterol as well as triglycerides were similar in h-apoA-I, TripA-I, and PBS-treated mice (Supplemental Fig. 1).

378

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381

Table 1 Effect of h-apoA-I and TripA-I on plasma lipids PBS

h-apoA-I

TripA-I

Before treatment Total cholesterol (mmol/L) Free cholesterol (mmol/L) Esterified cholesterol (mmol/L)

22.1 ± 1.3 6.9 ± 0.4 15.1 ± 0.9

22.1 ± 1.6 6.9 ± 0.5 15.2 ± 1.1

21.1 ± 0.8 6.7 ± 0.3 14.5 ± 0.5

Termination Total cholesterol (mmol/L) Free cholesterol (mmol/L) Esterified cholesterol (mmol/L) Phospholipid (mmol/L) Triglyceride (mmol/L)

17.8 7.3 10.6 4.2 0.8

± ± ± ± ±

1.0 0.4 0.6 0.1 0.1

15.8 6.8 9.0 3.9 1.6

± ± ± ± ±

1.3 0.6 0.7 0.2 0.4

16.1 6.9 9.2 4.3 1.6

± ± ± ± ±

1.0 0.3 0.7 0.1 0.2*

Blood samples at termination were taken 24 h after the last injection. Values are mean ± S.E.M. * P = 0.003 for PBS vs. TripA-I.

3.4. Effects of long-term treatment with h-apoA-I and TripA-I on vascular gene expression in uremic apoE−/− mice Vascular inflammation with increased expression of VCAM-1, ICAM-1 and matrix metalloproteinase (MMP)3 and -12 as well as decreased expression of genes involved with muscle cell function are prominent in aortas from uremic apoE−/− mice [32]. Uremia also increases the plasma levels of soluble (s) VCAM-1 and sICAM-1 in apoE−/− mice [3], further supporting that uremia causes vascular inflammation. To examine whether long-term treatment with h-apoA-I and TripA-I could attenuate these inflammatory effects, we measured plasma levels of sICAM-1 and sVCAM1 as well as mRNA expression in the distal portion of the thoracic aorta. Plasma sVCAM-1 and sICAM-1 were similar in the three treatment groups (Supplemental Fig. 2). Also, neither h-apoA-I nor TripA-I affected the mRNA expression of VCAM-1, ICAM-1, MMP-3, MMP-12, or ␣-smooth muscle cell actin in the distal portion of the thoracic aorta (Supplemental Table 2).

4. Discussion Uremic patients are in particular need of effective treatments to prevent the development and progression of atherosclerosis. Previous studies have raised the possibility that apoA-I based therapies can be used to rapidly diminish atherosclerosis. Therefore, we examined the effects of repeated injections of h-apoA-I and a trimerized apoA-I variant, TripA-I, on atherosclerosis in uremic apoE−/− mice. The main finding of the present study was that two weekly injections of h-apoA-I for 7 weeks failed to affect the size of atherosclerotic plaques in aorta, but appeared to change plaque composition in the aortic root. Treatment with TripA-I had a similar effect despite a ∼3-times longer plasma halflife. Why did treatment with apoA-I fail to reduce lesion formation in uremic apoE−/− mice? Overexpression of a human apoA-I transgene in atherosclerosis-susceptible

mice generally reduces formation of atherosclerosis in aorta [26–30,33–36]. Two studies have tested the effect of repeated apoA-I injections on atherosclerosis in vivo. One saw a ∼50% reduction in lesion formation in cholesterol-fed rabbits [37]. The other saw no effect on lesion formation in the aorta of cholic acid/cholesterol-fed mice [38]. In contrast to the relative paucity of studies using apoA-I injections, several studies have addressed the usefulness of apoA-IMilano -injections to reduce atherosclerosis. ApoA-IMilano is a naturally occurring human apoA-I variant with a single amino acid substitution of arginine in position 173 with cysteine [39]. The cysteine substitution results in formation of apoA-I dimers, and in cell cultures apoA-IMilano appears more effective in mediating cholesterol efflux from foam cells than native apoA-I [40]. Two studies in rabbits, two in mice, and one in humans all found that injection(s) of apoA-IMilano have anti-atherogenic effects (e.g. reductions of intima thickness, macrophage staining, and surface lesion and cross-sectional oil-red-O-staining areas) [41–45]. We chose to examine the effects of non-lipidated hapoA-I and TripA-I because non-lipidated h-apoA-I and TripA-I rapidly associated with plasma lipoproteins in uremic apoE−/− mice in vivo. In accordance with these observations, lipid-free apoA-I was rapidly lipidated when incubated with endothelial cells or injected into rabbits [46] and the plasma metabolism of lipidated and non-lipidated apoA-I was similar upon injection into rabbits [47]. Moreover, the phospholipid component of the apoA-I/phospholipid complexes may have anti-atherogenic effects on its own [15,41]. Nevertheless, Lee et al. reported that 99% of lipid-free apoA-I rapidly associates with medium-sized (8.6 nm) HDL particles whereas only 37% of lipidated pre-␤ HDL remodelled to mature HDL upon intravenous injections [48]. The latter observation may imply that lipidated apoA-I affect reverse cholesterol transport and atherosclerosis differently than non-lipidated apoA-I. ABCA-1 mediated formation of phospholipid and cholesterol-containing pre-␤ HDL particles occurs in at least two steps, where the phospholipids are transferred onto apoA-I before cholesterol enrichment [10]. Hence, in theory, the rapid association of non-lipidated apoAI with plasma HDL may bypass formation of cholesterol

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381

efflux promoting phospholipid-containing apoA-I particles. As such, the ability of injected lipid-free apoA-I to promote cholesterol efflux may be smaller than that of phospholipidated apoA-I. It is also possible that the pro-atherogenic effects of uremia simply override beneficial effects of apoA-I treatment that would be seen in less aggressive classical atherosclerosis. Hence, we cannot exclude that apoA-I treatment had been more effective if initiated earlier when the aggressive uremic atherosclerosis was less advanced or if the treatment period had been longer. Also, uremia has multiple effects on HDL metabolism [49,50] and as such, may dampen effects of apoA-I that are anti-atherogenic in the non-uremic state. Indeed, A. Fogelman and others have delineated that HDL can be converted from protective anti-inflammatory particles to dysfunctional pro-inflammatory particles that may even be pro-atherogenic [51]. The anti-inflammatory effects associated with apoA-I injections in non-uremic animals is supported by two studies. Thus, three injections of 25 mg non-lipidated apoA-I into rabbits 24 h before, immediately before and 24 h after placement of a collar around the carotid artery, respectively, resulted in markedly reduced endothelial expression of adhesion molecules and infiltration of neutrophils [15]. Furthermore, a single intravenous infusion of 1, 2 or 8 mg/kg lipid-free apoA-I 0, 3 or 9 h after insertion of a non-occlusive collar around the carotid artery in normocholesterolemic rabbits resulted in a dose-dependent decrease in neutrophil infiltration and activation [16]. Uremia markedly increases the expression of inflammatory genes in aortas of apoE−/− mice [32]. However, the aortic mRNA expression of selected inflammatory genes and plasma concentrations of sVCAM-1 and sICAM-1 were similar in the h-apoA-I, TripA-I, and PBS groups further supporting the conclusion that h-apoA-I and TripA-I had little effect on endothelial inflammatory status that would further affect progression of atherogenesis in uremic mice. We used rather high doses of h-apoA-I and TripAI (100 mg/kg) compared to previous in vivo studies and observed that mouse apoA-I plasma concentrations decreased temporarily upon injection of h-apoA-I and TripA-I, although the liver expression of mouse apoA-I mRNA was unchanged. Lowering of mouse apoA-I also occur in human apoA-I transgenic mice [26,27], probably reflecting displacement of mouse apoA-I from the HDL particles. Thus, it cannot be excluded that replacement of endogenous mouse apoAI with h-apoA-I and TripA-I in mouse HDL reduced their endogenous anti-atherogenic potential. However, this idea needs further examination in future studies. Although there was no effect on total lesion areas in the aorta, treatment with h-apoA-I increased the relative content of smooth muscle cells in aortic root plaques. This effect may be potentially beneficial since smooth muscle cell enrichment leads to a more stable, less rupture prone plaque. In accord with the present finding, overexpression of human apoA-I also increases plaque smooth muscle cell content in mice [35]. Importantly, a potential plaque-stabilising effect

379

in the aortic root was also seen in LDL-receptor-deficient mice treated with lipidated h-apoA-I or TripA-I [21]. Moreover, treatment with apoA-I Milano –phospholipid complexes reduces intimal thickening after balloon injury in rabbits [41] and in-stent stenosis in pigs [52]. In conclusion, the present data show that long-term treatment with non-lipidated h-apoA-I and TripA-I does not reduce atherosclerotic lesion size in uremic apoE−/− mice, but that h-apoA-I might affect the composition of the plaques towards a more stable phenotype.

Acknowledgements Tina Estrup Axen, Kirsten Hansen and Karen Rasmussen provided excellent technical assistance. This work was supported by ‘The Danish Medical Research Council’, ‘The Danish Heart Foundation’ and ‘Fru Ruth E K¨onig-Petersens forskningsfond’.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.atherosclerosis. 2008.04.041.

References [1] Foley RN, Herzog CA, Collins AJ. Smoking and cardiovascular outcomes in dialysis patients, the United States Renal Data System Wave 2 study. Kidney Int 2003;63:1462–7. [2] Bro S, Bentzon JF, Falk E, et al. Chronic renal failure accelerates atherogenesis in apolipoprotein-E-deficient mice. J Am Soc Nephrol 2003;14:2466–74. [3] Bro S, Moeller F, Andersen CB, Olgaard K, Nielsen LB. Increased expression of adhesion molecules in uremic atherosclerosis in apolipoprotein-E-deficient mice. J Am Soc Nephrol 2004;15:1495–503. [4] Buzello M, Tornig J, Faulhaber J, et al. The apolipoprotein-E knockout mouse, a model documenting accelerated atherogenesis in uremia. J Am Soc Nephrol 2003;14:311–6. [5] Massy ZA, Ivanovski O, Nguyen-Khoa T, et al. Uremia accelerates both atherosclerosis and arterial calcification in apolipoprotein-E knockout mice. J Am Soc Nephrol 2005;16:109–16. [6] Wanner C, Krane V, Marz W, et al. Atorvastatin in patients with type 2 diabetes mellitus undergoing hemodialysis. N Engl J Med 2005;353:238–48. [7] Stampfer MJ, Sacks FM, Salvini S, Willett WC, Hennekens CH. A prospective study of cholesterol, apolipoproteins, and the risk of myocardial infarction. N Engl J Med 1991;325:373–81. [8] Buring JE, O’Connor GT, Goldhaber SZ, et al. Decreased HDL2 and HDL3 cholesterol, Apo A-I and Apo A-II, and increased risk of myocardial infarction. Circulation 1992;85:22–9. [9] Rader DJ. High-density lipoproteins as an emerging therapeutic target for atherosclerosis. JAMA 2003;290:2322–4. [10] Rye KA, Barter PJ. Formation and metabolism of prebeta-migrating, lipid-poor apolipoprotein A-I. Arterioscler Thromb Vasc Biol 2004;24:421–8. [11] Ohashi R, Mu H, Wang X, Yao Q, Chen C. Reverse cholesterol transport and cholesterol efflux in atherosclerosis. QJM 2005;98:845–56.

380

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381

[12] Eriksson M, Carlson LA, Miettinen TA, Angelin B. Stimulation of fecal steroid excretion after infusion of recombinant proapolipoprotein A-I, potential reverse cholesterol transport in humans. Circulation 1999;100:594–8. [13] Zhang Y, Zanotti I, Reilly MP, et al. Overexpression of apolipoprotein A-I promotes reverse transport of cholesterol from macrophages to feces in vivo. Circulation 2003;108:661–3. [14] Moore RE, Navab M, Millar JS, et al. Increased atherosclerosis in mice lacking apolipoprotein A-I attributable to both impaired reverse cholesterol transport and increased inflammation. Circ Res 2005;97: 763–71. [15] Nicholls SJ, Dusting GJ, Cutri B, et al. Reconstituted high-density lipoproteins inhibit the acute pro-oxidant and proinflammatory vascular changes induced by a periarterial collar in normocholesterolemic rabbits. Circulation 2005;111:1543–50. [16] Puranik R, Bao S, Nobecourt E, et al. Low dose apolipoprotein AI rescues carotid arteries from inflammation in vivo. Atherosclerosis 2008;196:240–7. [17] Kozyraki R, Fyfe J, Kristiansen M, et al. The intrinsic factor-vitamin B12 receptor, cubilin, is a high-affinity apolipoprotein A-I receptor facilitating endocytosis of high-density lipoprotein. Nat Med 1999;5:656–61. [18] Rye KA, Clay MA, Barter PJ. Remodelling of high density lipoproteins by plasma factors. Atherosclerosis 1999;145:227–38. [19] Julve J, Escol´a-Gil JC, Ribas V, et al. Mechanisms of HDL deficiency in mice overexpressing human apoA-II. J Lipid Res 2002;43: 1734–42. [20] Lee JY, Timmins JM, Mulva A, et al. HDLs in apoA-I transgenic ABCA-I knockout mice are remodeled normally in plasma but are hypercatabolized by the kidney. J Lipid Res 2005;46: 2233–45. [21] Graversen JH, Laurberg JM, Andersen MH, et al. Trimerization of apolipoprotein A-I retards plasma clearance and preserves antiatherosclerotic properties. J Cardiovasc Pharmacol 2008 [E-pub ahead of print]. [22] Jonas A. Reconstitution of high-density lipoproteins. Methods Enzymol 1986;128:553–82. [23] Bartels ED, Lauritsen M, Nielsen LB. Hepatic expression of microsomal triglyceride transfer protein and in vivo secretion of triglyceride-rich lipoproteins are increased in obese diabetic mice. Diabetes 2002;51:1233–9. [24] Ruiz JI, Ochoa B. Quantification in the subnanomolar range of phospholipids and neutral lipids by monodimensional thin-layer chromatography and image analysis. J Lipid Res 1997;38:1482–9. [25] Christoffersen C, Bollano E, Lindegaard ML, et al. Cardiac lipid accumulation associated with diastolic dysfunction in obese mice. Endocrinology 2003;144:3483–90. [26] Rubin EM, Krauss RM, Spangler EA, Verstuyft JG, Clift SM. Inhibition of early atherogenesis in transgenic mice by human apolipoprotein AI. Nature 1991;353:265–7. [27] Rubin EM, Ishida BY, Clift SM, Krauss RM. Expression of human apolipoprotein A-I in transgenic mice results in reduced plasma levels of murine apolipoprotein A-I and the appearance of two new high density lipoprotein size subclasses. Proc Natl Acad Sci USA 1991;88: 434–8. [28] Paszty C, Maeda N, Verstuyft J, Rubin EM. Apolipoprotein AI transgene corrects apolipoprotein-E deficiency-induced atherosclerosis in mice. J Clin Invest 1994;94:899–903. [29] Belalcazar LM, Merched A, Carr B, et al. Long-term stable expression of human apolipoprotein A-I mediated by helper-dependent adenovirus gene transfer inhibits atherosclerosis progression and remodels atherosclerotic plaques in a mouse model of familial hypercholesterolemia. Circulation 2003;107:2726–32. [30] Choudhury RP, Rong JX, Trogan E, et al. High-density lipoproteins retard the progression of atherosclerosis and favorably remodel lesions without suppressing indices of inflammation or oxidation. Arterioscler Thromb Vasc Biol 2004;24:1904–9.

[31] Nanjee MN, Doran JE, Lerch PG, Miller NE. Acute effects of intravenous infusion of ApoA1/phosphatidylcholine discs on plasma lipoproteins in humans. Arterioscler Thromb Vasc Biol 1999;19:979–89. [32] Bro S, Borup R, Andersen CB, et al. Uremia-specific effects in the arterial media during development of uremic atherosclerosis in apolipoprotein-E-deficient mice. Arterioscler Thromb Vasc Biol 2006;26:570–5. [33] Plump AS, Scott CJ, Breslow JL. Human apolipoprotein A-I gene expression increases high density lipoprotein and suppresses atherosclerosis in the apolipoprotein-E-deficient mouse. Proc Natl Acad Sci USA 1994;91:9607–11. [34] Tangirala RK, Tsukamoto K, Chun SH, et al. Regression of atherosclerosis induced by liver-directed gene transfer of apolipoprotein A-I in mice. Circulation 1999;100:1816–22. [35] Rong JX, Li J, Reis ED, et al. Elevating high-density lipoprotein cholesterol in apolipoprotein-E-deficient mice remodels advanced atherosclerotic lesions by decreasing macrophage and increasing smooth muscle cell content. Circulation 2001;104: 2447–52. [36] Valenta DT, Bulgrien JJ, Banka CL, Curtiss LK. Overexpression of human ApoAI transgene provides long-term atheroprotection in LDL receptor-deficient mice. Atherosclerosis 2006;189:255–63. [37] Miyazaki A, Sakuma S, Morikawa W, et al. Intravenous injection of rabbit apolipoprotein A-I inhibits the progression of atherosclerosis in cholesterol-fed rabbits. Arterioscler Thromb Vasc Biol 1995;15:1882–8. [38] Garber DW, Datta G, Chaddha M, et al. A new synthetic class A amphipathic peptide analogue protects mice from diet-induced atherosclerosis. J Lipid Res 2001;42:545–52. [39] Kaul S, Shah PK. ApoA-I Milano/phospholipid complexes emerging pharmacological strategies and medications for the prevention of atherosclerotic plaque progression. Curr Drug Targets Cardiovasc Haematol Disord 2005;5:471–9. [40] Franceschini G, Calabresi L, Chiesa G, et al. Increased cholesterol efflux potential of sera from ApoA-IMilano carriers and transgenic mice. Arterioscler Thromb Vasc Biol 1999;19:1257–62. [41] Ameli S, Hultgardh-Nilsson A, Cercek B, et al. Recombinant apolipoprotein A-I Milano reduces intimal thickening after balloon injury in hypercholesterolemic rabbits. Circulation 1994;90:1935–41. [42] Shah PK, Nilsson J, Kaul S, et al. Effects of recombinant apolipoprotein A-I(Milano) on aortic atherosclerosis in apolipoprotein-E-deficient mice. Circulation 1998;97:780–5. [43] Shah PK, Yano J, Reyes O, et al. High-dose recombinant apolipoprotein A-I(milano) mobilizes tissue cholesterol and rapidly reduces plaque lipid and macrophage content in apolipoprotein-E-deficient mice, potential implications for acute plaque stabilization. Circulation 2001;103:3047–50. [44] Chiesa G, Monteggia E, Marchesi M, et al. Recombinant apolipoprotein A-I(Milano) infusion into rabbit carotid artery rapidly removes lipid from fatty streaks. Circ Res 2002;90:974–80. [45] Nissen SE, Tsunoda T, Tuzcu EM, et al. Effect of recombinant ApoAI Milano on coronary atherosclerosis in patients with acute coronary syndromes, a randomized controlled trial. JAMA 2003;290:2292–300. [46] Rohrer L, Cavelier C, Fuchs S, et al. Binding, internalization and transport of apolipoprotein A-I by vascular endothelial cells. Biochim Biophys Acta 1761;2006:186–94. [47] Kee P, Rye KA, Taylor JL, Barrett PH, Barter PJ. Metabolism of apoA-I as lipid-free protein or as component of discoidal and spherical reconstituted HDLs, studies in wild-type and hepatic lipase transgenic rabbits. Arterioscler Thromb Vasc Biol 2002;22:1912–7. [48] Lee JY, Lanningham-Foster L, Boudyguina EY, et al. Prebeta high density lipoprotein has two metabolic fates in human apolipoprotein A-I transgenic mice. J Lipid Res 2004;45:716–28. [49] Vaziri ND. Dyslipidemia of chronic renal failure, the nature, mechanisms, and potential consequences. Am J Physiol Renal Physiol 2006;290:F262–72.

T.X. Pedersen et al. / Atherosclerosis 202 (2009) 372–381 [50] Bang CA, Bro S, Bartels ED, Pedersen TX, Nielsen LB. Effect of uremia on HDL composition, vascular inflammation, and atherosclerosis in wild-type mice. Am J Physiol Renal Physiol 2007 [E-pub ahead of print]. [51] Ansell BJ, Fonarow GC, Fogelman AM. The paradox of dysfunctional high-density lipoprotein. Curr Opin Lipidol 2007;18:427–34.

381

[52] Kaul S, Rukshin V, Santos R, et al. Intramural delivery of recombinant apolipoprotein A-IMilano/phospholipid complex (ETC-216) inhibits in-stent stenosis in porcine coronary arteries. Circulation 2003;107:2551–4.