- Email: [email protected]
Increased expression of TFPI in human carotid stenosis
Thrombosis Research 155 (2017) 31–37
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Full Length Article
Increased expression of TFPI in human carotid stenosis Benedicte Stavik a,b,⁎, Sverre Holm a,h,1, Sandra Espada a,b,e,1, Nina Iversen c, Bjørnar Sporsheim i, Vigdis Bjerkeli a,f, Tuva Børresdatter Dahl a, Per Morten Sandset a,b,f, Mona Skjelland d, Terje Espevik i, Grethe Skretting a,b, Bente Halvorsen a,f,g a
Research Institute of Internal Medicine, Oslo University Hospital Rikshospitalet, Oslo, Norway Department of Haematology, Oslo University Hospital Rikshospitalet, Oslo, Norway Department of Medical Genetics, Oslo University Hospital Rikshospitalet, Oslo, Norway d Department of Neurology, Oslo University Hospital Rikshospitalet, Oslo, Norway e Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway f Institute of Clinical Medicine, University of Oslo, Oslo, Norway g K.G. Jebsen Inflammatory Research Centre, University of Oslo, Oslo, Norway h Hospital for Rheumatic Diseases, Lillehammer, Norway i Centre of Molecular Inflammation Research, Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway b c
a r t i c l e
i n f o
Article history: Received 25 January 2017 Received in revised form 4 April 2017 Accepted 25 April 2017 Available online 28 April 2017 Keywords: Atherothrombosis Tissue factor pathway inhibitor Carotid plaque Anticoagulation Vascular biology
a b s t r a c t Introduction: Tissue factor (TF) pathway inhibitor (TFPI) is the physiological inhibitor of TF induced blood coagulation and two isoforms exists, TFPIα and TFPIβ. In atherosclerotic plaques, TFPI may inhibit TF activity and thrombus formation, which is the main cause of ischemic stroke in carotid artery disease. We aimed to identify the isoforms of TFPI present in human carotid plaques and potential sources of TFPI. Materials and methods: Human atherosclerotic plaques from carotid endarterectomies were used for mRNA and immunohistochemistry analyses. hPBMCs isolated from buffy coats and THP-1 cells were differentiated and polarized into M1 or M2 macrophages, and subsequently cultured with or without cholesterol crystals (CC). mRNA and protein expression were measured with qRT-PCR and ELISA, respectively, and procoagulant activity was assessed using a two-stage chromogenic assay. Results: TFPIα and TFPIβ mRNA levels were significantly increased in carotid plaques, whereas TF levels were unchanged as compared to healthy arteries. Antibodies against total TFPI showed elevated levels compared to antibodies against free TFPIα, both by immunohistochemical and ELISA detection in plaques. The antibody against total TFPI also co-localized with CD68 and the M1 and M2 markers CD80 and CD163, respectively. The TFPI mRNA expression was elevated and the procoagulant activity was decreased in M2 compared to M1 polarized human macrophages. TFPI was present in early foam cell formation and CC treatment increased the TFPI mRNA expression even further in M2 macrophages. Conclusions: Our data indicate that both isoforms of TFPI are present in advanced plaques and that anti-inflammatory M2 macrophages may be a potential source of TFPI. © 2017 Elsevier Ltd. All rights reserved.
1. Introduction Atherosclerosis is the main precursor to carotid artery disease (CAD), which is the leading cause of ischemic stroke. It is a progressive state where influx and retention of cholesterol-containing low density lipoproteins (LDLs) in the intima of arteries trigger the recruitment of monocytes, resulting in a low grade chronic inflammation and plaque formation [1,2]. Advanced plaques can be classified as stable or unstable. ⁎ Corresponding author at: Oslo University Hospital Rikshospitalet, Research Institute of Internal Medicine, BOX 4950 Nydalen, 0424 Oslo, Norway. E-mail address: [email protected] (B. Stavik). 1 Equal contribution.
http://dx.doi.org/10.1016/j.thromres.2017.04.024 0049-3848/© 2017 Elsevier Ltd. All rights reserved.
Stable plaques are recognized by a thick fibrous cap, high degree of calcification, and substantial smooth muscle cell (SMC) proliferation. They are usually asymptomatic, however, they can attenuate the blood flow if growing into the artery lumen, thereby causing severe stenosis and ischemia. Unstable plaques typically have a thin fibrous cap, a lipid-rich necrotic core, substantial inflammation and macrophage infiltration, and neovascularization, and may cause ischemic stroke through stenosis due to plaque growth or rupture-induced thrombus formation [3,4]. Lately, the presence of different macrophage subtypes in atherosclerotic plaques has been extensively studied. While the classically activated and pro-inflammatory M1 macrophages are mainly found in the rupture prone shoulder area and in the necrotic core of the plaque, the alternatively activated anti-inflammatory M2 macrophages are
32
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
mainly located in the adventitia [5]. Consistent with this, the levels of M1 macrophage markers correlate with lesion progression and are associated with an unstable plaque phenotype [6,7]. Upon plaque rupture, cells that express tissue factor (TF) will be exposed to the blood leading to the formation of a thrombus. TF pathway inhibitor (TFPI) is the physiological inhibitor of TF induced thrombus formation and thus an essential regulator of the coagulation system. It is mainly synthesized by endothelial cells (ECs), but in atherosclerotic plaques it has also been located to macrophages, medial and neointimal SMCs, and T-cells [8,9]. Two isoforms of TFPI are produced by alternative splicing, i.e., TFPIα and TFPIβ, and they both possess anti-coagulant activity [10]. TFPIα is the full length isoform containing three Kunitz type inhibitor domains and a polar C-terminal end. It is mainly cell associated, but some is also secreted from the cells. In plasma, only a small portion of the secreted TFPIα circulates freely as a full length molecule. The remaining is found truncated at the C-terminus and/or complexed with lipoproteins, such as LDL [11]. TFPIβ is a shorter isoform lacking the third Kunitz domain and has a glycosylphosphatidylinositol attachment sequence in its C-terminal end exclusively associating it to the cell surface. The majority of previous clinical studies showed increased circulating TFPI levels in plasma of patients with atherothrombotic disease [12– 15], and also in atherosclerotic vessels compared to normal vessels [8]. Moreover, studies have demonstrated co-localization of TFPIα with TF in plaques where it is believed to play an important role in attenuating TF activity [16–18]. In fact, TFPI deficiency increased vessel occlusion time and atherothrombosis when plaque rupture was induced in atherosclerotic mice [19,20]. Interestingly, TFPI levels also affected the atherosclerotic burden in such mice [19,21]. Thus, it has been suggested that increased TFPI levels in atherosclerotic disease may represent a compensatory mechanism for elevated procoagulant activity and play a protective role not only in atherothrombosis, but also in atherogenesis [22]. In this study, we aimed to identify which TFPI isoforms are present in human carotid plaques, and to identify potential cellular sources of TFPI that may contribute to the protective effect of TFPI in the development of atherosclerotic disease.
normalizing the Ct values against the endogenous control phosphomannomutase 1 (PMM1) or β-actin, and using untreated cells or control arteries (n = 10) as a control reference. 2.4. ELISA and Immunostaining Free and total TFPI ELISA kits (Asserachrom®, Diagnostica Stago, Asnière, France) were used to measure the TFPI antigen in 55 plaque lysates and in plasma samples from 63 carotid stenosis patients and 15 matched controls, according to the manufacturer's protocols. To adjust for the amount of vessel wall lysed, these results were corrected against total protein that was measured in lysates by the modified Lowry assay (Bio-Rad DC Protein Assay, Bio-Rad Laboratories, Hercules, CA, USA). The free TFPI kit measures only free TFPIα as the antibody detects residues 161–240 of TFPI, while the total TFPI kit measures both TFPIβ and free, truncated and lipoprotein-bound forms of TFPIα, as the antibody detects residues 1–160. Detailed protocols for the immunohistochemistry and immunofluorescence methods are described in the supplementary data. The TFPIα specific antibody used for immunostaining detects residues 292–304 of TFPI that are specific for TFPIα, while the total TFPI antibody detects residues 22–87 that are common for TFPIα and TFPIβ. 2.5. Procoagulant activity
2. Materials and methods
The procoagulant activity of the cells was measured as previously described [25]. In short, polarized macrophages in 24-well plates were washed and incubated for 1 h at 37 °C with 10 nM FVIIa (Novo Nordisk, Bagsvaerd, Denmark) and 175 nM FX (Aniara Diagnostica, Mason, OH, USA) before 50 μL of supernatant from the intact cells were transferred to stop solution on ice, loaded on a 96-well microtiter plate and incubated with CS-11(22) substrate (Aniara Diagnostica). The absorbance at 405 nm was recorded at 37 °C for 45 min at 15 s intervals using a Spectra Max Plus 384 microtiter plate reader (Molecular Devices, Sunnyvale, CA, USA). The maximum velocities (Vmax) in mU/min were used to calculate the amount of FXa generated, using a standard curve obtained with known concentrations of FXa (Aniara Diagnostica).
2.1. Carotid endarterectomy specimens
2.6. Statistics
Carotid plaques from patients with internal carotid stenosis (≥70%) were collected during endarterectomy (Oslo University Hospital Rikshospitalet). A detailed description of the tissue sampling and preparation is given in the supplementary data. Patients were classified as asymptomatic or symptomatic according to the absence or presence of the clinical symptoms stroke, transient ischemic attack (TIA) or amaurosis fugax within the past 2 months or 3–6 months prior to surgery. Ultrasound plaque appearance in terms of echogenicity was classified according to consensus criteria [23,24]. Carotid stenosis and echolucency were diagnosed and classified by precerebral color Duplex ultrasound and CT angiography. A detailed description of the patients' characteristics is provided in the supplementary material (Table S2).
The distribution of the data was assessed before statistical differences between the diseased or treated samples and controls were calculated using the unpaired Student's t/Mann Whitney tests, or the oneway ANOVA/Kruskal-Wallis tests (Bonferroni corrected). Correlations were analyzed using the Pearson or Spearman's tests in GraphPad Prism 5.0 (GraphPad, San Diego, CA, USA). Odds ratios were calculated using logistic regression in SPSS (version 22.0; SPSS Inc., Chicago, IL, USA). n denotes the total number of biological replicates from several independent experiments, and a p-value of b0.05 was considered statistically significant. * = p b 0.05, ** = p b 0.01, and *** = p b 0.001.
2.2. hPBMCs and THP-1 cells
The protocols were approved by the Regional Committee for Medical and Health Research Ethics - South East Norway, approval no S-0923a 2009/6065. All studies were conducted according to the declaration of Helsinki, and signed informed consent for participation in the study was obtained from all individuals.
A detailed description of the isolation and culturing of hPBMCs and THP-1 cells is given in the supplementary data.
2.7. Ethical approval
2.3. Quantitative real-time PCR (qRT-PCR) 3. Results RNA isolation and cDNA preparation are described in the supplementary data. Gene expression was examined by qRT-PCR on the ABI PRISM 7900 Sequence Detection System (Thermo Fisher Scientific, Waltham, MA, USA) according to the protocol, using sequence specific assays (Table S1). All samples were run in triplicates. The relative mRNA expression was calculated using the comparative Ct method,
3.1. Expression of TFPI isoforms in human carotid plaques Human carotid plaque samples from 68 subjects were analyzed for mRNA expression of TFPI and TF. The total TFPI mRNA levels were 2-fold higher (p b 0.0001) in plaques than in control arteries
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
(Fig. S1). To distinguish between the two splice variants of TFPI found in humans, we used isoform specific assays detecting only TFPIα or TFPIβ. The mean relative mRNA expression of TFPIα was increased 2.2-fold (p b 0.0001), while the TFPIβ mRNA expression was increased 1.8-fold (p b 0.0013) compared to control arteries (Fig. 1A and B). In the plaques, the average Ct values of the two isoforms were 28.6 and 36.1, respectively. The TF mRNA expression in the plaques did not differ significantly from the control arteries (Fig. 1C), and the results showed no significant differences in the TFPI and TF mRNA levels according to symptomatology (Fig. S2). The gender of the patients had no significant influence on the TFPI or TF mRNA levels (Table S3). To detect TFPI protein in the plaques, we stained the human carotid plaque sections with antibodies specific for free TFPIα or total TFPI. Low staining was observed with the free TFPIα specific antibody (Fig. 2A, upper left), however, the antibody against total TFPI stained plaque sections extensively (Fig. 2A, upper right). Substantial staining was also observed with antibody against TF (Fig. 2A, lower right) in the same area of the plaque as TFPIα (Fig. 2A, lower right vs. left). Next, we measured the TFPI protein levels in 55 human carotid plaque samples using ELISA kits detecting different forms of the TFPI protein. The free TFPI kit detects only free TFPIα, while the total TFPI kit detects both free and lipoprotein-bound TFPIα, and also TFPIβ. The median (range) free and total TFPI antigen levels in the plaque lysates were 2.75 (1.36–5.26) and 3.98 (2.44–7.88) ng/ mg cell protein, respectively (Fig. 2B). When we classified the samples according to the clinical status of the patients, no significant differences between the subgroups were detected (data not shown). We then looked at the frequency of patients in the subgroups with TFPI levels above the median (Q2). Among patients experiencing symptoms in the last two months and in echolucent plaques, there
33
Fig. 2. TFPI protein was detected in human carotid plaques. (A) Immunohistochemical staining of TFPIα (upper and lower left), total TFPI (upper right), and TF (lower right) counterstained with hematoxylin in atherosclerotic carotid plaques. 100/400× magnification. Images are representative of four independent stainings. (B-D) Free and total TFPI antigen levels were measured in human carotid plaques homogenates with ELISA kits. Results were corrected against total protein measurements and expressed as ng TFPI/mg homogenate protein, and divided according to the clinical status of the patients. Box plot with median (n = 55) ± 10–90 percentile, bar graphs of occurrence of quartiles (Q) in percentage.
was an increased OR for high total TFPI levels compared to asymptomatic (OR 1.17, 95% CI 0.61–7.30) and echolucent (OR 2.11, 95% CI 0.61–7.3) plaques (Fig. 2C and Table S5). Oppositely, a reduced OR for high free TFPI levels was seen in plaques from patients experiencing symptoms in the last two months (OR 0.44 95% CI 0.11–1.76) and in echolucent plaques (OR 0.53 95% CI 0.15–2.73) compared to asymptomatic and echogenic plaques (Fig. 2D and Table S5). We also detected a small, but non-significant increase in the free TFPI levels in plasma from patients with carotid stenosis compared to healthy controls (Fig. S3), with no significant differences between asymptomatic and symptomatic patients (data not shown). The gender of the patients had no significant influence on the free TFPI levels (Table S3). 3.2. Plaque macrophages expressed TFPI
Fig. 1. TFPIα and TFPIβ mRNA levels were increased in human carotid plaques. mRNA expression of TFPIα (A), TFPIβ (B), and TF (C) in human carotid plaques and healthy control arteries analyzed using qRT-PCR. Lines show means (n ≥ 9; healthy controls, n ≥ 67; plaques) ± SD. Control arteries were obtained from the common iliac artery of deceased organ donors.
To identify the potential cellular sources of TFPI in the plaques, we first correlated TFPI mRNA expression levels with the mRNA expression levels of different markers for immune cells (CD45 and CD3g), macrophages (CD68) and lipid filled cells (ADRP). No correlation was found between the mRNA levels of TFPI and the B cell markers CD45 and CD3g (data not shown), however, both TFPIα and TFPIβ mRNA levels correlated significantly with CD68 and ADRP mRNA levels. The same correlations were observed between TF and CD68 and ADRP mRNA levels (Fig. S4). To verify that TFPI was expressed in macrophages, plaque sections were double stained with antibodies against total TFPI and CD68. The result showed that the staining of TFPI and CD68 overlapped, as seen on overlay images (Fig. 3A). TF, on the other hand, costained with CD68 in some, but not all CD68 positive cells (Fig. 3B).
34
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
Fig. 3. TFPI and TF are expressed in macrophages in human carotid plaques. Co-immunofluorescence using antibodies against total TFPI (A, C and D), or TF (B), and CD68 (A and B), CD80 (C) or CD163 (D). Sections were counterstained with DAPI. 400× magnification. Images are representative of three independent stainings. Scale bar = 25 μm.
To further investigate the expression of TFPI in plaque macrophage subtypes, we correlated the TFPI and TF mRNA levels with markers of M1 and M2. TFPIα levels, but not TFPIβ levels correlated significantly with both CCR7 and CD163, markers of M1 and M2 polarized macrophages, respectively, while the TF mRNA levels only correlated with CD163 mRNA levels (Fig. S5). To verify this at the protein level, we double stained plaque sections for TFPI and CCR7, CD80, another marker of M1 macrophages, and CD163. The results showed that TFPI co-localized with both M1 (data not shown and Fig. 3C) and M2 macrophages (Fig. 3D). 3.3. TFPI and TF expression in polarized macrophages Next, we investigated whether the TFPI expression was more prominent in either of the macrophage subtypes by differentiating and polarizing human monocytic cells into M1 and M2 in vitro. A specific and distinct macrophage polarization was obtained as M1 macrophages showed a 9- and 1.7-fold increase in TNF-α and IL-6 mRNA levels, respectively, compared to M2 macrophages, while M2 macrophages showed a 2.3-, 4-, and 33-fold increase in PPAR-γ, IL-1ra, and CD163 mRNA levels, respectively, compared to M1 macrophages (Table S4). Polarization of the human monocyte-derived macrophages (MDMs) resulted in a 7.7-fold increase in basal TFPI mRNA levels in M2 macrophages, compared to M1 macrophages (Fig. 4A). This result was also observed in polarized THP-1 cells (Fig. S6A-B). Moreover, basal TF mRNA levels were also significantly increased 2.4-fold in M2 compared to M1 macrophages (Fig. 4B and Fig. S6C).
3.4. Procoagulant activity in polarized macrophages To investigate if the increased expression of TFPI in the M2 macrophages also could affect the procoagulant activity of the cells, FXa activity of the polarized macrophages was investigated using a two-stage chromogenic substrate assay where human FVIIa and FX were added to the cells. TF present on the cell surface initiated the reaction by binding to FVIIa, which then activated FX to FXa. When we analyzed the procoagulant activity on the surface of polarized MDMs, a 2-fold increase in FXa activity in M1 polarized macrophages was obtained compared to the M2 polarized macrophages (Fig. 4C). In the THP-1 cells, a 5-fold increase in FXa activity in M1 polarized macrophages compared to M2 polarized macrophages was seen (Fig. 4D).
3.5. Cholesterol crystals induce TFPI expression We next investigated the expression of total TFPI in sections of carotid plaques and found that TFPI was strongly expressed in cells that resembled an early phase of foam cell development close to the lumen of the plaque (Fig. 5A, right). In more central parts of the plaque, imprints of CC, shown as CC-clefts, started to appear inside foam cell-like structures that were weakly positive for TFPI (Fig. 5A, lower left). To further explore this potential link between CCs and TFPI, we treated polarized MDMs with CCs and measured the TFPI mRNA expression. CC treatment increased both TFPIα and
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
Fig. 4. Basal TFPI and TF mRNA levels and procoagulant activity in polarized macrophages. (A–C) Human monocytes were isolated from buffy coats, differentiated into macrophages using GM- or M-CSF (50–100 ng/mL) and polarized into M1 or M2 subtypes with LPS (100 ng/mL) plus INFγ (20 ng/mL) or IL-4 (20 ng/mL), respectively. Resting MDMs were only incubated with conditional medium for 6 days and denoted M0. mRNA levels of total TFPI (A) and TF (B) were measured using qRT-PCR. Mean levels (n = 9) + SD of three independent experiments are show. (C–D) FXa generation was measured using a two-stage chromogenic substrate assay, where FVIIa and FX were added to the cells. After 1 h incubation at 37 °C, the reaction was stopped and the substrate (CS-11(22)) added. The maximum velocities (Vmax) in mU/min were used to calculate the amount of FXa generated, using a standard curve obtained with known concentrations of FXa in μg/L. (C) Human monocyte-derived macrophages (mean values (n = 6) + SD of two independent experiments are shown) and (D) THP-1 macrophages (monocytes treated with PMA for 6 h prior to M1 and M2 polarization using LPS (10 ng/mL) plus INFγ (5 ng/mL) or IL-4 and IL-13 (25 ng/mL), respectively, mean values (n = 15) + SD of three independent experiments are shown).
TFPIβ mRNA levels approx. 2-fold, but only in M2 polarized macrophages (Fig. 5B and C). No differences in the uptake of CC were observed in the cells (data not shown). Similar results were also obtained following treatment with oxLDL or VLDL (data not shown). In contrast, the TF mRNA levels were significantly decreased by 63% in M2 polarized macrophages after CC treatment (Fig. 5D).
4. Discussion Stroke is the second leading cause of death worldwide and N85% of these events are ischemic. Rupture of unstable atherosclerotic plaques results in exposure of TF expressing cells to the blood and thrombus formation, which is the main cause of CAD induced stenosis and stroke. TFPI is the natural inhibitor of TF induced thrombus formation and both TFPIα and TFPIβ isoforms possess anti-coagulant activity [10]. In the current study, we detected elevated levels of both TFPIα and TFPIβ mRNA in human carotid plaques, however, the differences in Ct values between the two isoforms indicated that TFPIα was more prominent. The TFPIα specific antibodies used in the IHC analyzes and ELISAs showed the presence of free full-length TFPIα in the plaques. Moreover, total TFPI was detected both with IHC and ELISA, and indicate also the presence of truncated and/or lipoprotein associated TFPIα. It is possible that the total TFPI antibody also detected TFPIβ although the mRNA levels indicated that very little TFPIβ was produced in the plaques.
35
The total TFPI protein levels in the plaques averaged 4 ng/mg homogenate protein, which is in line with what others have reported [26,27]. The TFPI protein levels in the plaques were not strongly associated with the patient clinical status, although there was a trend towards higher total TFPI levels in the recently symptomatic and lipid filled unstable plaques than in the asymptomatic and calcified plaques (data not shown). This is consistent with previous data showing elevated levels of total TFPI in plasma of patients with symptomatic vascular disease [14]. Although the ORs for high TFPI levels within the clinical subpopulations were not significant, they showed an interesting difference between total and free TFPI. While the ORs for high total TFPI levels were increased in the recently symptomatic and lipid filled unstable plaques, they were reduced for high free TFPI levels. As TFPI levels are known to affect atherothrombosis in humans [28], these findings might indicate that free TFPI levels may better reflect the symptomatic status of the patients and plaque stability than the total TFPI levels. In atherosclerotic lesions, the expression of M1 and M2 markers increases with plaque severity, and their presence is associated with the fate of the plaque [5,29]. In our study, we detected total TFPI protein in both M1 and M2 macrophages in human plaque samples, and the mRNA measurements indicated that M2 polarized human macrophages expressed substantially more TFPI than the M1 macrophages. Recently, Jiang and co-workers reported that M-CSF stimulated macrophages from mice expressed more TFPI compared to GM-CSF stimulated macrophages [30], which is consistent with our findings. They also found that M-CSF stimulated macrophages were significantly less procoagulant active, most likely due to the higher TFPI expression, as they found no differences in TF expression. In line with this, we also observed decreased procoagulant activity in the human M2 polarized macrophages compared to the M1 polarized macrophages. Although the TF mRNA levels actually were increased in the M2 compared to M1 polarized macrophages, the increase in TFPI mRNA levels was larger, which could contribute to altered TF/TFPI surface ratio and decreased procoagulant activity. We also observed a reduction in TFPI mRNA levels when macrophages were polarized towards a M1 phenotype. Increased procoagulant potential in the M1 macrophages, which has been located to the rupture prone shoulder region of the fibrous cap of the plaque [5], as a result of a shift in the TF/TFPI surface ratio, may have significant impact on the thrombus formation in the event of a plaque rupture. Moreover, as animal studies have shown a beneficial effect of TFPI also in plaque development [19,21], an increased procoagulant potential in M1 macrophages due to lower TFPI expression may also contribute to plaque instability and atherosclerotic progression. CCs are a hallmark of advanced atherosclerosis, but recently it was discovered that they are present also in the early stages of atherosclerotic development [31] coinciding with the first appearances of immune cells. Although M2 macrophages seem to be outnumbered in the plaque core, they accumulate oxLDL more rapidly than M1 macrophages and are more susceptible for foam cell formation [32]. Therefore, they may play an important part in lipid removal in the initial development of the plaque. Thus, the detection of TFPI in foam cells with developing CC clefts and the observation that accumulation of CCs inside macrophages can result in TFPI production might further indicate that TFPI may have an important role in dampening of plaque development. In conclusion, our data indicate that the anti-inflammatory M2 macrophages are a major cellular source of TFPI, possibly contributing to an anti-coagulant phenotype. These results suggest that an alternative M2 activation of macrophages is favourable in advanced atherosclerotic disease, and might potentially have a dampening effect on both atherothrombosis and atherogenesis.
Conflict of interest None declared.
36
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
Fig. 5. Cholesterol crystals (CC) increased TFPI expression in M2 macrophages. (A) Immunohistochemistry of human carotid plaque sections stained with an antibody detecting total TFPI. Rectangles of selected areas are shown in higher magnifications. Arrows indicate CC-clefts. One representative plaque section is shown. Scale bar = 200 μm. (B–C) Human monocytes were isolated from buffy coats, differentiated into macrophages using GM- or M-CSF (50–100 ng/mL) and polarized into M1 or M2 subtypes with LPS (100 ng/mL) plus INFγ (20 ng/mL) or IL-4 (20 ng/mL), respectively, before treatment with cholesterol crystals (CC, 500 μg/mL) for 24 h. mRNA expression of total TFPIα (B), TFPIβ (C) and TF (D) were analyzed using qRT-PCR. Mean levels (n = 7) + SD of three independent experiments are shown.
Financial support This work was funded by support from the Oslo University Hospital and the University of Oslo, and research grants from the South-Eastern Norway Regional Health Authority (Grant # 2014090), Hamar, Norway.
Author contributions SH, SE, BS, VB, TBD, and BS performed the experiments; SH, SE, and NI analyzed the results and edited the paper; BS analyzed the results and wrote the paper; PMS, TE, MS, GS and BH interpreted results and edited the paper; GS, BS, and BH designed the research. All authors read and approved the final manuscript.
Acknowledgements The authors would like to thank Mari Tinholt (Oslo University Hospital) for help with the statistical analysis of TFPI antigen results. The immunohistochemistry in Fig. 5A was performed at the Cellular and Molecular Imaging Core Facility (CMIC), Norwegian University of Science and Technology (NTNU), and the authors would like to thank Ingunn Nervik (CMIC) for her contribution. CMIC is funded by the Faculty of Medicine at NTNU and Central Norway Regional Health Authority. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.thromres.2017.04.024.
References [1] P. Libby, P.M. Ridker, G.K. Hansson, Progress and challenges in translating the biology of atherosclerosis, Nature 473 (2011) 317–325. [2] R. Ross, Atherosclerosis - an inflammatory disease, N. Engl. J. Med. 340 (1999) 115–126. [3] B. Halvorsen, K. Otterdal, T.B. Dahl, M. Skjelland, L. Gullestad, et al., Atherosclerotic plaque stability - what determines the fate of a plaque? Prog. Cardiovasc. Dis. 51 (2008) 183–194. [4] A.V. Finn, M. Nakano, J. Narula, F.D. Kolodgie, R. Virmani, Concept of vulnerable/unstable plaque, Arterioscler. Thromb. Vasc. Biol. 30 (2010) 1282–1292. [5] J.L. Stöger, M.J.J. Gijbels, S. van der Velden, M. Manca, C.M. van der Loos, et al., Distribution of macrophage polarization markers in human atherosclerosis, Atherosclerosis 225 (2012) 461–468. [6] J. Khallou-Laschet, A. Varthaman, G. Fornasa, C. Compain, A.T. Gaston, et al., Macrophage plasticity in experimental atherosclerosis, PLoS One 5 (2010), e8852. [7] H.J. Medbury, V. James, J. Ngo, K. Hitos, Y. Wang, et al., Differing association of macrophage subsets with atherosclerotic plaque stability, Int. Angiol. 32 (2013) 74–84. [8] J. Crawley, F. Lupu, A.D. Westmuckett, N.J. Severs, V.V. Kakkar, et al., Expression, localization, and activity of tissue factor pathway inhibitor in normal and atherosclerotic human vessels, Arterioscler. Thromb. Vasc. Biol. 20 (2000) 1362–1373. [9] A.F. Drew, P. Davenport, J. Apostolopoulos, P.G. Tipping, Tissue factor pathway inhibitor expression in atherosclerosis, Lab. Investig. 77 (1997) 291–298. [10] S.A. Maroney, P.E. Ellery, J.P. Wood, J.P. Ferrel, N.D. Martinez, et al., Comparison of the inhibitory activities of human tissue factor pathway inhibitor (TFPI)α and TFPIβ, J. Thromb. Haemost. 11 (2013) 911–918. [11] J.B. Hansen, N.E. Huseby, P.M. Sandset, B. Svensson, V. Lyngmo, et al., Tissue-factor pathway inhibitor and lipoproteins. Evidence for association with and regulation by LDL in human plasma, Arterioscler. Thromb. 14 (1994) 223–229. [12] M. Falciani, A.M. Gori, S. Fedi, L. Chiarugi, I. Simonetti, et al., Elevated tissue factor and tissue factor pathway inhibitor circulating levels in ischaemic heart disease patients, Thromb. Haemost. 79 (1998) 495–499. [13] S.S. Habib, Plasma tissue factor pathway inhibitor levels in angiographically defined coronary artery disease among saudis, Oman Med. J. 28 (2013) 191–194. [14] G. Novo, N. Caplice, R. Tantillo, F. Bonura, R. Simari, et al., TFPI antigen and activity levels in patients with asymptomatic atherosclerosis and target organ acute and chronic complications, Int. Angiol. 24 (2005) 366–371. [15] P.M. Sandset, P.A. Sirnes, U. Abildgaard, Factor VII and extrinsic pathway inhibitor in acute coronary disease, Br. J. Haematol. 72 (1989) 391–396.
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37 [16] K. Kaikita, M. Takeya, H. Ogawa, H. Suefuji, H. Yasue, et al., Co-localization of tissue factor and tissue factor pathway inhibitor in coronary atherosclerosis, J. Pathol. 188 (1999) 180–188. [17] N.M. Caplice, C.S. Mueske, L.S. Kleppe, R.D. Simari, Presence of tissue factor pathway inhibitor in human atherosclerotic plaques is associated with reduced tissue factor activity, Circulation 98 (1998) 1051–1057. [18] J. Crawley, F. Lupu, A.D. Westmuckett, N.J. Severs, V.V. Kakkar, et al., Expression, localization, and activity of tissue factor pathway inhibitor in normal and atherosclerotic human vessels, Arterioscler. Thromb. Vasc. Biol. 20 (2000) 1362–1373. [19] R.J. Westrick, P.F. Bodary, Z. Xu, Y.C. Shen, G.J. Broze, et al., Deficiency of tissue factor pathway inhibitor promotes atherosclerosis and thrombosis in mice, Circulation 103 (2001) 3044–3046. [20] S. Pan, L.S. Kleppe, T.A. Witt, C.S. Mueske, R.D. Simari, The effect of vascular smooth muscle cell-targeted expression of tissue factor pathway inhibitor in a murine model of arterial thrombosis, Thromb. Haemost. 92 (2004) 495–502. [21] S. Pan, T.A. White, T.A. Witt, A. Chiriac, C.S. Mueske, et al., Vascular-directed tissue factor pathway inhibitor overexpression regulates plasma cholesterol and reduces atherosclerotic plaque development, Circ. Res. 105 (2009) 713–720. [22] K. Winckers, H. ten Cate, T.M. Hackeng, The role of tissue factor pathway inhibitor in atherosclerosis and arterial thrombosis, Blood Rev. 27 (2013) 119–132. [23] Carotid artery plaque composition—relationship to clinical presentation and ultrasound B-mode imaging. European Carotid Plaque Study Group, Eur. J. Vasc. Endovasc. Surg. 10 (1995) 23–30. [24] E.G. Grant, C.B. Benson, G.L. Moneta, A.V. Alexandrov, J.D. Baker, et al., Carotid artery stenosis: gray-scale and Doppler US diagnosis—Society of Radiologists in Ultrasound Consensus Conference, Radiology 229 (2003) 340–346.
37
[25] B. Stavik, S. Espada, X.Y. Cui, N. Iversen, S. Holm, et al., EPAS1/HIF-2 alpha-mediated downregulation of tissue factor pathway inhibitor leads to a pro-thrombotic potential in endothelial cells, BBA-Mol. Basis Dis. 2016 (1862) 670–678. [26] N.M. Caplice, C.S. Mueske, L.S. Kleppe, T.E. Peterson, G.J. Broze, et al., Expression of tissue factor pathway inhibitor in vascular smooth muscle cells and its regulation by growth factors, Circ. Res. 83 (1998) 1264–1270. [27] A. Migdalski, M. Kotschy, A. Jawien, Tissue factor, tissue factor pathway inhibitor and vascular endothelial growth factor-A in carotid atherosclerotic plaques, Eur. J. Vasc. Endovasc. Surg. 30 (2005) 41–47. [28] J.J. Badimon, M. Lettino, V. Toschi, V. Fuster, M. Berrozpe, et al., Local inhibition of tissue factor reduces the thrombogenicity of disrupted human atherosclerotic plaques, Circulation 99 (1999) 1780. [29] A. Mantovani, C. Garlanda, M. Locati, Macrophage diversity and polarization in atherosclerosis: a question of balance, Arterioscler. Thromb. Vasc. Biol. 29 (2009) 1419–1423. [30] P. Jiang, D. Xue, Y. Zhang, L. Ye, Y. Liu, et al., The extrinsic coagulation cascade and tissue factor pathway inhibitor in macrophages: a potential therapeutic opportunity for atherosclerotic thrombosis, Thromb. Res. 133 (2014) 657–666. [31] P. Duewell, H. Kono, K.J. Rayner, C.M. Sirois, G. Vladimer, et al., NLRP3 inflammasomes are required for atherogenesis and activated by cholesterol crystals that form early in disease, Nature 464 (2010) 1357–1361. [32] L.J.H. van Tits, R. Stienstra, P.L. van Lent, M.G. Netea, L.A.B. Joosten, et al., Oxidized LDL enhances pro-inflammatory responses of alternatively activated M2 macrophages: a crucial role for Krüppel-like factor 2, Atherosclerosis 214 (2011) 345–349.
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Full Length Article
Increased expression of TFPI in human carotid stenosis Benedicte Stavik a,b,⁎, Sverre Holm a,h,1, Sandra Espada a,b,e,1, Nina Iversen c, Bjørnar Sporsheim i, Vigdis Bjerkeli a,f, Tuva Børresdatter Dahl a, Per Morten Sandset a,b,f, Mona Skjelland d, Terje Espevik i, Grethe Skretting a,b, Bente Halvorsen a,f,g a
Research Institute of Internal Medicine, Oslo University Hospital Rikshospitalet, Oslo, Norway Department of Haematology, Oslo University Hospital Rikshospitalet, Oslo, Norway Department of Medical Genetics, Oslo University Hospital Rikshospitalet, Oslo, Norway d Department of Neurology, Oslo University Hospital Rikshospitalet, Oslo, Norway e Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway f Institute of Clinical Medicine, University of Oslo, Oslo, Norway g K.G. Jebsen Inflammatory Research Centre, University of Oslo, Oslo, Norway h Hospital for Rheumatic Diseases, Lillehammer, Norway i Centre of Molecular Inflammation Research, Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway b c
a r t i c l e
i n f o
Article history: Received 25 January 2017 Received in revised form 4 April 2017 Accepted 25 April 2017 Available online 28 April 2017 Keywords: Atherothrombosis Tissue factor pathway inhibitor Carotid plaque Anticoagulation Vascular biology
a b s t r a c t Introduction: Tissue factor (TF) pathway inhibitor (TFPI) is the physiological inhibitor of TF induced blood coagulation and two isoforms exists, TFPIα and TFPIβ. In atherosclerotic plaques, TFPI may inhibit TF activity and thrombus formation, which is the main cause of ischemic stroke in carotid artery disease. We aimed to identify the isoforms of TFPI present in human carotid plaques and potential sources of TFPI. Materials and methods: Human atherosclerotic plaques from carotid endarterectomies were used for mRNA and immunohistochemistry analyses. hPBMCs isolated from buffy coats and THP-1 cells were differentiated and polarized into M1 or M2 macrophages, and subsequently cultured with or without cholesterol crystals (CC). mRNA and protein expression were measured with qRT-PCR and ELISA, respectively, and procoagulant activity was assessed using a two-stage chromogenic assay. Results: TFPIα and TFPIβ mRNA levels were significantly increased in carotid plaques, whereas TF levels were unchanged as compared to healthy arteries. Antibodies against total TFPI showed elevated levels compared to antibodies against free TFPIα, both by immunohistochemical and ELISA detection in plaques. The antibody against total TFPI also co-localized with CD68 and the M1 and M2 markers CD80 and CD163, respectively. The TFPI mRNA expression was elevated and the procoagulant activity was decreased in M2 compared to M1 polarized human macrophages. TFPI was present in early foam cell formation and CC treatment increased the TFPI mRNA expression even further in M2 macrophages. Conclusions: Our data indicate that both isoforms of TFPI are present in advanced plaques and that anti-inflammatory M2 macrophages may be a potential source of TFPI. © 2017 Elsevier Ltd. All rights reserved.
1. Introduction Atherosclerosis is the main precursor to carotid artery disease (CAD), which is the leading cause of ischemic stroke. It is a progressive state where influx and retention of cholesterol-containing low density lipoproteins (LDLs) in the intima of arteries trigger the recruitment of monocytes, resulting in a low grade chronic inflammation and plaque formation [1,2]. Advanced plaques can be classified as stable or unstable. ⁎ Corresponding author at: Oslo University Hospital Rikshospitalet, Research Institute of Internal Medicine, BOX 4950 Nydalen, 0424 Oslo, Norway. E-mail address: [email protected] (B. Stavik). 1 Equal contribution.
http://dx.doi.org/10.1016/j.thromres.2017.04.024 0049-3848/© 2017 Elsevier Ltd. All rights reserved.
Stable plaques are recognized by a thick fibrous cap, high degree of calcification, and substantial smooth muscle cell (SMC) proliferation. They are usually asymptomatic, however, they can attenuate the blood flow if growing into the artery lumen, thereby causing severe stenosis and ischemia. Unstable plaques typically have a thin fibrous cap, a lipid-rich necrotic core, substantial inflammation and macrophage infiltration, and neovascularization, and may cause ischemic stroke through stenosis due to plaque growth or rupture-induced thrombus formation [3,4]. Lately, the presence of different macrophage subtypes in atherosclerotic plaques has been extensively studied. While the classically activated and pro-inflammatory M1 macrophages are mainly found in the rupture prone shoulder area and in the necrotic core of the plaque, the alternatively activated anti-inflammatory M2 macrophages are
32
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
mainly located in the adventitia [5]. Consistent with this, the levels of M1 macrophage markers correlate with lesion progression and are associated with an unstable plaque phenotype [6,7]. Upon plaque rupture, cells that express tissue factor (TF) will be exposed to the blood leading to the formation of a thrombus. TF pathway inhibitor (TFPI) is the physiological inhibitor of TF induced thrombus formation and thus an essential regulator of the coagulation system. It is mainly synthesized by endothelial cells (ECs), but in atherosclerotic plaques it has also been located to macrophages, medial and neointimal SMCs, and T-cells [8,9]. Two isoforms of TFPI are produced by alternative splicing, i.e., TFPIα and TFPIβ, and they both possess anti-coagulant activity [10]. TFPIα is the full length isoform containing three Kunitz type inhibitor domains and a polar C-terminal end. It is mainly cell associated, but some is also secreted from the cells. In plasma, only a small portion of the secreted TFPIα circulates freely as a full length molecule. The remaining is found truncated at the C-terminus and/or complexed with lipoproteins, such as LDL [11]. TFPIβ is a shorter isoform lacking the third Kunitz domain and has a glycosylphosphatidylinositol attachment sequence in its C-terminal end exclusively associating it to the cell surface. The majority of previous clinical studies showed increased circulating TFPI levels in plasma of patients with atherothrombotic disease [12– 15], and also in atherosclerotic vessels compared to normal vessels [8]. Moreover, studies have demonstrated co-localization of TFPIα with TF in plaques where it is believed to play an important role in attenuating TF activity [16–18]. In fact, TFPI deficiency increased vessel occlusion time and atherothrombosis when plaque rupture was induced in atherosclerotic mice [19,20]. Interestingly, TFPI levels also affected the atherosclerotic burden in such mice [19,21]. Thus, it has been suggested that increased TFPI levels in atherosclerotic disease may represent a compensatory mechanism for elevated procoagulant activity and play a protective role not only in atherothrombosis, but also in atherogenesis [22]. In this study, we aimed to identify which TFPI isoforms are present in human carotid plaques, and to identify potential cellular sources of TFPI that may contribute to the protective effect of TFPI in the development of atherosclerotic disease.
normalizing the Ct values against the endogenous control phosphomannomutase 1 (PMM1) or β-actin, and using untreated cells or control arteries (n = 10) as a control reference. 2.4. ELISA and Immunostaining Free and total TFPI ELISA kits (Asserachrom®, Diagnostica Stago, Asnière, France) were used to measure the TFPI antigen in 55 plaque lysates and in plasma samples from 63 carotid stenosis patients and 15 matched controls, according to the manufacturer's protocols. To adjust for the amount of vessel wall lysed, these results were corrected against total protein that was measured in lysates by the modified Lowry assay (Bio-Rad DC Protein Assay, Bio-Rad Laboratories, Hercules, CA, USA). The free TFPI kit measures only free TFPIα as the antibody detects residues 161–240 of TFPI, while the total TFPI kit measures both TFPIβ and free, truncated and lipoprotein-bound forms of TFPIα, as the antibody detects residues 1–160. Detailed protocols for the immunohistochemistry and immunofluorescence methods are described in the supplementary data. The TFPIα specific antibody used for immunostaining detects residues 292–304 of TFPI that are specific for TFPIα, while the total TFPI antibody detects residues 22–87 that are common for TFPIα and TFPIβ. 2.5. Procoagulant activity
2. Materials and methods
The procoagulant activity of the cells was measured as previously described [25]. In short, polarized macrophages in 24-well plates were washed and incubated for 1 h at 37 °C with 10 nM FVIIa (Novo Nordisk, Bagsvaerd, Denmark) and 175 nM FX (Aniara Diagnostica, Mason, OH, USA) before 50 μL of supernatant from the intact cells were transferred to stop solution on ice, loaded on a 96-well microtiter plate and incubated with CS-11(22) substrate (Aniara Diagnostica). The absorbance at 405 nm was recorded at 37 °C for 45 min at 15 s intervals using a Spectra Max Plus 384 microtiter plate reader (Molecular Devices, Sunnyvale, CA, USA). The maximum velocities (Vmax) in mU/min were used to calculate the amount of FXa generated, using a standard curve obtained with known concentrations of FXa (Aniara Diagnostica).
2.1. Carotid endarterectomy specimens
2.6. Statistics
Carotid plaques from patients with internal carotid stenosis (≥70%) were collected during endarterectomy (Oslo University Hospital Rikshospitalet). A detailed description of the tissue sampling and preparation is given in the supplementary data. Patients were classified as asymptomatic or symptomatic according to the absence or presence of the clinical symptoms stroke, transient ischemic attack (TIA) or amaurosis fugax within the past 2 months or 3–6 months prior to surgery. Ultrasound plaque appearance in terms of echogenicity was classified according to consensus criteria [23,24]. Carotid stenosis and echolucency were diagnosed and classified by precerebral color Duplex ultrasound and CT angiography. A detailed description of the patients' characteristics is provided in the supplementary material (Table S2).
The distribution of the data was assessed before statistical differences between the diseased or treated samples and controls were calculated using the unpaired Student's t/Mann Whitney tests, or the oneway ANOVA/Kruskal-Wallis tests (Bonferroni corrected). Correlations were analyzed using the Pearson or Spearman's tests in GraphPad Prism 5.0 (GraphPad, San Diego, CA, USA). Odds ratios were calculated using logistic regression in SPSS (version 22.0; SPSS Inc., Chicago, IL, USA). n denotes the total number of biological replicates from several independent experiments, and a p-value of b0.05 was considered statistically significant. * = p b 0.05, ** = p b 0.01, and *** = p b 0.001.
2.2. hPBMCs and THP-1 cells
The protocols were approved by the Regional Committee for Medical and Health Research Ethics - South East Norway, approval no S-0923a 2009/6065. All studies were conducted according to the declaration of Helsinki, and signed informed consent for participation in the study was obtained from all individuals.
A detailed description of the isolation and culturing of hPBMCs and THP-1 cells is given in the supplementary data.
2.7. Ethical approval
2.3. Quantitative real-time PCR (qRT-PCR) 3. Results RNA isolation and cDNA preparation are described in the supplementary data. Gene expression was examined by qRT-PCR on the ABI PRISM 7900 Sequence Detection System (Thermo Fisher Scientific, Waltham, MA, USA) according to the protocol, using sequence specific assays (Table S1). All samples were run in triplicates. The relative mRNA expression was calculated using the comparative Ct method,
3.1. Expression of TFPI isoforms in human carotid plaques Human carotid plaque samples from 68 subjects were analyzed for mRNA expression of TFPI and TF. The total TFPI mRNA levels were 2-fold higher (p b 0.0001) in plaques than in control arteries
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
(Fig. S1). To distinguish between the two splice variants of TFPI found in humans, we used isoform specific assays detecting only TFPIα or TFPIβ. The mean relative mRNA expression of TFPIα was increased 2.2-fold (p b 0.0001), while the TFPIβ mRNA expression was increased 1.8-fold (p b 0.0013) compared to control arteries (Fig. 1A and B). In the plaques, the average Ct values of the two isoforms were 28.6 and 36.1, respectively. The TF mRNA expression in the plaques did not differ significantly from the control arteries (Fig. 1C), and the results showed no significant differences in the TFPI and TF mRNA levels according to symptomatology (Fig. S2). The gender of the patients had no significant influence on the TFPI or TF mRNA levels (Table S3). To detect TFPI protein in the plaques, we stained the human carotid plaque sections with antibodies specific for free TFPIα or total TFPI. Low staining was observed with the free TFPIα specific antibody (Fig. 2A, upper left), however, the antibody against total TFPI stained plaque sections extensively (Fig. 2A, upper right). Substantial staining was also observed with antibody against TF (Fig. 2A, lower right) in the same area of the plaque as TFPIα (Fig. 2A, lower right vs. left). Next, we measured the TFPI protein levels in 55 human carotid plaque samples using ELISA kits detecting different forms of the TFPI protein. The free TFPI kit detects only free TFPIα, while the total TFPI kit detects both free and lipoprotein-bound TFPIα, and also TFPIβ. The median (range) free and total TFPI antigen levels in the plaque lysates were 2.75 (1.36–5.26) and 3.98 (2.44–7.88) ng/ mg cell protein, respectively (Fig. 2B). When we classified the samples according to the clinical status of the patients, no significant differences between the subgroups were detected (data not shown). We then looked at the frequency of patients in the subgroups with TFPI levels above the median (Q2). Among patients experiencing symptoms in the last two months and in echolucent plaques, there
33
Fig. 2. TFPI protein was detected in human carotid plaques. (A) Immunohistochemical staining of TFPIα (upper and lower left), total TFPI (upper right), and TF (lower right) counterstained with hematoxylin in atherosclerotic carotid plaques. 100/400× magnification. Images are representative of four independent stainings. (B-D) Free and total TFPI antigen levels were measured in human carotid plaques homogenates with ELISA kits. Results were corrected against total protein measurements and expressed as ng TFPI/mg homogenate protein, and divided according to the clinical status of the patients. Box plot with median (n = 55) ± 10–90 percentile, bar graphs of occurrence of quartiles (Q) in percentage.
was an increased OR for high total TFPI levels compared to asymptomatic (OR 1.17, 95% CI 0.61–7.30) and echolucent (OR 2.11, 95% CI 0.61–7.3) plaques (Fig. 2C and Table S5). Oppositely, a reduced OR for high free TFPI levels was seen in plaques from patients experiencing symptoms in the last two months (OR 0.44 95% CI 0.11–1.76) and in echolucent plaques (OR 0.53 95% CI 0.15–2.73) compared to asymptomatic and echogenic plaques (Fig. 2D and Table S5). We also detected a small, but non-significant increase in the free TFPI levels in plasma from patients with carotid stenosis compared to healthy controls (Fig. S3), with no significant differences between asymptomatic and symptomatic patients (data not shown). The gender of the patients had no significant influence on the free TFPI levels (Table S3). 3.2. Plaque macrophages expressed TFPI
Fig. 1. TFPIα and TFPIβ mRNA levels were increased in human carotid plaques. mRNA expression of TFPIα (A), TFPIβ (B), and TF (C) in human carotid plaques and healthy control arteries analyzed using qRT-PCR. Lines show means (n ≥ 9; healthy controls, n ≥ 67; plaques) ± SD. Control arteries were obtained from the common iliac artery of deceased organ donors.
To identify the potential cellular sources of TFPI in the plaques, we first correlated TFPI mRNA expression levels with the mRNA expression levels of different markers for immune cells (CD45 and CD3g), macrophages (CD68) and lipid filled cells (ADRP). No correlation was found between the mRNA levels of TFPI and the B cell markers CD45 and CD3g (data not shown), however, both TFPIα and TFPIβ mRNA levels correlated significantly with CD68 and ADRP mRNA levels. The same correlations were observed between TF and CD68 and ADRP mRNA levels (Fig. S4). To verify that TFPI was expressed in macrophages, plaque sections were double stained with antibodies against total TFPI and CD68. The result showed that the staining of TFPI and CD68 overlapped, as seen on overlay images (Fig. 3A). TF, on the other hand, costained with CD68 in some, but not all CD68 positive cells (Fig. 3B).
34
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
Fig. 3. TFPI and TF are expressed in macrophages in human carotid plaques. Co-immunofluorescence using antibodies against total TFPI (A, C and D), or TF (B), and CD68 (A and B), CD80 (C) or CD163 (D). Sections were counterstained with DAPI. 400× magnification. Images are representative of three independent stainings. Scale bar = 25 μm.
To further investigate the expression of TFPI in plaque macrophage subtypes, we correlated the TFPI and TF mRNA levels with markers of M1 and M2. TFPIα levels, but not TFPIβ levels correlated significantly with both CCR7 and CD163, markers of M1 and M2 polarized macrophages, respectively, while the TF mRNA levels only correlated with CD163 mRNA levels (Fig. S5). To verify this at the protein level, we double stained plaque sections for TFPI and CCR7, CD80, another marker of M1 macrophages, and CD163. The results showed that TFPI co-localized with both M1 (data not shown and Fig. 3C) and M2 macrophages (Fig. 3D). 3.3. TFPI and TF expression in polarized macrophages Next, we investigated whether the TFPI expression was more prominent in either of the macrophage subtypes by differentiating and polarizing human monocytic cells into M1 and M2 in vitro. A specific and distinct macrophage polarization was obtained as M1 macrophages showed a 9- and 1.7-fold increase in TNF-α and IL-6 mRNA levels, respectively, compared to M2 macrophages, while M2 macrophages showed a 2.3-, 4-, and 33-fold increase in PPAR-γ, IL-1ra, and CD163 mRNA levels, respectively, compared to M1 macrophages (Table S4). Polarization of the human monocyte-derived macrophages (MDMs) resulted in a 7.7-fold increase in basal TFPI mRNA levels in M2 macrophages, compared to M1 macrophages (Fig. 4A). This result was also observed in polarized THP-1 cells (Fig. S6A-B). Moreover, basal TF mRNA levels were also significantly increased 2.4-fold in M2 compared to M1 macrophages (Fig. 4B and Fig. S6C).
3.4. Procoagulant activity in polarized macrophages To investigate if the increased expression of TFPI in the M2 macrophages also could affect the procoagulant activity of the cells, FXa activity of the polarized macrophages was investigated using a two-stage chromogenic substrate assay where human FVIIa and FX were added to the cells. TF present on the cell surface initiated the reaction by binding to FVIIa, which then activated FX to FXa. When we analyzed the procoagulant activity on the surface of polarized MDMs, a 2-fold increase in FXa activity in M1 polarized macrophages was obtained compared to the M2 polarized macrophages (Fig. 4C). In the THP-1 cells, a 5-fold increase in FXa activity in M1 polarized macrophages compared to M2 polarized macrophages was seen (Fig. 4D).
3.5. Cholesterol crystals induce TFPI expression We next investigated the expression of total TFPI in sections of carotid plaques and found that TFPI was strongly expressed in cells that resembled an early phase of foam cell development close to the lumen of the plaque (Fig. 5A, right). In more central parts of the plaque, imprints of CC, shown as CC-clefts, started to appear inside foam cell-like structures that were weakly positive for TFPI (Fig. 5A, lower left). To further explore this potential link between CCs and TFPI, we treated polarized MDMs with CCs and measured the TFPI mRNA expression. CC treatment increased both TFPIα and
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
Fig. 4. Basal TFPI and TF mRNA levels and procoagulant activity in polarized macrophages. (A–C) Human monocytes were isolated from buffy coats, differentiated into macrophages using GM- or M-CSF (50–100 ng/mL) and polarized into M1 or M2 subtypes with LPS (100 ng/mL) plus INFγ (20 ng/mL) or IL-4 (20 ng/mL), respectively. Resting MDMs were only incubated with conditional medium for 6 days and denoted M0. mRNA levels of total TFPI (A) and TF (B) were measured using qRT-PCR. Mean levels (n = 9) + SD of three independent experiments are show. (C–D) FXa generation was measured using a two-stage chromogenic substrate assay, where FVIIa and FX were added to the cells. After 1 h incubation at 37 °C, the reaction was stopped and the substrate (CS-11(22)) added. The maximum velocities (Vmax) in mU/min were used to calculate the amount of FXa generated, using a standard curve obtained with known concentrations of FXa in μg/L. (C) Human monocyte-derived macrophages (mean values (n = 6) + SD of two independent experiments are shown) and (D) THP-1 macrophages (monocytes treated with PMA for 6 h prior to M1 and M2 polarization using LPS (10 ng/mL) plus INFγ (5 ng/mL) or IL-4 and IL-13 (25 ng/mL), respectively, mean values (n = 15) + SD of three independent experiments are shown).
TFPIβ mRNA levels approx. 2-fold, but only in M2 polarized macrophages (Fig. 5B and C). No differences in the uptake of CC were observed in the cells (data not shown). Similar results were also obtained following treatment with oxLDL or VLDL (data not shown). In contrast, the TF mRNA levels were significantly decreased by 63% in M2 polarized macrophages after CC treatment (Fig. 5D).
4. Discussion Stroke is the second leading cause of death worldwide and N85% of these events are ischemic. Rupture of unstable atherosclerotic plaques results in exposure of TF expressing cells to the blood and thrombus formation, which is the main cause of CAD induced stenosis and stroke. TFPI is the natural inhibitor of TF induced thrombus formation and both TFPIα and TFPIβ isoforms possess anti-coagulant activity [10]. In the current study, we detected elevated levels of both TFPIα and TFPIβ mRNA in human carotid plaques, however, the differences in Ct values between the two isoforms indicated that TFPIα was more prominent. The TFPIα specific antibodies used in the IHC analyzes and ELISAs showed the presence of free full-length TFPIα in the plaques. Moreover, total TFPI was detected both with IHC and ELISA, and indicate also the presence of truncated and/or lipoprotein associated TFPIα. It is possible that the total TFPI antibody also detected TFPIβ although the mRNA levels indicated that very little TFPIβ was produced in the plaques.
35
The total TFPI protein levels in the plaques averaged 4 ng/mg homogenate protein, which is in line with what others have reported [26,27]. The TFPI protein levels in the plaques were not strongly associated with the patient clinical status, although there was a trend towards higher total TFPI levels in the recently symptomatic and lipid filled unstable plaques than in the asymptomatic and calcified plaques (data not shown). This is consistent with previous data showing elevated levels of total TFPI in plasma of patients with symptomatic vascular disease [14]. Although the ORs for high TFPI levels within the clinical subpopulations were not significant, they showed an interesting difference between total and free TFPI. While the ORs for high total TFPI levels were increased in the recently symptomatic and lipid filled unstable plaques, they were reduced for high free TFPI levels. As TFPI levels are known to affect atherothrombosis in humans [28], these findings might indicate that free TFPI levels may better reflect the symptomatic status of the patients and plaque stability than the total TFPI levels. In atherosclerotic lesions, the expression of M1 and M2 markers increases with plaque severity, and their presence is associated with the fate of the plaque [5,29]. In our study, we detected total TFPI protein in both M1 and M2 macrophages in human plaque samples, and the mRNA measurements indicated that M2 polarized human macrophages expressed substantially more TFPI than the M1 macrophages. Recently, Jiang and co-workers reported that M-CSF stimulated macrophages from mice expressed more TFPI compared to GM-CSF stimulated macrophages [30], which is consistent with our findings. They also found that M-CSF stimulated macrophages were significantly less procoagulant active, most likely due to the higher TFPI expression, as they found no differences in TF expression. In line with this, we also observed decreased procoagulant activity in the human M2 polarized macrophages compared to the M1 polarized macrophages. Although the TF mRNA levels actually were increased in the M2 compared to M1 polarized macrophages, the increase in TFPI mRNA levels was larger, which could contribute to altered TF/TFPI surface ratio and decreased procoagulant activity. We also observed a reduction in TFPI mRNA levels when macrophages were polarized towards a M1 phenotype. Increased procoagulant potential in the M1 macrophages, which has been located to the rupture prone shoulder region of the fibrous cap of the plaque [5], as a result of a shift in the TF/TFPI surface ratio, may have significant impact on the thrombus formation in the event of a plaque rupture. Moreover, as animal studies have shown a beneficial effect of TFPI also in plaque development [19,21], an increased procoagulant potential in M1 macrophages due to lower TFPI expression may also contribute to plaque instability and atherosclerotic progression. CCs are a hallmark of advanced atherosclerosis, but recently it was discovered that they are present also in the early stages of atherosclerotic development [31] coinciding with the first appearances of immune cells. Although M2 macrophages seem to be outnumbered in the plaque core, they accumulate oxLDL more rapidly than M1 macrophages and are more susceptible for foam cell formation [32]. Therefore, they may play an important part in lipid removal in the initial development of the plaque. Thus, the detection of TFPI in foam cells with developing CC clefts and the observation that accumulation of CCs inside macrophages can result in TFPI production might further indicate that TFPI may have an important role in dampening of plaque development. In conclusion, our data indicate that the anti-inflammatory M2 macrophages are a major cellular source of TFPI, possibly contributing to an anti-coagulant phenotype. These results suggest that an alternative M2 activation of macrophages is favourable in advanced atherosclerotic disease, and might potentially have a dampening effect on both atherothrombosis and atherogenesis.
Conflict of interest None declared.
36
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37
Fig. 5. Cholesterol crystals (CC) increased TFPI expression in M2 macrophages. (A) Immunohistochemistry of human carotid plaque sections stained with an antibody detecting total TFPI. Rectangles of selected areas are shown in higher magnifications. Arrows indicate CC-clefts. One representative plaque section is shown. Scale bar = 200 μm. (B–C) Human monocytes were isolated from buffy coats, differentiated into macrophages using GM- or M-CSF (50–100 ng/mL) and polarized into M1 or M2 subtypes with LPS (100 ng/mL) plus INFγ (20 ng/mL) or IL-4 (20 ng/mL), respectively, before treatment with cholesterol crystals (CC, 500 μg/mL) for 24 h. mRNA expression of total TFPIα (B), TFPIβ (C) and TF (D) were analyzed using qRT-PCR. Mean levels (n = 7) + SD of three independent experiments are shown.
Financial support This work was funded by support from the Oslo University Hospital and the University of Oslo, and research grants from the South-Eastern Norway Regional Health Authority (Grant # 2014090), Hamar, Norway.
Author contributions SH, SE, BS, VB, TBD, and BS performed the experiments; SH, SE, and NI analyzed the results and edited the paper; BS analyzed the results and wrote the paper; PMS, TE, MS, GS and BH interpreted results and edited the paper; GS, BS, and BH designed the research. All authors read and approved the final manuscript.
Acknowledgements The authors would like to thank Mari Tinholt (Oslo University Hospital) for help with the statistical analysis of TFPI antigen results. The immunohistochemistry in Fig. 5A was performed at the Cellular and Molecular Imaging Core Facility (CMIC), Norwegian University of Science and Technology (NTNU), and the authors would like to thank Ingunn Nervik (CMIC) for her contribution. CMIC is funded by the Faculty of Medicine at NTNU and Central Norway Regional Health Authority. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.thromres.2017.04.024.
References [1] P. Libby, P.M. Ridker, G.K. Hansson, Progress and challenges in translating the biology of atherosclerosis, Nature 473 (2011) 317–325. [2] R. Ross, Atherosclerosis - an inflammatory disease, N. Engl. J. Med. 340 (1999) 115–126. [3] B. Halvorsen, K. Otterdal, T.B. Dahl, M. Skjelland, L. Gullestad, et al., Atherosclerotic plaque stability - what determines the fate of a plaque? Prog. Cardiovasc. Dis. 51 (2008) 183–194. [4] A.V. Finn, M. Nakano, J. Narula, F.D. Kolodgie, R. Virmani, Concept of vulnerable/unstable plaque, Arterioscler. Thromb. Vasc. Biol. 30 (2010) 1282–1292. [5] J.L. Stöger, M.J.J. Gijbels, S. van der Velden, M. Manca, C.M. van der Loos, et al., Distribution of macrophage polarization markers in human atherosclerosis, Atherosclerosis 225 (2012) 461–468. [6] J. Khallou-Laschet, A. Varthaman, G. Fornasa, C. Compain, A.T. Gaston, et al., Macrophage plasticity in experimental atherosclerosis, PLoS One 5 (2010), e8852. [7] H.J. Medbury, V. James, J. Ngo, K. Hitos, Y. Wang, et al., Differing association of macrophage subsets with atherosclerotic plaque stability, Int. Angiol. 32 (2013) 74–84. [8] J. Crawley, F. Lupu, A.D. Westmuckett, N.J. Severs, V.V. Kakkar, et al., Expression, localization, and activity of tissue factor pathway inhibitor in normal and atherosclerotic human vessels, Arterioscler. Thromb. Vasc. Biol. 20 (2000) 1362–1373. [9] A.F. Drew, P. Davenport, J. Apostolopoulos, P.G. Tipping, Tissue factor pathway inhibitor expression in atherosclerosis, Lab. Investig. 77 (1997) 291–298. [10] S.A. Maroney, P.E. Ellery, J.P. Wood, J.P. Ferrel, N.D. Martinez, et al., Comparison of the inhibitory activities of human tissue factor pathway inhibitor (TFPI)α and TFPIβ, J. Thromb. Haemost. 11 (2013) 911–918. [11] J.B. Hansen, N.E. Huseby, P.M. Sandset, B. Svensson, V. Lyngmo, et al., Tissue-factor pathway inhibitor and lipoproteins. Evidence for association with and regulation by LDL in human plasma, Arterioscler. Thromb. 14 (1994) 223–229. [12] M. Falciani, A.M. Gori, S. Fedi, L. Chiarugi, I. Simonetti, et al., Elevated tissue factor and tissue factor pathway inhibitor circulating levels in ischaemic heart disease patients, Thromb. Haemost. 79 (1998) 495–499. [13] S.S. Habib, Plasma tissue factor pathway inhibitor levels in angiographically defined coronary artery disease among saudis, Oman Med. J. 28 (2013) 191–194. [14] G. Novo, N. Caplice, R. Tantillo, F. Bonura, R. Simari, et al., TFPI antigen and activity levels in patients with asymptomatic atherosclerosis and target organ acute and chronic complications, Int. Angiol. 24 (2005) 366–371. [15] P.M. Sandset, P.A. Sirnes, U. Abildgaard, Factor VII and extrinsic pathway inhibitor in acute coronary disease, Br. J. Haematol. 72 (1989) 391–396.
B. Stavik et al. / Thrombosis Research 155 (2017) 31–37 [16] K. Kaikita, M. Takeya, H. Ogawa, H. Suefuji, H. Yasue, et al., Co-localization of tissue factor and tissue factor pathway inhibitor in coronary atherosclerosis, J. Pathol. 188 (1999) 180–188. [17] N.M. Caplice, C.S. Mueske, L.S. Kleppe, R.D. Simari, Presence of tissue factor pathway inhibitor in human atherosclerotic plaques is associated with reduced tissue factor activity, Circulation 98 (1998) 1051–1057. [18] J. Crawley, F. Lupu, A.D. Westmuckett, N.J. Severs, V.V. Kakkar, et al., Expression, localization, and activity of tissue factor pathway inhibitor in normal and atherosclerotic human vessels, Arterioscler. Thromb. Vasc. Biol. 20 (2000) 1362–1373. [19] R.J. Westrick, P.F. Bodary, Z. Xu, Y.C. Shen, G.J. Broze, et al., Deficiency of tissue factor pathway inhibitor promotes atherosclerosis and thrombosis in mice, Circulation 103 (2001) 3044–3046. [20] S. Pan, L.S. Kleppe, T.A. Witt, C.S. Mueske, R.D. Simari, The effect of vascular smooth muscle cell-targeted expression of tissue factor pathway inhibitor in a murine model of arterial thrombosis, Thromb. Haemost. 92 (2004) 495–502. [21] S. Pan, T.A. White, T.A. Witt, A. Chiriac, C.S. Mueske, et al., Vascular-directed tissue factor pathway inhibitor overexpression regulates plasma cholesterol and reduces atherosclerotic plaque development, Circ. Res. 105 (2009) 713–720. [22] K. Winckers, H. ten Cate, T.M. Hackeng, The role of tissue factor pathway inhibitor in atherosclerosis and arterial thrombosis, Blood Rev. 27 (2013) 119–132. [23] Carotid artery plaque composition—relationship to clinical presentation and ultrasound B-mode imaging. European Carotid Plaque Study Group, Eur. J. Vasc. Endovasc. Surg. 10 (1995) 23–30. [24] E.G. Grant, C.B. Benson, G.L. Moneta, A.V. Alexandrov, J.D. Baker, et al., Carotid artery stenosis: gray-scale and Doppler US diagnosis—Society of Radiologists in Ultrasound Consensus Conference, Radiology 229 (2003) 340–346.
37
[25] B. Stavik, S. Espada, X.Y. Cui, N. Iversen, S. Holm, et al., EPAS1/HIF-2 alpha-mediated downregulation of tissue factor pathway inhibitor leads to a pro-thrombotic potential in endothelial cells, BBA-Mol. Basis Dis. 2016 (1862) 670–678. [26] N.M. Caplice, C.S. Mueske, L.S. Kleppe, T.E. Peterson, G.J. Broze, et al., Expression of tissue factor pathway inhibitor in vascular smooth muscle cells and its regulation by growth factors, Circ. Res. 83 (1998) 1264–1270. [27] A. Migdalski, M. Kotschy, A. Jawien, Tissue factor, tissue factor pathway inhibitor and vascular endothelial growth factor-A in carotid atherosclerotic plaques, Eur. J. Vasc. Endovasc. Surg. 30 (2005) 41–47. [28] J.J. Badimon, M. Lettino, V. Toschi, V. Fuster, M. Berrozpe, et al., Local inhibition of tissue factor reduces the thrombogenicity of disrupted human atherosclerotic plaques, Circulation 99 (1999) 1780. [29] A. Mantovani, C. Garlanda, M. Locati, Macrophage diversity and polarization in atherosclerosis: a question of balance, Arterioscler. Thromb. Vasc. Biol. 29 (2009) 1419–1423. [30] P. Jiang, D. Xue, Y. Zhang, L. Ye, Y. Liu, et al., The extrinsic coagulation cascade and tissue factor pathway inhibitor in macrophages: a potential therapeutic opportunity for atherosclerotic thrombosis, Thromb. Res. 133 (2014) 657–666. [31] P. Duewell, H. Kono, K.J. Rayner, C.M. Sirois, G. Vladimer, et al., NLRP3 inflammasomes are required for atherogenesis and activated by cholesterol crystals that form early in disease, Nature 464 (2010) 1357–1361. [32] L.J.H. van Tits, R. Stienstra, P.L. van Lent, M.G. Netea, L.A.B. Joosten, et al., Oxidized LDL enhances pro-inflammatory responses of alternatively activated M2 macrophages: a crucial role for Krüppel-like factor 2, Atherosclerosis 214 (2011) 345–349.