Interferon-alpha promotes abnormal vasculogenesis in lupus: a potential pathway for premature atherosclerosis

Abstract

Individuals with systemic lupus erythematosus (SLE) have a striking increase in premature atherosclerosis of unclear etiology. Accelerated endothelial cell apoptosis occurs in SLE and correlates with endothelial dysfunction. Endothelial progenitor cells (EPCs) and myelomonocytic circulating angiogenic cells (CACs) are crucial in blood vessel repair after vascular damage, and decreased levels or abnormal function of EPCs/CACs are established atherosclerosis risk factors. We investigated if vascular repair is impaired in SLE. We report that SLE patients display abnormal phenotype and function of EPCs/CACs. These abnormalities are characterized by significant decreases in the number of circulating EPCs (310 +/- 50 EPCs/mL of blood in SLE versus 639 +/- 102 in controls) and significant impairments in the capacity of EPCs/CACs to differentiate into mature ECs and synthesize adequate levels of the proangiogenic molecules vascular endothelial growth factor (VEGF) and hepatic growth factor (HGF). These abnormalities are triggered by interferon-alpha (IFN-alpha), which induces EPC and CAC apoptosis and skews myeloid cells toward nonangiogenic phenotypes. Lupus EPCs/CACs have increased IFN-alpha expression and their supernatants promote higher induction of IFN-inducible genes. Importantly, neutralization of IFN pathways restores a normal EPC/CAC phenotype in lupus. SLE is characterized by an imbalance between endothelial cell damage and repair triggered by type I IFNs, which might promote accelerated atherosclerosis.

Figure 1

SLE EPCs are decreased in peripheral blood. (A) Dot plots of 1 representative control and 1 representative SLE patient displaying no. of EPCs/mL of blood. Total events for acquisition were as follows: for controls, 127 000 ± 18 000; for SLE, 87 000 ± 6400 (mean ± SEM). (B) Results represent the mean (± SEM). (C) Decreased EPC numbers in SLE correlate with disease activity. EPCs are plotted as a function of SLEDAI score in individual patients. EPC levels in controls also exceed those in lupus patients with SLEDAI scores of zero (639 EPC/mL blood in control vs 370 EPC/mL blood in SLEDAI = 0, P < .05). (D) Lack of association between circulating EPCs and daily prednisone dose in SLE. Error bars are 95% CI whiskers.

Figure 2

Lupus PBMCs fail to form mature EC monolayers. (A) Control PBMCs were plated into fibronectin-coated wells in complete EC media. On day 15, cells were stained with anti–VWF-FITC, UEA-1–Texas red, and Hoechst 33342 and analyzed by fluorescent microscopy. Images are from 1 representative healthy control and show single fluorophores and a merged image (bottom; × 20 objective magnification). (B) On day 15, wells were examined for EC monolayer development. Representative images of PBMC-derived cells from a healthy control and a lupus patient (× 5 objective magnification). See “In vitro differentiation into mature ECs” for more image acquisition information. (C) The y-axis represents the percentage of controls (13/14) or lupus patients (7/33) that formed an EC monolayer.

Figure 3

EC monolayer phenotype in SLE and controls. (A) No differences in acetylated-LDL uptake were detected between SLE and controls, although there were significantly fewer cells in SLE cultures at day 15 (× 10 objective magnification). See “In vitro differentiation into mature ECs” for more image acquisition information. (B) Day 15 EC monolayers were analyzed for expression of HLA-DP, -R, and -Q and CD14. Dot plots display a representative healthy control and lupus patient. MHC class II and CD14 expression is consistent with CD14-derived CACs and did not differ between SLE and controls. (C) mRNA expression of proangiogenic factors on day 15 ECs from controls or SLE, normalized to GAPDH. HGF and VEGF-B expression was lower in lupus patients (1-tailed t test, P < .05 for HGF and P = .06 for VEGF-B). Results are mean (± SEM) of 8 patients and 4 controls. (D-E) Proangiogenic molecules are decreased in lupus sera. Results represent mean (± SEM) of 9 controls and 47 SLE patients.

Figure 4

Increased IFN-α expression in lupus EPCs/CACs. (A) Lupus serum prevents monolayer formation by healthy control EPCs/CACs. Control PBMCs were plated on fibronectin-coated wells with complete EC media in the presence or absence of 20% allogeneic control or SLE sera. At day 15, EC monolayer formation was assessed. Images are representative of 7 of 8 SLE serum samples and 4 of 4 control serum samples used in 5 allogeneic control PBMCs. Images represent bright field (top), diI–ac-LDL (middle), and UEA-1–FITC (bottom) of allogeneic control cells exposed to 1 representative SLE serum sample (left) and 1 representative control serum sample (right; × 10 objective magnification). See “In vitro differentiation into mature ECs” for more image acquisition information. (B) Graphs represent percentage IFN-α expression (± SEM) of 13 SLE and 6 controls in PBMCs cultured under proangiogenic conditions. (C) IFN-α expression on day 1–cultured cells of a representative control and SLE patient. (D) EPC/CAC supernatants and autologous serum from SLE patients induce higher expression of IFN-α–inducible genes in epithelial cell lines than controls. Results represent fold induction of IFN-inducible genes (mRNA) and are presented as mean (± SEM) of supernatants or sera from 8 controls and 23 SLE patients. Data are normalized to HPRT-1.

Figure 5

IFN-α treatment prevents EC monolayer formation in healthy controls. (A) Control PBMCs were plated on fibronectin-coated wells with complete EC media in the presence or absence of graded concentrations of recombinant IFN-α. At day 15, EC monolayer formation was assessed. Images are from experiments from 4 representative controls. Similar results were seen in 4 SLE patients using similar concentrations of IFN-α. Magnifications are × 10. See “In vitro differentiation into mature ECs” for more image acquisition information. (B-E) IFN-α induces CAC apoptosis and skewing toward other myeloid cell subsets. Representative dot plots display percentage apoptotic CD14⁺ CACs in untreated and IFN-α–treated cells. Graphs represent mean (± SEM) percentage expression of myeloid subsets on the EC monolayer of 5 controls after IFN-α treatment. * indicates P < .05.

Figure 6

IFN-α induces EPC apoptosis. (A,B) Representative dot plots show increased BM and circulating CD133⁺ cytotoxicity by IFN-α. (C) Graph is representative of 2 independent experiments from healthy controls.

Figure 7

IFN-α blockade promotes EC monolayer formation in SLE. Lupus PBMCs were cultured under proangiogenic conditions in the presence or absence of (A) blocking antihuman IFN-α mAb or isotype control or (B) antihuman type I IFN-R or isotype control (all 2 μg/mL). At day 15, anti–IFN-α– or anti–type I IFN-R–treated cells, but not controls, formed EC monolayers. Results are representative of independent experiments in SLE patients who failed to form EC monolayers, which corrected with anti–IFN-α (n = 9) or with anti–type I IFN-R (n = 6). (C) TLR7 and/or TLR9 blockade promotes EC monolayer formation in SLE. Lupus PBMCs were cultured in EGM20 supplemented with a control ODN, a TLR7 antagonist, a TLR9 antagonist, or a TLR7/9 antagonist (1 μM). Images were acquired after 15 days in culture. Results show cells from a representative SLE patient. All magnifications are × 20. See “In vitro differentiation into mature ECs” for more image acquisition information.