Increased levels of the homeostatic chemokine CXCL13 in human atherosclerosis – Potential role in plaque stabilization  Linda M. Smedbakken, Bente Halvorsen, Isabelle Daissormont, Trine Ranheim, Annika E. Michelsen, Mona Skjelland, Ellen Lund Sagen, Lasse Folkersen, Kirsten Krohg-Sørensen, David Russell, Sverre Holm, Thor Ueland, Børre Fevang, Ulf Hedin, Arne Yndestad, Lars Gullestad, Göran K. Hansson, Erik A. Biessen, Pål Aukrust  Atherosclerosis  Volume 224, Issue 1, Pages 266-273 (September 2012) DOI: 10.1016/j.atherosclerosis.2012.06.071 Copyright © 2012 Elsevier Ireland Ltd Terms and Conditions

Fig. 1 Levels of CXCL13 and CXCR5 in patients with carotid atherosclerosis. Panel A shows plasma levels of CXCL13 in patients with symptomatic (ie, symptoms within the latest 6 months) (n = 89) and asymptomatic (n = 41) carotid plaques, and in 20 healthy controls. Panel B shows the mRNA levels of CXCL13 and CXCR5 in atherosclerotic carotid plaques (n = 106) and in control tissue, obtained from iliac arteries of organ donors (n = 10). Panel C and D show CXCL13 mRNA levels in relation to time between latest symptom and surgery (0–1 month [mo], n = 37; 1–3 mo, n = 57; ≥3 mo, n = 12) (C) and in relation to plaque echogenicity (high, n = 27 and low, n = 33; note that ultrasound data were not available from all patients) (D). mRNA was quantified by means of Affymetrix Gene Array analysis and are given in arbitrary unit. The full data set from the affymetrix gene array has recently been published [23]. Data are medians, interquartile range and outliers (Tukey box and whiskers). ***p < 0.001 versus controls. In panel C, the p value represents the results of the Kruskal–Wallis test comparing all three groups of individuals. 0–1 mo versus 1–3 mo, p = 0.01; 0–1 versus ≥3 mo, p = 0.002. Atherosclerosis 2012 224, 266-273DOI: (10.1016/j.atherosclerosis.2012.06.071) Copyright © 2012 Elsevier Ireland Ltd Terms and Conditions

Fig. 2 CXCL13 and CXCR5 co-expression with vascular SMC and macrophages in advanced human atherosclerotic lesions. CXCL13 and CXCR5 are both highly expressed by medial SMC (CXCL13(red)/aSMA(blue) A (10×) and B (40×) and CXCR5(red)/aSMA(blue) C (10×) and D (40×)). Macrophages inside the lesions are strongly positive for CXCL13 and CXCR5 (CXCL13(blue)/CD68(red) E (10×) and F (40×) and CXCR5(blue)/CD68(red) G (10×) and H (40×)). CXCL13 is expressed by macrophages in IPH lesions as well (I (10×) and J (40×)). As isotype control for CXCL13 staining, adjacent slides were stained with a goat IgG control antibody (K), and for CXCR5, a mouse IgG2b antibody was used (L) (20×).(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Atherosclerosis 2012 224, 266-273DOI: (10.1016/j.atherosclerosis.2012.06.071) Copyright © 2012 Elsevier Ireland Ltd Terms and Conditions

Fig. 3 Regulation of CXCL13 in monocytes. Panels A–B show the effect of LPS (5 ng/ml), Pam3Cys (1 μg/ml), relaseate from un-stimulated (uPRL) and thrombin-activated (sPRL) platelets, and IL-1β (1 ng/ml) on mRNA (A) and protein (B) levels of CXCL13 in THP-1 monocytes after culturing for 6 and 24 h, respectively. Panel C shows the effect of LPS (5 ng/ml), Pam3Cys (1 μg/ml), and IL-1β (1 ng/ml) on the mRNA levels of CXCL13 in primary monocytes from 7 healthy controls after culturing for 6 h. Panel D shows the effect of relaseate from uPRL and sPRL on the mRNA levels of CXCL13 in primary monocytes from 7 healthy controls after culturing for 6 h. Panels E and F show the effect of sPRL on the mRNA levels of CXCL13 in THP-1 monocytes (E) and primary monocytes from 7 healthy controls (F), with or without co-culturing with neutralizing antibodies (Ab) against RANTES or isotype-matched control antibodies, after culturing for 6 h. The isotype-matched control antibody had no effect on the sPRL-mediated induction of CXCL13. Note, in panels A, B, D, E, and F, but not C, the cells were pre-activated with TNFα (5 ng/ml) for 96 h (48 h in primary monocytes) prior to experimental start. mRNA levels were assessed by real-time RT-PCR in relation to the control gene β-actin, and CXCL13 protein levels in cell supernatants were measured by EIA. Data are mean ± SEM (n = 4–8). *p < 0.05, **p < 0.01, and ***p < 0.001 versus controls (vehicle). †p < 0.05 versus sPRL without neutralizing antibodies against RANTES. In panel D, ###p < 0.01 reflects comparison with un-stimulated cells, ie, without TNFα pre-activation. Atherosclerosis 2012 224, 266-273DOI: (10.1016/j.atherosclerosis.2012.06.071) Copyright © 2012 Elsevier Ireland Ltd Terms and Conditions

Fig. 4 CXCL13 exerts anti-apoptotic effect in monocytes/macrophages and SMC. PMA differentiated THP-1 macrophages or primary monocytes from 7 healthy controls were stimulated for 24 h (18 h in primary monocytes) with or without 7-ketocholesterol (7-keto, 40 μg/ml), camptothecin (Campto, 10 μM), rhCXCL13 (200 ng/ml), or a combination thereof. Panel A (n = 7) shows the degree of early apoptosis assessed by flow cytometric detection of Annexin V-FITC expressed as percentage of Annexin V positive cells. Panel B shows the degree of apoptosis, as assessed by EIA measurements of histone-associated DNA fragments, in primary monocytes under similar experimental condition. Panel C (n = 4) shows the degree of apoptosis, as assessed by EIA measurements of histone-associated DNA fragments, in SMC that were exposed to 7-ketocholesterol (40 μg/ml) and camptothecin (10 μM) for 12 h with and without rhCXCL13 (200 ng/ml). Data are mean ± SEM. *p < 0.05, **p < 0.01 and ***p < 0.001 versus 7-ketocholesterol without rhCXCL13; †p < 0.05 and †††p < 0.001 versus camptothecin without CXCL13. Atherosclerosis 2012 224, 266-273DOI: (10.1016/j.atherosclerosis.2012.06.071) Copyright © 2012 Elsevier Ireland Ltd Terms and Conditions

Fig. 5 The effect of CXCL13 in monocytes and atherosclerotic carotid plaques. Panels A and B show the effect of rhCXCL13 (200 ng/ml) on mRNA levels of arginase-1 (Arg-1), TGF-β, and IL-10 in THP-1 monocytes (A) and primary monocytes from 7 healthy controls (B) after culturing for 6 h. Prior to experiment start, the cells were pre-activated with TNFα (5 ng/ml) for 96 h. Panels C–F show the effect of rhCXCL13 (500 ng/ml) on mRNA (C) and protein (D–F) levels of arginase-1, IL-10 and TGF-β in freshly isolated carotid plaque samples obtained from patients that had been suffering from symptoms within 1 month and that were undergoing carotid endarterectomy (n = 7). Panel A–C shows mRNA levels of these mediators as assessed by real-time RT-PCR in relation to the control gene β-actin. Panel D shows protein levels of arginase-1 in plaque lysates as assessed by western blotting. The upper part shows blots from two representative carotid plaques. Panel D–E shows protein levels of IL-10 (E) and TGF-β1 (F) as assessed by EIA in plaque lysates (plaque associated) and plaque supernatants (released). In all experiments, including the plaque experiments, controls represent un-stimulated samples given vehicle only. Data are mean ± SEM. *p ≤ 0.05 and **p < 0.01 versus control (vehicle). Atherosclerosis 2012 224, 266-273DOI: (10.1016/j.atherosclerosis.2012.06.071) Copyright © 2012 Elsevier Ireland Ltd Terms and Conditions