Contribution of TLR4 signaling in intermittent hypoxia-mediated atherosclerosis progression

Sixty ApoE^−/− mice on the C57BL/6 background (Male, 8 weeks old) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). Mice were fed on a high-fat diet (HFD) (0.25% cholesterol and 15% cocoa butter) starting at 8 weeks of age.

Three short hairpin RNA (shRNA) sequences targeting mouse TLR4 gene (NM_021297.3) were designed and constructed. After RNAi screening, the most effective shRNA sequences (sense: 5′-GATCC GCACT CTTGA TTGCA GTTTC ATTCA AGAGA TGAAA CTGCA ATCAA GAGTG CTTTT TTG-3′ anti-sense: 5′-AATTC AAAAA AGCAC TCTTG ATTGC AGTTT CATCT CTTGA ATGAA ACTGC AATCA AGAGT GCG-3′), which exhibited 89.5% of TLR4mRNA reduction, was selected for the study. Recombinant shRNA-TLR4 lentivirus (LV-TLR4i) was generated and purified. LV-enhanced green fluorescent protein (LV-EGFP) viral suspension was prepared, which was purchased from Geneche (Shanghai, China).

All ApoE^−/−mice were randomly divided into four groups: Sham (n = 15), IH (n = 15), IH + LV-EGFP (lentivirus-enhanced green fluorescent protein) (n = 15) and IH + LV-TLR4i (TLR4 interference mediated by lentivirus) (n = 15) groups. The animal grouping and timeline of the experimental protocol are presented in Fig. 1. At 20 weeks of age, IH + LV-EGFP and IH + LV-TLR4i group mice were injected respectively with LV-EGFP and LV-TLR4i (2.0 × 10⁷ infection unit per mouse) through the tail vein, while normal saline was injected into the other two groups of mice which served as vehicle controls.

One week after gene transfection, mice were exposed to either Sham or IH. A computer-controlled solenoid valve system was used to control the oxygen concentration by regulating the infusion of nitrogen, room air and oxygen into the exposure chamber. During each cycle of IH, the fractional oxygen concentration of the exposure chamber was reduced to a nadir of 5.5–6.5%, stabilized at that level for 10 s and increased to 20–21% in the subsequent 30 s [9, 10, 16]. The Sham group mice were exposed to a similar environment but only room air was used during the IH exposure time (08:00–17:00 daily) to coincide with the animals’ diurnal sleep period.

Oxygen concentration was measured by oxygen Detector AR8100 (SMART SENSOR, Hong Kong).

At these two time points of Pre- and Post-IH, blood pressure was measured by the tail-cuff method (CODA2, Kent Scientific, Torrington, CT) [19].

At the end of exposure protocols, mice fasted for 8 h before euthanasia. Arterial blood was obtained by direct cardiac puncture under 1% pentobarbital anesthesia. The heart with the aortic root, liver, kidney and the artery from the aortic arch to the left and right common iliac artery were quickly harvested and fixed in 4% paraformaldehyde (pH 7.0–7.4) or frozen in liquid nitrogen. Serial cryostat Sects. (5 μm thickness) of the aortic root were prepared by the freezing microtome (Leica CM1950, Nussloch, Germany), and were used for histological and immunohistochemical staining.

To measure overall load and distribution of AS in the inner wall of the aorta, en face lesion was stained with Oil-Red O. Moreover, the cryostat sections were stained with hematoxylin and eosin (H&E) according to a standard protocol of our laboratory. Oil-Red O staining and Sirius red staining were used to detect the content of lipids and collagen of aortic root plaques, respectively. TLR4, NF-κB p65 (nuclear factor κappa B p65), α-SMA (α-smooth muscle actin), MOMA-2 (macrophages/monocytes) expression in aortic root were detected by immunohistochemical staining. Briefly, cryosections were prepared, hydrated, blocked in blocking solution, and incubated with anti-TLR4 (Abcam, ab120684-1, 1:200), anti-p65 (Abcam, ab7970, 1:250), anti-α-SMA (Abcam, ab5694; 1:200), anti-MOMA-2 (Abcam, ab33451, 1:250) as we previously described [20]. Histological and immunohistochemical staining were analyzed with Image Pro-Plus 6.0. (IPP 6.0, Media Cybernetics, Rockville, MD, USA). Atherosclerotic plaque instability index was calculated according to the standard plaque stabilization score formula: (Oil Red O⁺ area plus MOMA-2⁺ area)/(α-SMA⁺ area plus collagen I⁺ area) [21].

Serum lipid profiles, including serum total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), total triglycerides (TG) and glucose concentrations were measured by enzymatic assay using an automatic biochemical analyzer (Roche Cobas Integra 800, Basel, Switzerland).

Mice serum levels of interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α) were measured using ELISA kits (eBiosciences) according to the instructions.

Human umbilical vein endothelial cells (HUVECs) were purchased from CHI SCIENTIFIC Biotechnology (Jiangyin, China). Cells were cultured in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Gibco Inc.), penicillin (100 U/mL) and streptomycin (100 μg/mL).

Before exposure to hypoxia/reoxygenation (H/R), we first performed cell transfection. Lentivirus expressing either nonsense shRNA or anti-TLR4 shRNA were transfected into the HUVECs. HUVECs were plated at a density of 3 × 10⁵ cells per 35 mm plate and cultured for 12 h before transfection. Then medium containing packaged lentivirus was added and 12 h later was replaced with fresh medium. After 48 h, the interference effect of TLR4 protein was detected by using western blot for each experiment.

Hypoxia was induced in a Whitley H35 Hypoxystation equipment of CO₂/O₂ incubator for hypoxia research. Briefly, cells were cultured in 35 mm dishes at 1% O₂ for 6 h and then reoxygenationed at the time periods of 2, 12 and 24 h, respectively. Control cells were cultured in normoxia in the same incubator and harvested at the specified times.

Proteins were extracted from treated HUVECs and ApoE^−/− mice aortas. All samples were lysed on ice for 30 min in a RIPA buffer (Beyotime, Nantong, China) after being ground with a pestle. Equal amounts of protein were separated by SDS-PAGE and transferred to PVDF membranes. After blocking by 5% non-fat milk, the membranes were incubated with primary antibodies for TLR4 (Abcam, ab120684-1, 1:500), p65 (Abcam, ab7970, 1:1000), interleukin-1β (IL-1β)(Abcam, ab9722, 1:500) and p-p65 (Ser536) (Cell signaling, #3033, 1:1000) overnight at 4 °C. After washing and incubating with an HRP-conjugated secondary antibody (1:10,000), the immunoreactive bands were visualized using chemiluminescence (Millipore Corporation, Billerica, MA, USA). Sample loadings were normalized to β-actin expression.

Total RNA was extracted using RNAiso Plus reagent (TaKaRa Biotech Corporation, Dalian, China) and treated with RNase-free gDNA Eraser to remove traces of genomic DNA. Complementary DNA was generated with a PrimeScript™ RT reagent kit with gDNA Eraser (TaKaRa Biotech Co.). A SYBR^® Premix Ex Taq™ Kit (TaKaRa Biotech Co.) was used for a real-time polymerase chain reaction (PCR) reaction examination. The sequences of primers (5′–3′) were (1) for TLR4: ATGGC ATGGC TTACA CCACC (forward) and GAGGC CAATT TTGTC TCCAC A (reverse); (2) for β-actin: CACTG TGCCC ATCTA CGA (forward) and GTAGT CTGTC AGGTC CCG (reverse). The relative expression of the target gene was normalized against β-actin and the data were analyzed by the 2^−ΔΔCT method [22].

The data are shown as mean ± SEM. Continuous data for two group differences were analyzed by Student’s t test and differences among groups were analyzed by one-way ANOVA with post hoc analysis. A p value < 0.05 was considered statistically significant. Data were analyzed using SPSS 17.0 software (SPSS Inc., Chicago, IL, USA).

A growing number of in vivo studies show that IH exposure triggers multiple proatherogenic factors [9‐12]. However, the impact of IH on plaque growth and its underlying mechanisms remains poorly understood. Thus, we observed the consequence of IH exposure in affecting atherosclerotic plaque growth by using special staining techniques. As expected, general Oil Red O staining showed that arteries, crossing from the aorta arch to the common iliac artery, contained extensive atherosclerotic lesions throughout the artery trunk after 5-weeks of IH exposure (Fig. 2a, b). Moreover, the average size of plaque areas in the IH group was increased 2.5- to threefold than in the Sham group (Fig. 2c, d).

To further explore whether TLR4 signaling was involved in IH-mediated atherosclerotic plaque growth, mice were transfected with TLR4 specific shRNA and then treated with IH exposure. Compared with the IH + LV-EGFP group, the IH + LV-TLR4i group mice showed fewer atherosclerotic lesions (Fig. 2a, b) and smaller plaque areas (Fig. 2c, d). Therefore, these findings showed that TLR4 signaling was involved in IH-triggered atherosclerotic plaque growth.

Finding that IH accelerated atherosclerotic plaque growth in ApoE^−/− mice, we further investigated the effect of IH on atherosclerotic plaque vulnerability. A vulnerable plaque is characterized by a large lipid core covered with a thin fibrous cap, containing sparse smooth muscle cells and extensive macrophages [23]. In the present study, IH exposure increased the content of macrophages and lipids but reduced that of smooth muscle cells and collagen I in the atherosclerotic plaque, thus contributing to increased plaque vulnerability in ApoE^−/− mice (Fig. 3a–d). The plaque vulnerability index in the IH group was increased in several times relative to the Sham group (Fig. 3e). However, mice transfected with TLR4-specific shRNA showed a more stable atherosclerotic plaque (Fig. 3a–e). These changes in the composition of plaques indicated that IH exposure promotes plaque instability and the potential for plaque rupture. In addition, TLR4 interference modified the composition of plaques and improved the stability of vulnerable plaque in IH-exposed ApoE^−/− mice.

TLR4 signaling is closely involved in the activation of the inflammatory response. To investigate whether TLR4 signaling involved in IH-mediated plaque development by aggravating inflammatory activity in vivo, we first examined TLR4 expression in the aortic plaque in IH-exposed ApoE^−/− mice. IH upregulated both protein (Fig. 4a, c) and mRNA (Fig. 4f) levels of TLR4 in the aorta. Meanwhile, the nuclear localization of p65 (Fig. 4b) and the phosphorylation of p65 (Fig. 4d) in the IH group were significantly increased than that in the Sham group. Moreover, IH upregulated the protein level of IL-1β (Fig. 4e) and the infiltration of macrophages (Fig. 3b) in atherosclerotic plaque, as well as increased serum levels of IL-6 and TNF-α in ApoE^−/− mice (Fig. 4g, h).

To explore possible mechanisms by which IH promoted the inflammatory response, we further elaborated the relationship between TLR4 signaling and inflammatory activity in IH-exposed ApoE^−/− mice. Compared with the IH + LV-EGFP group, TLR4 interference evidently reduced TLR4 expression (Fig. 4a, c, f) and activation of p65 (Fig. 4b, d) in the IH + LV-TLR4i group mice aortas. Moreover, inhibition of TLR4 could restore IL-1β, IL-6, and TNF-α to the levels of the Sham group (Fig. 4e, g, h). Likewise, TLR4 interference effectively blocked the increasing of macrophage infiltration in IH-exposed atherosclerotic plaque (Fig. 3b). Consequently, we concluded that IH accelerated plaque progression via TLR4-triggered vascular and systemic inflammatory responses.

The results of body weight and blood pressure of the four groups are presented in Table 1. There are no differences in body weight among the four groups at baseline (V₀). After 5 weeks of IH exposure, IH and IH + LV-EGFP groups were significantly reduced in body weight (V₁) compared with the Sham group (Table), but this effect of IH on body weight was reversed by TLR4 interference, as shown in Table 1.

At baseline, there were no differences in systolic blood pressure (SBP) and diastolic blood pressure (DBP) among the four groups. After IH exposure, SBP and DBP were significantly increased compared with the Sham group. However, there were no significant differences between the IH + LV-TLR4i group and the IH + LV-EGFP group (Table 1), indicating that TLR4 interference had little effect on IH-elevated blood pressure.

As most of the OSA patients are complicated with dyslipidemia, we investigated whether IH had an effect on glycolipid metabolisms. As shown in Fig. 5a–e, IH significantly increased serum levels of total cholesterol, LDL-C, HDL-C, glucose, with no influence on TG levels. TLR4 interference reversed the IH-induced increase of total cholesterol and LDL-C, but had little influence on the serum levels of HDL-C, TG and glucose. Therefore, IH could cause glycolipid metabolism disorders and TLR4 was involved in IH-mediated hypercholesterolemia.

To further prove that H/R exposure could trigger the activation of the proinflammatory TLR4/NF-κB signaling, we applied HUVECs as the carrier of in vitro study. Compared with the other H/R intervals, we discovered that the protein level of TLR4 was significantly higher in the H6/R12 (6 h of hypoxia and 12 h of reoxygenation) group (Fig. 6a). TLR4 expression both at the gene level (Fig. 6b) and protein level (Fig. 6c) was significantly higher in H6/R12 group than in the Control group. Additionally, H6/R12 exposure also promoted the phosphorylation of p65 (Fig. 6d). To further investigate the role of TLR4 in H/R-triggered p65 activation, we added treatment with TLR4-shRNA. We found that TLR4 interference reversed the increase of TLR4 expression (Fig. 6b, c) and the phosphorylation of p65 (Fig. 6d). These data indicated that H/R exposure significantly enhanced the activation of the proinflammatory TLR4/NF-κB signaling in HUVECs, and TLR4 interference mitigated inflammatory-related endothelial dysfunction.