Page 1 of 33

Accepted Preprint first posted on 23 July 2014 as Manuscript JME-14-0049

1

oxLDL induces injury and defenestration of human liver sinusoidal endothelial

2

cells via LOX-1

3

Qi Zhang1, Jing Liu2*, Jia Liu2, Wenhui Huang2, Limin Tian2, Jinxing Quan2, Yunfang Wang2,

4

Ruilan Niu1

5

1.The First Clinical College of Lanzhou University, Lanzhou City 730000, Gansu Province, China

6

2.Department of Endocrinology, Gansu Provincial Hospital, 204 West Donggang Road, Lanzhou City 730000, Gansu Province, China

7 8

Corresponding author’s postal and email address：

9

Department of Endocrinology, Gansu Provincial Hospital, 204 West Dong Gang Road, Lanzhou, 730000, China; [email protected].

10 11

Short title: oxLDL, LOX-1, defenestration, HLSECs, NAFLD, T2DM

12

Key words Oxidised low-density lipoprotein, lectin-like oxLDL receptor-1,

13

endothelial injury, defenestration, human liver sinusoidal endothelial cells

14

Word count of the full text: 4564

15

The first two authors contributed equally to this study.

16

* Corresponding author

17

【Abstract】 】Non-alcoholic fatty liver disease (NAFLD) is associated with hepatic

18

microangiopathy and liver inflammation caused by type 2 diabetes mellitus (T2DM).

19

Oxidised low-density lipoprotein (oxLDL) is involved in proinflammatory and

20

cytotoxic events in various microcirculatory systems. The lectin-like oxLDL

21

receptor-1 (LOX-1) plays a crucial role in oxLDL induced pathological

22

transformation. However, the underlying mechanism of oxLDL’s effects on liver 1

Copyright © 2014 by the Society for Endocrinology.

Page 2 of 33

23

microcirculation disturbances remains unclear. In this study, we investigated the

24

effects of oxLDL on LOX-1 expression and function, as well as on the fenestration

25

features of human liver sinusoidal endothelial cells (HLSECs) in vitro. Primary

26

HLSECs were obtained and cultured. The cells were treated with various

27

concentrations of oxLDL (25, 50, 100, and 200 µg/mL), and the cytotoxicity and

28

expression of LOX-1 were examined. Furthermore, LOX-1 knockdown was

29

performed using siRNA technology, and the changes in intracellular reactive oxygen

30

species (ROS), NF-κB (p65), endothelin-1 (ET-1), eNOS, and caveolin-1 levels were

31

measured. Cells were treated with 100 µg/mL oxLDL, and the fenestra morphology

32

was visualised using scanning electron microscopy. oxLDL significantly increased

33

LOX-1 expression at both the mRNA and protein levels in HLSECs in a dose- and

34

time-dependent manner. oxLDL stimulation increased ROS generation and NF-κB

35

activation, upregulated ET-1 and caveolin-1 expression, downregulated eNOS

36

expression and reduced the fenestra diameter and porosity. All of these

37

oxLDL-mediated effects were inhibited after LOX-1 knockdown. These data reveal a

38

mechanism by which oxLDL stimulates the production of LOX-1 through the

39

ROS/NF-κB signalling pathway and by which LOX-1-mediates oxLDL-induced

40

endothelial injury and the defenestration of HLSECs.

41

Introduction

42

Type 2 diabetes mellitus (T2DM) is potentially complicated by hepatic

43

microangiopathy and liver inflammation (Hudacko et al.2009) and is an independent

44

risk factor for the formation and development of non-alcoholic fatty liver disease 2

Page 3 of 33

45

(NAFLD)(Krishnan et al. 2013). Metabolic syndrome plays a major role in the

46

pathogenesis of both NAFLD and T2DM (El-Serag et al. 2004). NAFLD is

47

commonly observed in adults and adolescents with T2DM (Lv et al. 2013).

48

Cross-sectional studies have revealed that the prevalence of NAFLD in patients with

49

T2DM ranges from 42.1 to 75.2% in China (Zhou et al. 2007). However, the

50

cause–effect relationship between these conditions is complex and difficult to

51

decipher.

52

NAFLD is associated with hepatic sinusoid microcirculation (Farrell et al. 2008).

53

Liver sinusoidal endothelial cells (LSECs) play important roles in regulating hepatic

54

sinus function. LSECs line the capillaries of the microvasculature and possess

55

fenestrae that facilitate filtration between the liver parenchyma and sinusoids by

56

serving as a selectively permeable barrier (Braet et al. 1995,2002). Studies have

57

suggested that NAFLD is associated with dysfunction of LSECs (Braet F,2003).

58

Sinusoidal endothelial fenestrae (SEF), which represent complete pores in LSECs,

59

play important roles in liver physiology and disease. Alterations in LSEC fenestration

60

have implications for liver function, including lipoprotein metabolism (Le Couteur et

61

al.2002). Preclinical studies have demonstrated that LSECs undergo defenestration as

62

an early event that precedes liver fibrosis (Marcos et al.2012), and endothelin-1 (ET-1)

63

and caveolin-1 (Cav-1) induce defenestration in LSECs (Yokomori et al. 2006, 2009).

64

Oxidised low-density lipoprotein (oxLDL) is formed in large quantities in the

65

circulating blood of diabetic patients (Holvoet et al. 2008) and is a crucial risk factor

66

for atherosclerosis (AS) and diabetic microangiopathy (Witztum et al. 2001, 3

Page 4 of 33

67

Lopes-Virella et al.1999, Pirillo et al. 2013). oxLDL can lead to oxidative stress and

68

an inflammatory response (Zmijewski et al. 2005). oxLDL has also been reported to

69

facilitate the formation of more oxLDL and other pro-inflammatory cytokines through

70

a vicious cycle mediated by a specific receptor known as the lectin-like oxidised

71

low-density lipoprotein receptor-1 (LOX-1) (Mehta et al. 2006). LOX-1 is one of the

72

major scavenger receptors (SRs) responsible for binding, internalising, and degrading

73

oxLDL (Sawamura et al. 1997). LOX-1 activation is associated with many

74

pathophysiological events, including endothelial cell and vascular smooth muscle cell

75

proliferation, alterations in cell cycle signals, apoptosis, and autophagy (Yoshimoto et

76

al.2011, Zabirnyk et al.2010). Accumulating evidence indicates that oxLDL binding to

77

LOX-1 increases intracellular reactive oxygen species (ROS), activates the NF-κB

78

signal pathway, decreases intracellular NO, and increases endothelin-1 (ET-1), which

79

could lead to endothelial dysfunction (Cominacini et al. 2000, Morawietz et al. 2002,

80

Fleming et al. 2005).

81

oxLDL is involved in liver microcirculation disturbance and NAFLD (Oteiza et al.

82

2011,Walenbergh et al. 2013). oxLDL can activate hepatic Kupffer and stellate cells,

83

trigger inflammation and fibrogenesis, and induce sinusoidal endothelial dysfunction

84

(Dorman et al. 2006, Li et al. 2011.). Atherogenic blood-borne oxLDL is primarily

85

removed by cells lining the liver sinusoids (McCuskey, 2006). LSECs play an

86

important role in the blood clearance of oxLDL (Li et al. 2011). Various SRs can bind

87

and/or mediate endocytosis of oxLDL in LSECs (Steinberg et al. 1997). We

88

previously demonstrated that LOX-1 is present in LSECs. However, the role of 4

Page 5 of 33

89

LOX-1 in oxLDL-mediated LSEC dysfunction has not been specifically investigated.

90

In addition, whether changes in the fenestration of LSECs are the result of oxLDL

91

binding to LOX-1 is largely unknown.

92

In the present study, we explored models of oxLDL-induced injury and LOX-1

93

gene knockdown in HLSECs to investigate the effects of oxLDL on endothelium

94

function and the fenestration characteristics of HLSECs in vitro.

95

Materials and Methods

96

Reagents

97

DMEM and foetal bovine serum (FBS) were purchased from Gibco-BRL (Gibco, NY,

98

USA). The NO assay kit was purchased from Nanjing Jiancheng Bioengineering

99

Institute (China). The primer pairs for GAPDH, LOX-1, ET-1, endothelial nitric oxide

100

synthase (eNOS), and caveolin-1 were synthesised by TaKaRa Biotechnology Co.,

101

Ltd. (Dalian, China). TRIzol reagent, RNase inhibitor, LOX-1 siRNA vector, XfectTM

102

transfection reagent, and the RT-PCR kit were purchased from TaKaRa Bio, Inc.

103

(Dalian, China). Rabbit antibodies to LOX-1, CD31, vWF, ET-1, eNOS, caveolin-1,

104

NF-κB, and GAPDH and goat anti-rabbit HRP-conjugated secondary antibodies were

105

purchased from Abcam (Hong Kong) Ltd. 2’,7’-dichlorofluorescein diacetate

106

(DCFH-DA) was obtained from Invitrogen (Carlsbad, CA, USA). OxLDL was

107

purchased from Peking Union-Biology Co., Ltd (Beijing, China). All other chemicals

108

of analytical grade were purchased from commercial suppliers.

109

Primary human LSEC isolation, culture, and identification Human liver tissue was obtained from a single 53-year-old Chinese woman

110 5

Page 6 of 33

111

undergoing resection for liver cysts under sterile conditions. Informed consent was

112

obtained from the patient, and the present study was approved by the ethics research

113

committee of The First Clinical College of Lanzhou University. HLSECs were

114

isolated from 25 g of human liver tissue as described previously (Daneker et al.1998).

115

HLSECs were cultured in DMEM supplemented with 10% (v/v) FBS (Gibco, NY,

116

USA), penicillin, streptomycin and glutamine (100 U/ml, 100 mg/mL, 292 mg/ml,

117

respectively), as well as hepatocyte growth factor (HGF) and vascular endothelial

118

growth factor (VEGF) (both 10 ng/ml) under standard cell culture conditions

119

(humidified atmosphere with 5% CO2 at 37°C). Cells were identified using an

120

anti-CD31 antibody via immunocytochemistry and characteristic morphology in

121

culture. Cells were cultured to confluence and were used in all experiments.

122

Experimental protocol

123

First, to investigate the effects of oxLDL on LOX-1 expression in HLSECs, the

124

cells were treated with oxLDL (25,50, 100, or 200 µg/ml) for 12, 24, or 48 h. Second,

125

to further examine the role of the LOX-1 in oxLDL-induced functional injury of

126

HLSECs, LOX-1 expression was silenced. Cells were successfully transfected with

127

LOX-1 siRNA or pretreated with 250 µg/ml polyinosinic cytidylic acid (PCA, a

128

LOX-1 inhibitor). The cells were divided into 4 experimental groups (the control,

129

LOX-1 siRNA, scrambled siRNA (control siRNA) and LOX-1 inhibitor groups) and

130

exposed to oxLDL (100 µg/ml) for 24 h. The levels of LOX-1, ROS, p65, eNOS, ET-1,

131

Cav-1, and GAPDH were then determined; the characteristics of fenestration were

132

simultaneously detected. 6

Page 7 of 33

133

Cell viability measurement

134

Cells were seeded at a density of 5 × 104 cells/mL in 96-well plates with various

135

concentrations (25, 50, 100, 200 µg/ml) of oxLDL for 6, 12, 24, or 48 h, and the cell

136

viability was measured using the MTT assay. Briefly, the culture supernatant was

137

removed at the indicated time points after the treatment, and the cells were washed

138

with PBS and incubated with MTT (5 mg/mL) in the culture medium at 37°C for an

139

additional 3 h. After MTT removal, the coloured formazan was dissolved in 100 µL of

140

DMSO. The absorption values were measured at 490 nm using a Thermo Scientific

141

Microplate Reader (Multiskan Spectrum, USA). The viability of HLSECs in each

142

well is presented as the percentage of control cells.

143

Designing the LOX-1-siRNA sequence and constructing the LOX-1 silencing

144

plasmid

145

For LOX-1 knockdown, LOX-1 siRNAs were designed based on human LOX-1

146

cDNA. The mRNA sequences of human LOX-1 were retrieved from GenBank (ID:

147

NM_002543.3). The interference sequences were designed using the siRNA Design

148

Support

149

5’-CCCTTGCTCGGAAGCTGAATGAGAA-3’ and

target

150

5’-TTCTCATTCAGCTTCCGAGCAAGGG-3’.

Two

151

oligonucleotides (a top strand and a bottom strand) were synthesised (Table 1). Each

152

oligonucleotide was composed of a target sequence, the hairpin loop TTCAAGAGA,

153

the terminator sequence TTTTTT, and restriction sites (BamHI/EcoRI) added at both

154

ends of the sequences. The complementary oligonucleotides were annealed, and the 7

System

R5:

target

sense

sequence,

antisense

sequence,

complementary

Page 8 of 33

155

pSIREN-RetroQ-DsRed-Express vector was linearised by BamHI/EcoRI enzyme

156

cleavage.

157

pSIREN-RetroQ-DsRed-Express using T4 DNA ligase. The connected product

158

(pSIREN-RetroQ-DsRed-Express-LOX-1 shRNA) was transformed into JM109

159

competent cells (Code No. D9052) using the heat shock method. Subsequently,

160

positive clones were screened, and plasmids were extracted for PCR identification and

161

sequencing identification. Similarly, a control plasmid was constructed by

162

randomising the shRNA sequence (5'-GGAGAGCATCTCATTCGATGCAGAC-3').

163

Cell transfection

The

annealed

oligonucleotide

was

ligated

into

164

Cells were seeded in 6-well plates. When confluent, cells were transfected with

165

pSIREN-RetroQ-DsRed-Express-LOX-1 shRNA (7.5 µg) or control plasmid

166

(scrambled siRNA) using XfectTM transfection reagent (TaKaRa) according to the

167

manufacturer's instructions. The transfection efficiency was monitored by

168

fluorescence intensity and LOX-1 expression levels. A transfection rate above 70%

169

was used for the other experiments. The cell phenotype and characteristics were

170

evaluated after successful transfection.

171

Immunofluorescence (IF)

172

The cells were grown in 4-well chamber slides, fixed, quenched, blocked and

173

incubated with antibody against vWF (1:500, Abcam), or CD31 (1:500, Abcam)

174

overnight at 4 °C. Fluorescently-tagged secondary antibodies were used at 1:1000.

175

Nuclear counterstaining was performed with DAPI. Cells were mounted and imaged

176

by confocal microscopy (LEXT OLS4000). 8

Page 9 of 33

177

Measurement of intracellular reactive oxygen species

178

Intracellular ROS was measured with using the fluorescent signal H2DCF-DA.

179

Cells cultured in 6-well plates were incubated with 100 µg/ml oxLDL. Next, the cells

180

were harvested after reaching the target time and incubated with 10 µmol/L

181

DCFH-DA for 30 minutes at 37°C. DCFH-DA is converted by intracellular esterases

182

to DCFH, which is oxidised into the high fluorescent dichlorofluorescein (DCF) in the

183

presence of a proper oxidant. HLSECs were washed with ice-cold PBS three times,

184

and placed in the dark. Subsequently, the DCF fluorescence was measured with flow

185

cytometry at 488 nm and 525 nm. The results were analysed using the Cell Quest

186

software.

187

Quantitative RT-PCR (qRT-PCR)

188

Total RNA was extracted with TRIzol following the manufacturer’s instructions.

189

RNA concentrations were determined using a spectrophotometer (Beckman

190

Instruments, California, USA). Approximately 2 µg RNA was reverse transcribed to

191

cDNA using the Prime Script®RT reagent kit. Quantitative RT-PCR assays were

192

performed using the LightCycler Real-Time PCR System (Roche 480, Switzerland).

193

Samples were denatured at 95°C for 30 s followed by 40 PCR cycles of 95°C for 5 s,

194

60°C for 30 s, and 72°C for 60 s. A melting curve was used to confirm the formation

195

of the intended PCR products. The results are expressed as the fold difference relative

196

to the level of GAPDH using the 2-∆∆CT method. Each reaction was performed in

197

triplicate. The LOX-1 (NM_002543.3) primers were as follows: forward,

198

5’-AGCAAATGGAACTTCA CCACCAG-3’; and reserve, 5’-CTTGGCATCCAAAG 9

Page 10 of 33

199

ACAAGCAC-3’. GAPDH (NM_002046.2) was used as an endogenous control with

200

the following primers: forward, 5’-CCACCCATGGCAAATTCCATGGCA-3’; and

201

reverse, 5’-TCTAGACGGCAGGTCAGGTCCACC-3’.

202

Western blot analysis

203

Cells were lysed for 30 min at 4°C in lysis buffer. The extracted total protein

204

concentrations were measured using the bicinchoninic acid (BCA) reagent. Equal

205

amounts of lysate proteins were loaded and separated by SDS-PAGE and transferred

206

to polyinylidene fluoride (PVDF) membranes. After incubation in a blocking solution

207

(5% non-fat milk, Sigma), the membranes were immunoblotted with primary

208

anti-NF-kB (phos-p65) (1:1500), anti-LOX-1 (1:2000, Abcam), anti-eNOS (1:2000),

209

anti-ET-1 (1:2000), anti-caveolin-1 (1:2000), or anti-GAPDH (1:2500) overnight at

210

4°C. Membranes were washed and then incubated in a 1:3000 dilution of the specific

211

secondary antibodies for 1.5 h at room temperature, and the membranes were washed.

212

Following the enhanced chemiluminescence development, the exposed film was later

213

developed. The films were scanned with Image Quant 350, and the relative densities

214

of protein bands were analysed with Image J software. The density of each protein

215

band was normalised to that of GAPDH.

216

Scanning electron microscopy

217

LSECs cultured on glass coverslips were fixed in 1.2% glutaraldehyde buffered

218

with 1 M cacodylate buffer (pH 7.4) for 1 h at 4°C and post-fixed with 1% osmium

219

tetroxide in cacodylate buffer (pH 7.4) for 1 h at 4°C. After dehydration in a graded

220

series of ethanol solutions, the cultured cells were dried in a critical point apparatus 10

Page 11 of 33

221

(HCR-2, Hitachi, Japan) and coated with gold in a vacuum coating unit. The fenestrae

222

were observed for each experimental variable under a scanning electron microscope

223

(SU8010, Hitachi, Japan) at a 15-kV acceleration voltage and ×20000 magnification.

224

Statistical analysis

225

All data are expressed as the mean ± standard deviation from at least three

226

independent experiments each performed in triplicate. All data were analysed by

227

one-way ANOVA followed by Bonferroni comparison using SPSS 17.0. A P-value
100 µg/ml had noticeable cytotoxic effects. 11

Page 12 of 33

243

Furthermore, exposure to 200 µg/mL oxLDL for 12 h caused a significant (23.3%)

244

decrease in cell viability; increased exposure for 48 hours decreased the viability of

245

HLSECs (by 42% total from time 0).

246

oxLDL increased LOX-1 expression in HLSECs

247

To determine whether LOX-1 is involved in oxLDL-induced inflammatory

248

responses in HLSECs, we examined the effect of oxLDL on LOX-1 expression.

249

HLSECs were treated with 25, 50, 100, and 200 µg/mL oxLDL for 6, 12, 24, or 48 h.

250

Both qRT-PCR (Fig. 2B) and Western blotting (Fig. 2C) data revealed that LOX-1

251

mRNA and protein expression gradually increased in response to oxLDL. Hence,

252

oxLDL induced LOX-1 expression in a time- and dose-dependent manner. LOX-1

253

expression peaked at 100 µg/mL and 24 h (3.1-fold versus control, P < 0.05);

254

therefore, this concentration was used for all subsequent experiments.

255

The transfected HLSECs maintain specific markers and limited fenestrae

256

To confirm endothelial phenotype and Characterisation of transfected HLSECs,

257

the classic EC markers, von Willebrand factor (vWF) and CD31 were analyzed by

258

using IF staining and nuclear counterstaining, and fenestrae were visualised using

259

scanning electron microscopy. Compared with the isolated HLSECs, we found that

260

transfected cells maintained similar expression of vWF and CD31 (Fig. 3A and B),

261

and maintained limited transcytoplasmic fenestrations (Fig. 3C, arrows).

262

LOX-1 siRNA successfully suppressed LOX-1 mRNA and protein expression

263

To ensure the efficiency of LOX-1 gene silencing, we designed 3 pairs of siRNA

264

for LOX-1 cDNA. After transfection, we screened for the most efficacious siRNA for 12

Page 13 of 33

265

LOX-1 silencing by using RT-PCR and Western blotting. As shown in Fig. 3D, red

266

fluorescence intensity was clearly increased relative to the control, indicating that

267

the HLSECs had successfully been transfected with the LOX-1 siRNA. The LOX-1

268

mRNA and protein expression were significantly decreased compared with the

269

positive control (100 µg/mL oxLDL, P < 0.05), and the control plasmid had no effect.

270

The inhibition efficiency of LOX-1 siRNA on mRNA and protein expression was 77

271

and 74%, respectively (Fig. 3E and F). Moreover, PCA, which is a LOX-1 inhibitor,

272

imparted a weak reduction in LOX-1 expression compared with LOX-1 siRNA

273

treatment; the inhibition efficiency of PCA on LOX-1 mRNA and protein expression

274

was 21 and 32%, respectively (Fig. 3E and F).

275

oxLDL stimulated ROS formation and NF-κB activation in HLSECs through

276

LOX-1

277

To confirm that LOX-1 is involved in oxLDL-induced ROS/NF-κB activation,

278

we investigated the influence of oxLDL (100 µg/ml) on ROS formation and NF-κB

279

activation in HLSECs under control conditions, LOX-1 siRNA transfection, and

280

LOX-1 inhibitor treatment for 24 h. As demonstrated in Fig. 4A, oxLDL markedly

281

enhanced ROS formation and activated p65 in HLSECs (Fig. 4B and C). Interestingly,

282

LOX-1 silencing significantly attenuated oxLDL-induced ROS formation (Fig. 4A)

283

and inhibited p65 expression and NF-κB activation (Fig. 4B and C) in HLSECs. PCA

284

had similar but milder effects (Fig. 4A, B, and C). These results indicate that oxLDL

285

induced marked oxidative stress and that LOX-1 mediated the activation of the

286

ROS/NF-κB pathway in oxLDL-treated HLSECs. 13

Page 14 of 33

287

oxLDL stimulation reciprocally decreases eNOS via LOX-1

288

We investigated the influence of oxLDL on HLSEC eNOS expression by using

289

Western blotting. As shown in Figure 4D, eNOS expression at 24 h was decreased by

290

0.55-fold after exposure to 100 µg/mL oxLDL in HLSECs compared with the control.

291

Similarly, LOX-1 silencing and LOX-1 inhibition elevated oxLDL-induced eNOS

292

expression by 45.5 and 16%, respectively (Fig. 4D). Taken together, these results

293

indicate that oxLDL downregulates eNOS expression via LOX-1 in cultured HLSECs.

294

oxLDL upregulates ET-1 and caveolin-1 via LOX-1 in HLSECs

295

To ascertain whether LOX-1 is involved in the effect of oxLDL on ET-1 and

296

caveolin-1 (Cav-1), we examined the effects of oxLDL on ET-1 and Cav-1 expression

297

in HLSECs. Cells were pretreated with oxLDL (100 µg/mL) for 24 h. oxLDL

298

exposure notably increased ET-1 and Cav-1 levels in HLSECs versus control (p