The detrimental effects of interferon-α on vasculogenesis in lupus are mediated by repression of IL-1 pathways: potential role in atherogenesis and renal vascular rarefaction.

Abstract

Systemic lupus erythematosus (SLE) is characterized by increased vascular risk due to premature atherosclerosis independent of traditional risk factors. We previously proposed that Interferon-α (IFN-α) plays a crucial role in premature vascular damage in SLE. IFN-α alters the balance between endothelial cell apoptosis and vascular repair mediated by endothelial progenitor cells (EPCs) and myeloid circulating angiogenic cells (CACs). Here we demonstrate that IFN-α promotes an antiangiogenic signature in SLE and control EPCs/CACs, characterized by transcriptional repression of IL-1α and β, IL-1 receptor 1 and vascular endothelial growth factor A (VEGF-A) and upregulation of IL-1 receptor antagonist (IL-1RN) and the decoy receptor IL1-R2. IL-1β promotes significant improvement in the functional capacity of lupus EPCs/CACs, therefore abrogating the deleterious effects of IFN-α. The beneficial effects from IL-1 are mediated, at least in part, by increases in EPC/CAC proliferation, decreases in EPC/CAC apoptosis, and by preventing the skewing of CACs towards non-angiogenic pathways. IFN-α induces STAT2 and 6 phosphorylation in EPCs/CACs and JAK inhibition abrogates the transcriptional antiangiogenic changes induced by IFN-α in these cells. Immunohistochemistry of renal biopsies from patients with lupus nephritis, but not ANCA-positive vasculitis, showed this pathway to be operational in vivo, with increased IL-1RN , downregulation of VEGF-A and glomerular and blood vessel decreased capillary density, compared to controls. Our study introduces a novel putative pathway by which type I IFNs may interfere with vascular repair in SLE through repression of IL-1-dependent pathways. This could promote atherosclerosis and loss of renal function in this disease.

INTRODUCTION

Systemic lupus erythematosus (SLE) is an autoimmune disease that primarily affects women of childbearing age(1). A significant proportion of patients with lupus develop renal disease. In addition to inflammatory nephritis, renal involvement in lupus can manifest as a fibrotic, atrophic nephropat hy with significant renal functional impairment and potential progression to end-stage disease. Although this manifestation can be the result of earlier unchecked inflammation, alternative mechanisms mediating progressive loss of renal function may be envisioned, and the precise relationship between acute inflammatory and chronic fibrotic nephropathy remains unclear (2, 3) .

Further, SLE is characterized by strikingly higher rates of premature atherosclerotic cardiovascular disease (CVD) (4, 5),(6, 7) not explained by Framingham risk factors (6, 8, 9). While immune dysregulation may play the dominant role in atherogenesis(10), the exact mechanisms leading to enhanced CV risk in lupus remain to be determined. Our group previously reported that SLE patients without traditional CV risk factors display a striking imbalance between endothelial cell damage and repair. This is manifested by an increase in circulating apoptotic endothelial cells uncoupled from proper endothelial repair, as shown by a significant decrease in the numbers and function of bone marrow derived endothelial progenitor cells (EPCs) and circulating myeloid angiogenic cells (CACs) (11, 12). High levels of circulating apoptotic endothelial cells in SLE strongly correlate with endothelial dysfunction (12), a surrogate marker of future atherosclerosis development(13). Additional studies have al so reported aberrant phenotype and function of lupus EPCs/CACs in SLE patients(14).

Type I Interferons (IFNs), particularly IFN-α, have been proposed to play major pathogenic roles in SLE(15, 16). However, the possibility that they may play a prominent role in premature vascular damage in SLE had not been systematically investigated. Our group previously reported that IFN-α induces EPC/CAC apoptosis and skews myeloid cells away from CACs and towards nonangiogenic phenotypes including mature dendritic cells (DCs) (11). Importantly, neutralization of type I IFN pathways restores normal EPC/CAC phenotype and function in SLE (11). Further, the New Zealand Black/New Zealand White F₁ murine model of lupus, a mouse strain where type I IFNs are considered to play a prominent role in pathogenesis(17-20), is also characterized by endothelial dysfunction and aberrant EPC phenotype and function(21, 22),(23). All these observations support a potential role for type I IFNs in the development of premature atherosclerosis and altered vasculogenesis in SLE.

In murine and human systems, EPCs and CACs appear to be crucial in vasculogenesis and angiogenesis(24-28). Decreases in EPC/CAC numbers and function in disease states associated to enhanced vascular complications, correlate with an increased risk of atherosclerosis and vascular events (29-31). While type I IFNs have been implicated in the progression of SLE (32), in the severity of lupus nephritis and in the alteration of endothelial cell repair (11, 14), the molecular pathways by which they mediate these antiangiogenic effects have not been well characterized and studies have primarily focused on cancer cell lines(33-40). This study investigated the molecular pathways that are affected in EPCs/CACs by IFN-α exposure and further assessed the mechanisms by which this molecule interferes with vasculogenesis in SLE.

MATERIAL AND METHODS

Patient selection

The University of Michigan institutional review board approved this study. Subjects gave informed consent in accordance with the Declaration of Helsinki. To obtain peripheral blood, patients fulfilled the revised American College of Rheumatology criteria for SLE(41) and were enrolled from the University of Michigan outpatient Rheumatology clinic. Age- and gender- matched healthy controls were recruited by advertisement. Lupus disease activity was assessed by the SLE Disease Activity Index (SLEDAI)(42) (Supplementary Table I).

Cell isolation and culture and fluorescent microscopy

Human PBMCs, known to contain both EPCs and CACs (43-45), were isolated and cultured under proangiogenic stimulation as previously described (11), with a few modifications. Briefly, PBMCs (2.27 × 10⁶/cm²) were cultured in endothelial cell-specific enrichment medium (EBM2; Cambrex, East Rutherford, NJ) on fibronectin-coated wells (BD, Franklin Lakes, NJ). Typically, after 1-3 weeks in culture, these cells differentiate into mature endothelial cells that display typical mature endothelial cell markers and display functional characteristics of vascular cells (11, 43-45). For some of the experiments (gene expression studies), media was changed after 72 hours and fresh media was added with or without human recombinant IFN-α 2b (Schering, Kenilworth, NJ) at a final concentration of 1000 IU/ml. Cells were incubated in the presence or absence of IFN-α for 6 hours prior to RNA harvesting. This dose was chosen due to its effects in inhibiting vasculogenesis in control EPC/CAC cultures (11). Similar experiments were performed with EPCs directly obtained from control bone marrow. In brief, CD133+ cells obtained from healthy control bone marrows (AllCells, Emeryville, CA) were thawed following manufacturer’s recommendations and then cultured in StemSpan SFEM (Stemcell Technologies, Vancouver, BC, Canada) for 48 hours to allow for cell recovery and expansion, then under similar conditions as peripheral blood EPCs/CACs but at a density of 5×10⁵ cells/ml.

To assess the capacity of peripheral blood EPCs/CACs to differentiate into mature endothelial cells, media was changed 120 hours after plating, then every 3 days. On days 14 to 21, cells were incubated with markers of mature endothelial cells, including 1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (diI)–acetylated LDL (Biomedical Technologies, Stoughton, MA), and FITC–Ulex europaeus agglutinin-1 (UEA-1; Vector Labs, Burlingame, CA). In experiments to assess the effect of exogenous proangiogenic molecules effects on SLE and control EPC/CAC function, IL-1β (10ng/ml), IL-6 (10ng/ml), TNF-α (10ng/ml) (Peprotech, Rocky Hill, NJ) or vascular endothelial growth factor (1ug/ml) (VEGF, National Cancer Institute, Bethesda, MD) were added to EPC/CAC cultures and replenished every 48 hours.

To assess endothelial cell morphology and expression of endothelial cell markers, cells were analyzed by fluorescent microscopy using a Leica DMIRB fluorescent inverted microscope (Bannockburn, IL). Images were acquired at room temperature using live cells in PBS without mounting media. A total of 8 random fields of view were acquired and images were analyzed using the CellC program (http://www.cs.tut.fi/sgn/csb/cellc/) to quantify mature endothelial cells, which were considered as those that coexpress UEA-1 and acetylated LDL. Images were acquired with a 100X total magnification. The numeric aperture for the objective lens of the fluorescent microscope was 0.3. Images were acquired with an Olympus DP30BW camera (Olympus Corporation, Tokyo, Japan) using the acquisition software Olympus-BSW (Olympus). Final processing was done with Adobe Photoshop CS2 (San Jose, CA).

Phosphorylated STAT detection and inhibition of IFN-α signalling pathways

EPCs/CACs were cultured under proangiogenic stimulation as stated above for 60 hours, then media was harvested and changed to EGM without 20% FBS. After overnight incubation in serum-reduced media, cells were stimulated with 1000 IU IFN-α for 30 minutes followed by fixation with 4% paraformaldehyde. Cells were permeabilized with 90% methanol and incubated with anti-pSTAT2-FITC (R&D Systems, Minneapolis, MN), and anti-pSTAT6-FITC (Cell Signaling Technology, Danvers, MA). Immunofluorescence was acquired using FACSCalibur (BD Biosciences), followed by analysis with FlowJo (Treestar, Ashland, OR).

To inhibit IFN-α signalling, the pan-Janus Kinase (JAK) inhibitor 2-(1,1-Dimethylethyl)-9-fluoro-3,6-dihydro-7H-benz[h]-imidaz[4,5-f]isoquinolin-7-one (Pyridone 6) or the phosphoinositide-3-kinase (PI3K) inhibitor (5-(4-Fluoro-2-hydroxyphenyl)furan-2-ylmethylene)thiazolidine-2,4-dione were used (Calbiochem, Gibbstown, NJ). In brief, human EPCs/CACs were cultured under proangiogenic stimulation for 72 hours, followed by change of media containing 50μm of pyridone 6, 50nm of the PI-3K inhibitor, or vehicle (DMSO) for 1 hour before addition of 1000 U of recombinant IFN-α. EPC/CACs were incubated for 6 additional hours and then total RNA was isolated.

RNA isolation

Total RNA was isolated with Tri-pure (Roche, Indianapolis, IN) following the manufacturer’s recommendations. For microarray analysis, RNA was further purified and concentrated using an RNeasy micro kit and following the manufacturer’s recommendations (Qiagen, Valencia, CA). RNA samples were processed on an Agilent 2100 BioAnalyzer (Agilent Technologies, Santa Clara, CA) to assess RNA integrity.

Microarray data processing, analysis and pathway mapping

Affymetrix Human U133 Plus 2.0 Genechips (Affymetrix, Inc, Santa Clara, CA) were processed at the University of Michigan Microarray Core Facility following the manufacturer’s instructions (46). The samples analyzed and compared were: untreated and IFN-α-treated peripheral blood EPCs/CACs from healthy control and lupus patients (n=6 in each group); as well as untreated and IFN-α-treated bone-marrow EPCs from healthy controls (n=5 in each group). The CEL files were normalized in GenePattern pipeline (www.GenePattern.org) using the RMA (Robust MultiChip Average) method and the Human Entrez Gene custom CDF annotation version 10 (http://brainarray.mbni.med.umich.edu/Brainarray/default.asp). Of the 17527 gene IDs (corresponding to the 54675 Affymetrix probesets), the number of genes expressed above the Poly-A Affymetrix control expression baseline (negative controls) and used for further analyses were respectively: 15700, 15186 and 15909 in the healthy control peripheral blood cells (n=12), the SLE peripheral blood cells (n=12) and the healthy bone-marrow EPCs (n=10). Normalized data files are available on Gene Omnibus website, under the reference number GSE23203 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE23203). Statistical paired analyses were performed using Significance Analysis of Microarrays (SAM) comparing control IFN-α-treated to untreated peripheral blood cells, lupus IFN-α-treated to untreated peripheral blood cells and healthy IFN-α-treated to untreated bone-marrow EPCs. The significantly regulated genes between the groups (q-value <0.05 depicting the False Discovery Rate) were analyzed by building biological literature-based networks using Genomatix Bibliosphere Pathway Edition software (www.genomatix.de) (47, 48). Canonical pathways were analyzed using the Ingenuity Pathway Analysis Software (www.ingenuity.com).

Real-time quantitative PCR

Total RNA was transcribed into cDNA using oligo dT and MMLV (Invitrogen, Carlsbad, CA) with 1μg of RNA using a MyCycler thermocyler (Bio-Rad, Hercules, CA), and levels of expression of the following genes were measured: IL-1β: 5′-ATG TCT GGA ACT TTG GCC ATC TT-3′ (forward); 5′-AGA CAA TTA CAA AAG GCG AAG AAG ACT-3′ (reverse); IL-1α: 5′-CCA GAA GAA GAG GAG GTT GGT C-3′ (forward); 5′-CTG CCC AAG ATG AAG ACC AAC CA-3′ (reverse); IL-RN: 5′-CCT CAG ATA GAA GGT CTT CTG GTT AAC-3′ (forward); 5′-ATG CTG ACT CAA AGG AGA CGA TC-3′ (reverse); IL-1R1: 5′-GAA GCC TGA TGT TTC TCT GAC TAA AAT GA-3′ (forward); 5′-AGG TCT TGG AAA AAC AGT GTG GAT ATA AG-3′ (reverse); IL-1R2: 5′-AAA ATT TGC GGG TAT GAG ATG AAC G-3′ (forward); 5′-ACG TCT GCA CTA CTA GAA ATG CTT C-3′ (reverse); VEGF-A: 5′-GGT CTC GAT TGG ATG GCA GTA G-3′ (forward); 5′-CAC CCA TGG CAG AAG GAG GA-3′ (reverse); MX-1: 5′-TACCAGGACTACGAGATTG-3′ (forward) and 5′-TGCCAGGAAGGTCTATTAG-3′ (reverse); β- actin: 5′-CAT CAC GAT GCC AGT GGT ACG-3′ (forward); 5′-AAC CGC GAG AAG ATG ACC CAG-3′(reverse).

Real time PCR was carried out using an ABI PRISM 7900HT (Applied Biosystems) with the following cycling conditions: An initial denaturing/activation at 95°C for 10 minutes followed by 40 cycles of denaturing at 95°C for 30 seconds, annealing at 52°C for 30 seconds, and elongation at 72°C for 30 seconds.

Assessment of serum protein levels, EPC/CAC proliferation and apoptosis

Serum IL-1-receptor antagonist (IL-1RN) and IL-1β were quantified by ELISA (R&D Systems), following the manufacturer’s instructions. EPC/CAC proliferation and caspase 3/7 activation were assessed after 1 day in proangiogenic culture using the XTT cell proliferation Kit (Cayman, Ann Arbor, MI) and the Apo-ONE Caspase 3/7 Assay (Promega, Madison, WI) respectively, following the manufacturers’ instructions.

Assessment of myeloid cell phenotype

After 5 days in proangiogenic culture in the presence or absence of 10ng/ml IL-1β, cells were harvested using Cell Dissociation Buffer (Invitrogen, Carlsbad, CA) and incubated with the following fluorochrome-conjugated anti-human mAbs: anti-CD14, anti-CD11c, anti-class II MHC (Ancell, Bayport, MN) and anti-CD86 (BD Biosciences, Bedford, MA). Immunofluorescence was quantified by FACS with a FACSCalibur, followed by analysis with FlowJo.

Human kidney tissue and immunohistochemistry

Kidney tissue was obtained from 25 renal biopsies from subjects with clinical and histological diagnosis of lupus nephritis (5 with class II, 5 with class III, 5 with class IV, and 5 with class V, according to the new SLE nephritis classification(49)); ANCA-associated vasculitis (n=5) and, for comparison, normal samples from 5 control kidneys from tumor nephrectomies. Relevant clinico-histological parameters are given in Supplementary Table II.

Tissue samples for light microscopy were fixed in 4% buffered paraformaldehyde and embedded in paraffin. Routine stainings were performed according to standard techniques.

For immunoperoxidase staining, sections were rehydrated, then immersed in 10 mM citrate buffer (pH 6.0), treated with microwave irradiation at 500W for 10 min, and cooled at room temperature. After incubation with 0.5% avidin (Sigma Chimica, Gallarate, Milan, Italy) and 0.01% biotin (Sigma), to suppress endogenous avidin-binding activity, 3% H₂O₂ solution was applied to block endogenous peroxidase. After washing, sections were sequentially incubated with the primary Abs CD31, IL-1β and IL-1RN (Abcam, Cambridge, UK), then with the secondary biotinylated Ab (Zymed, HistoLine, Milan, Italy) and with peroxidase-labeled streptavidin (Zymed). Peroxidase activity was detected with 3,5-diaminobenzidine (Sigma), and sections were dehydrated and mounted in Bio Mount (Bio Optica, Milano, Italy). Specificity of Ab labelling was demonstrated by the lack of staining after substituting PBS and proper control immunoglobulins (Zymed) for the primary Ab.

Images were acquired by a Zeiss Axioscope 40FL microscope, equipped with AxioCam MRc5 digital video camera (Carl Zeiss SpA, Arese, Italy). Images were recorded using AxioVision software 4.3 and analyzed by the AxioVision analysis module (Carl Zeiss SpA). A color threshold procedure allowed selective highlighting of the stained areas in gray mode, and the software was programmed to calculate automatically the percentage of the area or the number of particles occupied by staining.

Renal biopsies were included in the study if a number of 9 or more glomeruli were present. For quantitative evaluation, all glomeruli and medium/large vessels in human biopsies and 20 glomeruli and 5 medium/large vessels in control kidneys were considered. As for the tubulo-interstitial compartment, consecutive images were taken at 200X magnification avoiding glomeruli. The whole specimen was scanned in renal biopsies and 20 fields were sampled in control kidneys.

Statistical analysis

For Affymetrix microarray studies, the Significance microarray analysis (SAM) was used to define genes significantly differentially regulated between the studied groups; a q-value below 0.05 was considered significant. For all other studies a paired Student’s T test or, in the case of the ELISA analysis, a Mann-Whitney T test was performed to determine differences; a p-value below 0.05 was considered significant.

RESULTS

IFN-α induces an antiangiogenic signature in control and lupus EPCs/CACs

The demographic and clinical characteristics of the patients and controls studied are included in Supplementary Table I. The effect of IFN-α on peripheral blood EPCs/CACs and on bone-marrow EPCs exposed to proangiogenic stimulation on steady state mRNA levels was assessed with Affymetrix genechip microarrays. IFN-α-treated vs. untreated control and lupus cells exposed to angiogenic stimulation showed no differential expression in IFN-α mRNA but did show a significant up-regulation of type I IFN-inducible genes (Table I).

Table I.

Affymetrix microarray expression data of studied genes in peripheral blood EPCs/CACs and bone marrow EPCs

			A. IFN-α-treated compared to non treated healthy control peripheral blood EPCs/CACs		B. IFN-α-treated compared to non treated SLEbreak/> peripheral blood EPCs/CACs		C. IFN-α-treated compared to non treated healthy control bone marrow EPCs
Entrez gene ID	Gene symbol	Gene name	Fold- change	q-valuea	Fold- change	q-valuea	Fold- change	q-valuea
Interferon genes
3439	IFN-α1	Interferon alpha 1	1.29	0.243	1.08	0.242	1.06	0.999
3440	IFN-α2	Interferon alpha 2	N/A	N/A	N/A	N/A	N/A	N/A
3454	IFNAR1	interferon (alpha, beta and omega) receptor 1	0.79	0.005*	0.79	0.002*	0.80	0.144
3455	IFNAR2	interferon (alpha, beta and omega) receptor 2	1.05	0.471	1.12	0.079	0.93	0.506
Interferon inducible genes
4599	MX1	Myxovirus resistance 1, interferon- inducible protein p78	2.74	0.000*	6.00	0.000*	6.09	0.003*
10561	IFI44	Interferon- induced protein 44	1.57	0.000*	3.42	0.000*	2.35	0.021*
10964	IFI44L	interferon- induced protein 44-like	1.56	0.001*	9.92	0.000*	3.75	0.020*
3434	IFIT1	interferon- induced protein with tetratricopeptide repeats 1	2.16	0.000*	8.44	0.000*	9.84	0.000*
Interleukin 1 and receptors
3552	IL-1α	Interleukin 1, alpha	0.50	0.003*	0.43	0.001*	0.64	0.214
3553	IL-1β	Interleukin 1, beta	0.70	0.009*	0.62	0.001*	0.84	0.162
3557	IL-1RN	Interleukin 1 receptor antagonist	4.18	0.000*	10.98	0.000*	3.37	0.024*
3554	IL-1R1	Interleukin 1 receptor, type 1	0.66	0.003*	0.76	0.007*	0.86	0.414
7850	IL-1R2	Interleukin 1 receptor, type 2	1.63	0.002*	1.48	0.008*	1.27	0.386
3556	IL1RAP	Interleukin 1 receptor accessory protein	0.75	0.009*	1.05	0.211	0.98	0.999
Markers of apoptosis
836	CASP3	Caspase 3	1.43	0.002*	0.95	0.000*	1.16	0.084
840	CASP7	Caspase 7	1.30	0.001*	2.56	0.000*	1.58	0.012*
Other genes
7422	VEGFA	Vascular endothelial growth factor A	0.66	0.002*	0.78	0.002*	1.04	0.999
5175	PECAM 1	platelet/endotheli al cell adhesion molecule	1.10	0.091	1.05	0.242	0.88	0.098

^aA q-value below 0.05 was considered as significant (in bold and with asterisk); N/A: not expressed above the Affymetrix control baseline.

For the healthy control EPCs/CACs, a total of 2850 genes (q-value <0.01) were found to be differentially regulated by IFN-α (Supplementary Table III). Assessing the association of these transcripts with canonical pathways using Ingenuity Pathway Analysis (IPA) identified interferon signaling as the top regulated pathway in IFN-α treated cells compared to non-treated (p-value=3.98E-05)(Supplementary Table IV). Transcriptional network analysis integrating differentially regulated mRNAs with literature mining and automated promoter analysis (Bibliosphere software suite, Genomatix), highlighted IL-1β and VEGF-A among important regulatory nodes (Supplementary Figure 1). IL-1β and VEGF-A were significantly down-regulated in IFN-α-treated control EPCs/CACs when compared to untreated cells (0.7 fold for both genes, q value < 0.05) (Table IA).

Based on these results, we verified if IL-1β was also regulated in lupus EPCs/CACs after IFN-α exposure. Affymetrix microarray analysis defined 1631 genes significantly up and down-regulated by exogenous IFN-α, with a fold-change ≥ 0.15 between the untreated and treated lupus cells (q-value <0.05) (Supplementary Table V). Interferon signaling also appeared as the top regulated pathway (p-value=5.75E-07)(Supplementary Table V). Similar, to the IFN-α-treated control cells, the literature-based network analysis of those genes highlighted a major IL-1β node. This IL-1 β node and the regulated gene nodes with a direct edge to IL-1β (201 genes in total) are displayed in Figure 1 and listed in the Supplementary Table VI. IL-1β was down-regulated 0.6-fold in IFN-α-treated SLE cells when compared to untreated SLE cells (Table IB). IL1-RN is a member of the IL-1 family that binds to IL-1 receptors without inducing a cellular response, thereby antagonizing the effects of IL-1α and IL-1β (50, 51). IL1-RN mRNA expression was up-regulated 11-fold in SLE cells after treatment with IFN-α, while IL-1α was down-regulated (0.43-fold) in the IFN-treated cells (Table IB, q-value <0.05). The expression of IL1-RN gene was doubled in SLE compared to the healthy control EPCs/CACs (fold-change 11.0 vs. fold-change 4.2, respectively) (Table IA and B). Similar to what was found in the control cells, treatment with IFN-α led to decreased expression of the proangiogenic molecule VEGF-A(52) (fold-change 0.78) in the lupus cells (Table IB). Thus, microarray analysis of EPCs/CACs from control and SLE patients exposed to proangiogenic stimuli identified an anti-angiogenic signature secondary to type I interferon treatment and a more pronounced antiangiogenic response in SLE than in control cells in response to this cytokine.

Figure 1. Transcriptional network displaying IL-1β as a major node regulated by IFN-α on EPCs/CACs.

Literature-based Genomatix Bibliosphere network was performed. The network generated from the regulated gene list between non-treated and IFNα-treated SLE PBMCs displayed IL-1β as one of the main nodes (1631 genes with a fold-change ≥ 0.15 and a q-value <0.05) (n=6 in each group). 1208 genes passed the Bibliosphere filter criterion, based on the co-citation of genes in PubMed abstract sentence linked by a function word (B2 filter). Only the 201 IL-1β directly connected regulated genes are displayed. The color code from red to blue indicates from up-to down regulation of genes in IFNα-treated SLE PBMCs compared to non-treated SLE PBMCs.

Healthy CD133+ bone marrow EPCs were found to have 710 genes (q-value <0.05) (Supplementary Table VII) regulated by IFN-α. Again the interferon signaling pathway was identified as the top regulated one (p-value=6.76E-10, IPA) (Supplementary Table VII). Although the IFN-α and β mRNA expression were not modified, the interferon inducible genes showed significant up-regulation in IFN-α-treated compared to untreated bone marrow cells (fold-changes from 2.35 to 9.84) (Table IC). As observed in the healthy control and lupus PBMCs, IL1-RN mRNA was 3.4 fold increased when cells were exposed to IFN-α compared to non-exposed cells.

Analysis by real time PCR confirmed that IL-1β, IL-1R1, and VEGF-A mRNA levels were significantly down-regulated and IL-1RN and the decoy receptor IL-1R2 significantly up-regulated in IFN-α-treated control and lupus peripheral blood EPCs/CACs when compared to untreated cells (p<0.05) (Figure 2A). Since higher levels and/or increased sensitivity to type I IFNs have been previously reported in SLE (53, 54), we then assessed if similar changes in IL-1 related molecules were seen at the protein level in vivo. A significant increase in circulating IL1-RN levels could be detected in the serum of SLE patients compared to healthy controls (p<0.02) (Figure 2B). Both circulating IL-1β and IL-1α protein levels were below the limit of detection in SLE and control serum (data not shown). We had previously reported that SLE patients have decreased serum levels of VEGF-A compared to healthy controls, further supporting the microarray and real-time PCR data (11).

Figure 2. IL-1 family members and VEGF-A are regulated by IFN-α in lupus and control EPC/CACs.

A. Regulation of IL-1β, IL-1R1, IL-1R2, IL-1RN and VEGF-A mRNA levels by IFN-α was confirmed by real-time PCR in EPCs/CACs from healthy controls and SLE patients. Results represent the mean ± SEM log fold change of IFN-α-treated over untreated cells. All treated samples were significantly altered with a p<0.05 when compared to untreated samples; (SLE n=11; control=17). B. SLE patients display elevated levels of IL-1RN in plasma (p=0.0116). Results represent mean± SEM plasma levels of IL-1RN in 35 SLE patients and 10 healthy controls.

IL-1β restores the capacity of lupus EPCs/CACS to differentiate into mature endothelial cells

We had previously reported that lupus EPCs/CACs, in contrast to control cells, fail to effectively differentiate into endothelial-like cells. Lupus EPCs/CACs cannot form a mature endothelial monolayer when cultured under proangiogenic stimulation on fibronectin-coated wells. We previously showed that IFN-α blockade restores the functional capacity of lupus EPCs/CACs (11). Given that IFN-α downmodulated IL-1 and VEGF pathways, we next tested if addition of IL-1β and/or VEGF-A to lupus EPC/CAC cultures would restore the capacity of these cells to differentiate into a mature endothelium. To this, end, recombinant IL-1β and/or VEGF-A were added to proangiogenic media and the number of mature endothelial cells was counted as determined by the uptake of ac-LDL and the binding of the UEA-1 lectin in control and SLE cells after 14-21 days in culture. We again confirmed a significant decrease in the capacity of SLE EPCs/CACs to differentiate to mature endothelial cells when compared to control samples. Addition of IL-1β significantly increased the number of mature endothelial cells by 3-fold in lupus cell cultures when compared to media alone (p=0.010) (Figure 3A-B). VEGF-A also increased the number of mature endothelial cells by 2-fold (p=0.045) (Figure 3A-B). The specificity of the response to molecules regulated by IFN-α was demonstrated, as treatment with other proangiogenic factors (IL-6 and TNF)(55-57) not regulated by IFN-α treatment in the gene expression analysis (data not shown), did not result in any improvement in EPC/CAC function (Figure 3A-B). In addition, the beneficial effect of IL-1 and VEGF-A was only seen in SLE samples but not in controls (Figure 3A-B).

Figure 3. IL-1β abrogates the abnormal phenotype of lupus EPC/CACs.

A. Addition of recombinant IL-1β to lupus EPC/CAC cultures improves the ability of these cells to differentiate into mature endothelial cells to levels comparable to those of healthy controls (**p<0.01, n=24),while addition of recombinant VEGF-A resulted in a more modest but significant improvement (*p<0.05, n=24). Addition of IL-6 or TNF-α to the same cultures had no beneficial effect when compared to untreated cultures (n=9). Control cells showed no significant improvement in their capacity to differentiate into mature endothelial cells with any of the treatments mentioned above (n=7, n=5, n=3, n=3 respectively). Results are expressed as a ratio of the number of endothelial cells/ high power field to the number of untreated lupus endothelial cells/high power field; *p<0.05. B. Representative images displaying the effects of these proteins on EPC/CAC cultures from 2 SLE patients at 2-3 weeks. Images acquired at 100X total magnification. Bar indicates 200 um. Mature endothelial cells show coexpression of FITC-UEA-1(green) and uptake of DiI-Ac-LDL (red).

IL-1β abrogates the defects in proliferation and viability of lupus EPCs/CACs

Various mechanisms may lead to improvements in lupus EPC/CAC function by addition of IL-1β, including increased proliferation or increased viability including modulation of caspase activity. IL-1β has been shown to increase proliferation of murine EPCs(58), rat bone marrow cells (59), and human T (60-62), and B cells (61, 63). IL-1β is also able to inhibit apoptosis and promote survival of monocytes(64) and granulocytes(65). With evidence for increased proliferation and viability possibly playing a role in the improvement in EPC/CAC function, we tested if supplementation with IL-1β and/or VEGF-A promoted increases in proliferation and/or protection from apoptosis. SLE and control cells were cultured with proangiogenic media in the presence or absence of IL-1β and/or VEGF-A for 24 hours. Lupus EPC/CAC proliferation was significantly increased by the addition of recombinant IL-1β (p=0.023) or VEGF-A (p=0.019) when compared to cells cultured with media alone (Figure 4A). Treatment with a combination of IL-1β and VEGF-A did not show a synergistic or additive effect on proliferation compared to cells treated with IL-1β or VEGF-A only (Figure 4A). When control EPCs/CACs where treated with IL-1β there was a trend of increased proliferation (p=0.084). Control cells treated with VEGF-A alone did not show any improvement in proliferation, but the combination of Il-1β and VEGF-A did significantly (p=0.016) improve proliferative capacity (Figure 4A).

Figure 4. IL-1β increases proliferation and inhibits apoptosis of lupus EPC/CACs.

A. Improvements in EPC/CAC proliferation by IL-1 and VEGF-A supplementation. Results represent mean ± SEM absorbance in duplicate wells; * p<0.05 (SLE: n=5 and controls: n=8). B. IL-1β inhibits activation of caspase 3 and 7 in lupus but not control EPC/CACs. Results represent mean fluorescent value of treated samples/mean fluorescent value of SLE media alone ± SEM; (SLE=5; control=8); * p<0.05.

Caspase 3 and 7 play a central role in the execution-phase of apoptosis (66) and activation of these molecules is used as a marker of programmed cell death. The mRNA levels of these caspases were significantly upregulated by IFN-α treatment (Table I) and caspase 3 was also a subnode in the literature based-Bibliosphere network (Figure 1). Untreated lupus EPC/CACs displayed significant increases in caspase 3/7 activation in the presence of media alone, when compared to control cells (Figure 4B) indicating increased apoptosis. Treatment of lupus EPC/CAC cultures with recombinant IL-1β significantly inhibited caspase 3/7 activation (p=0.045) (Figure 4B), while treatment with VEGF-A did not have any effect. A combination of recombinant IL-1β and VEGF-A decreased caspase 3/7 activity to a similar level as IL-1β treatment alone. These results indicate that both IL-1β and VEGF-A can improve EPC/CAC proliferation but only IL-1β has an effect in inhibiting enhanced apoptosis on these cells. In contrast, neither IL-1β nor VEGF-A addition had any effect on healthy control EPC/CAC apoptosis (Figure 4B).

IL-1-β decreases lupus DC differentiation

We previously reported that, in EPC/CAC cultures, IFN-α alters the expression of surface developmental markers on myelomonocytic cells, consistent with skewing the differentiation of these cells from angiogenic CACs to IFN-α–derived DCs. Indeed, IFN-α induces a significant reduction in DC-SIGN⁺CD14⁺ cells, CD86 up-regulation, and overall down-regulation of CD14 (11). To address if addition of IL-1β could inhibit this phenomenon and allow myeloid lupus cells to favor proangiogenic differentiation rather than DC differentiation, surface markers of mature DCs were examined in 5-day cultures of lupus PBMCs subjected to proangiogenic stimulation in the presence or absence of IL-1β. Addition of IL-1β led to a significant downregulation of the mature DC markers CD86 and MHC class II on SLE patients and a similar trend was observed in healthy (Table II). There was also a decrease in CD11c⁺ CD14⁻ cell population, indicating a decrease in total DCs. Additionally, expression of CD14 was increased in cultures, suggesting an increase in myeloid CACs. These results indicate that the skewing of myeloid cells into nonangiogenic pathways by IFN-α is mediated, at least in part, by downregulation of IL-1 pathways.

Table II.

Effect of IL-1β on myelomonocytic cell differentiation

	Controls			Lupus
	Untreated	+ IL1β	p-value	Untreated	+ IL1β	p-value
CD14	61.4 ± 6.23	75.9 ± 6.63	0.018	51.7 ± 6.82	75.3 ± 4.68	0.002
CD11c	51.8 ± 4.70	50.3 ± 6.56	0.779	46.6 ± 6.64	41.7 ± 7.61	0.492
DC-SIGN	21.7 ± 6.98	12.2 ± 1.87	0.318	20.9 ± 6.68	5.8 ± 1.82	0.033
CD86	57.5 ± 3.29	36.1 ± 7.21	0.117	40.5 ± 5.93	28.4 ± 5.62	0.079
MHC class II	55.2 ± 12.25	35.7 ± 16.28	0.037	22.0 ± 4.75	7.7 ± 3.16	0.001
CD11c+CD86+	44.8 ± 4.16	19.9 ± 1.72	0.011	32.1 ± 4.93	14.1 ± 3.52	0.006
CD14+DC−SIGN+	20.3 ± 6.65	11.1 ± 0.81	0.282	18.8 ± 7.22	5.7 ± 1.98	0.065
CD11c+CD14−	1.03 ± 0.31	0.18 ± 0.11	0.024	1.8 ± 0.46	0.6 ± 0.28	0.002

Numbers indicate mean percentage of positive cells ± standard error of the mean; p-value < 0.05 is considered significant.

IFN-α regulates IL-1 pathways in EPCs/CACs through JAK-STAT modulation

Various pathways potentially involved in the regulation of IL-1 family members and VEGF were investigated. Hypoxia-inducible factor (HIF)-1α is the main regulator of VEGF-A expression, mainly by posttranslational regulation via proteasomal degradation and stabilization by von Hippel-Lindau factor (67). HIF-1α is a transcription factor and 117 genes, including IL1-β and VEGF-A, of the 2850 in SLE EPCs/CACs have an ischemia response element contained in their promoter. A significant decrease of HIF-1α mRNA expression was detected in lupus EPC/CACs treated with IFN-α; the significance was not reached in control cells (Table III). However, there was no significant correlation between mRNA levels of HIF-1α and VEGF-A in each group studied (data not shown) indicating that it is unlikely that the changes in HIF-1α steady state mRNA levels induced by IFN-α were the main determinants of VEGF-A downregulation in EPCs/CACs.

Table III.

Affymetrix microarray expression data of STAT, HIF1A, HuR and HuR target genes in peripheral blood EPCs/CACs and bone marrow EPCs

			A. IFN-α-treated compared to non treated healthy control peripheral blood EPCs/CACs		B. IFN-α-treated compared to non treated SLE peripheral blood EPCs/CACs		C. IFN-α-treated compared to non treated healthy control bone marrow EPCs
Entrez gene ID	Gene symbol	Gene name	Fold- change	q-valuea	Fold- change	q-valuea	Fold- change	q-valuea
STAT genes
6772	STAT1	Signal transducer and activator of transcription 1	1.06	0.198	1.45	0.031*	2.61	0.000*
6773	STAT2	Signal transducer and activator of transcription 2	1.43	0.000*	1.69	0.000*	1.86	0.000*
6774	STAT3	Signal transducer and activator of transcription 3	1.02	0.480	1.54	0.001*	1.45	0.123
6775	STAT4	Signal transducer and activator of transcription 4	1.10	0.025*	1.49	0.036*	1.23	0.445
6776	STAT5A	Signal transducer and activator of transcription 5A	1.25	0.001*	1.21	0.003*	104	0.619
6777	STAT5B	Signal transducer and activator of transcription 5B	0.97	0.329	104	0.217	1.01	0.999
6778	STAT6	Signal transducer and activator of transcription 6	1.19	0.014*	1.22	0.000*	1.06	0.623
HIF1-α, HuR and HuR target genes
3091	HIF1A	Hypoxia factor 1A	0.96	0.243	0.96	0.036*	0.94	0.456
1994	ELAVL1	ELAV (embryonic lethal, abnormal vision Drosophila)-like 1, Hu antigen R	0.88	0.026*	0.85	0.002*	0.84	0.070
153	ADRB1	Adrenergic beta- 1 receptor	1.08	0.347	105	0.228	1.08	0.627
8900	CCNA1	Cyclin A1	450	0.000*	8.32	0.000*	6.12	0.000*
891	CCNB1	Cyclin B1	0.91	0.076	0.89	0.041*	0.90	0.150
595	CCND1	Cyclin D1	0.67	0.001*	146	0.012*	1.04	0.999
2353	FOS	FBJ murine osteosarcoma viral oncogene homolog	0.74	0.009*	0.67	0.006*	1.16	0.627
4609	MYC	V-myc myelocystomatos is viral oncogene homolog	0.86	0.102	1.07	0.999	0.88	0.999
3562	IL3	Interleukin 3	0.91	0.295	0.79	0.999	1.03	0.999
4613	MYCN	V-myc myelocystomatos is viral related oncogene, neuroblastoma derived	1.03	0.451	1.06	0.201	1.01	0.999
4763	NF1	Neurofibromin 1	05	0.368	0.86	0.016*	1.23	0.480
1026	CDKN1 A	Cyclin-dependent kinase inhibitor 1A	1.002	0.502	1.27	0.002*	1.07	0.548
5055	SERPIN B2	Serpin peptidase inhibitor, clade B, member 2	1.07	0.283	1.18	0.193	0.78	0.232
7124	TNF	Tumor necrosis factor	1.03	0.476	1.49	0.001*	1.09	0.651
7422	VEGFA	Vascular endothelial growth factor A	0.66	0.002*	0.78	0.002*	1.04	0.999
2596	GAP43	Growth associated protein 43	N/A	N/A	103	0211	0.97	0.548

^aA q-value below 0.05 was considered as significant (in bold and with asterisk); N/A: not expressed above the Affymetrix control baseline.

Regulation of VEGF-A occurs not only at the transcriptional level but also postranscriptionally(68) and includes mRNA stabilization by the Hu antigen R (HuR; or ELAV [embryonic lethal, abnormal vision]), a ubiquitously expressed RNA-binding protein(68). To determine whether HuR mRNA stabilization may play a role in VEGF-A changes observed secondary to IFN-α, we analyzed mRNA expression of known HuR targets(69) in the microarray expression profiles. Of the 14 HuR targets examined, only 4 (including VEGF-A mRNA) were significantly reduced in lupus EPCs/CACs treated with IFN-α, while 3 were reduced in the IFN-treated control EPCs/CACs (Table III). While HuR mRNA was significantly reduced in both control and SLE EPC/CACs upon exposure to IFN-α, several HuR targets were either upregulated or unchanged after IFN-α treatment. Overall, these results suggest that HuR-dependent RNA degradation is unlikely to be an important mechanism for VEGF-A reduction induced by IFN-α. Further, control bone marrow CD133+ EPCs did not show any significant HIF1A or HuR mRNA expression changes in response to IFN-α (Table III).

The large family of IFN-α/β proteins all bind to a single type of receptor, which is composed of two chains: IFNAR1 and IFNAR2. The intracellular domain of IFNAR1 associates with a member of the JAK kinase family, Tyk2, whereas IFNAR2 associates with Jak1. The major substrates for tyrosine phosphorylation subsequent to IFN receptor binding are members of the STAT family of transcription factors. These proteins are normally latent and reside in the cytoplasm in unstimulated cells. Once phosphorylated, STAT1 and STAT2 dimerize and assemble with IRF-9 to form the multimeric transcription factor ISGF3. ISGF3 binds to the IFN-stimulated response element of IFN-stimulated genes in the nucleus and activates their transcription(70). From the microarray analyses, the JAK/STAT pathway was one of the most significantly regulated canonical pathways in control and SLE EPCs/CACs after IFN-α treatment (p-value = 0.023 and 0.017 respectively). Indeed, SLE peripheral blood EPCs/CACs exposed to IFN-α significantly upregulated STAT 1, 2, 3, 4, 5A and 6 while control EPCs/CACs significantly upregulated STAT 2,4,5A and 6 (Table IIIA and B) and control bone marrow CD133+ EPCs significantly upregulated STAT 1 and 2 in response to IFN-α treatment (Table IIIC). Confirming the array data, when control and SLE peripheral blood EPC/CACs allowed to differentiate for 3 days were stimulated with IFN-α for 30 minutes, there were significant increases in phosphorylation of both STAT 2 and 6 (p<0.05, Figure 5). We then performed subanalysis on EPCs and CACs; In the EPCs, STAT 2 and 6 were significantly activated upon IFN-α stimulation (p=0.04 and 0.02, respectively, Figure 5A), while in CACs there was a significant activation of STAT2 (p<0.05, Figure 5B) and a trend towards STAT6 activation. We then tested if JAK inhibition would abrogate the downregulation of proangiogenic molecules seen in IFN-α treated cells. Indeed, the pan-JAK inhibitor pyridone 6 induced significant upregulation of IL-1α, IL-1β, IL-1R1 and VEGF-A (p=0.02, 0.001, 0.01 and 0.01, respectively, Figure 5C) and downregulation of IL-1RN and the canonical type I IFN-inducible gene MX1 (p=0.06 and 0.05, respectively, Figure 5D) in both IFN-α- treated control and lupus EPCs/CACs. In contrast, there was no change in level of expression of IL-1R2 mRNA after JAK inhibition.

Figure 5. IFN-α activates STAT2 and 6 on EPCs/CACs and JAK inhibition leads to downregulation of the antiangiogenic signature in SLE.

Phosphorylation of STAT 2 and STAT6 was detected by FACS in EPCs and CACs. Stimulation with IFN-α leads to STAT2 and 6 activation in EPCs (A) and STAT2 activation in CACs. (B). Results represent mean % ±SEM of cells positive for STAT phosphorylation in 7 SLE and 4 control subjects. C. JAK inhibition induces upregulation of IL-1α, IL-1β, IL-1R1 and VEGF-A mRNA and decreases mRNA levels of IL-1RN and MX1 (D). Results represent average fold-change+SEM over vehicle-treated cells; SLE n=10 and control n=5; *p<0.05.

The PI3 kinase pathway is also activated by IFN-α in various cell types and PI3K interacts with the JAK-STAT signaling pathway and aids in the activation of a limited number of genes(71). To assess if PI3K pathway was also involved in the induction of an antiangiogenic signature in IFN-treated EPCs/CACs, we performed similar experiments using a PI3K inhibitor. In contrast to what was observed with a JAK inhibitor, PI3K inhibition did not induce significant changes in any of the genes examined. Overall, these results indicate that the antiangiogenic signature induced by IFN-α on EPCs/CACs is signaled through JAK-STAT pathways.

Decreased endothelial density and altered IL-1 pathways are observed in SLE in vivo

While our group and others have suggested decreased vasculogenesis in human SLE(11, 14) and murine lupus models also show similar abnormalities in vivo(23), there was no previous evidence on the occurrence of this phenomenon in vivo in humans. To this end, we proceeded to examine the glomeruli and renal blood vessels of lupus nephritis biopsies and compared blood vessel density and protein expression of IL-1RN and VEGF-A in these compartments to those of control kidneys as well as renal biopsies from patients with another immune-mediated disease, ANCA-positive vasculitis (Supplementary Table II). Immunohistochemistry analysis revealed that the glomerular compartment of patients with lupus nephritis had a significant decrease in VEGF-A expression when compared with control biopsies or biopsies from patients with vasculitis and similar degree of kidney function. This was observed in classes II-V lupus nephritis but was most pronounced in classes IV and V (Figure 6). Confirming the in vitro data on IFN-treated EPCs/CACs and the serum findings in SLE, IL-1RN was detected in the renal blood vessels of patients with lupus nephritis (particularly class V) but not in the blood vessels of control patients or patients with ANCA-positive vasculitis. IL-1β was not detected by IHC in any of the tissues examined. The density of the peritubular and interstitial vessels was evaluated. CD31 expression, used as a marker of endothelial cell density, was significantly decreased in the glomerular compartment of lupus nephritis biopsies (particularly class V) when compared to control biopsies and the levels of this molecule in the renal blood vessels were significantly decreased in all subsets of lupus nephritis compared to control or ANCA-positive vasculitis (Figure 6). These results indicate that decreased vasculogenesis is also occurring in vivo in SLE and that changes in molecules involved in IL-1 signaling and VEGF-A are also observed at the tissue level in this disease.

Figure 6. Kidneys from patients with lupus nephritis display decreased VEGF-A, increased IL-1RN and decreased glomerular and capillary density.

A, and D. Representative photographs show that expression of IL-1RN is negative in the endothelium and vessel wall of a control kidney biopsy. In contrast, the vascular endothelium is clearly stained in a case of class V SLE nephritis and numerous positive cells are present in the vessel wall, particularly in the external layers. A mild vascular positivity for IL1RN can be detected in a medium sized vessel from a case of renal ANCA⁺ vasculitis (Immunoperoxidase, 400X). B and E. Evident loss of podocyte staining for VEGF-A in glomeruli from class IV lupus nephritis, as compared with control kidneys and ANCA⁺ vasculitis. (Immunoperoxidase, 200X). C and F. Compared to glomeruli from a control kidney and a case of vasculitis, glomerular expression of CD31 is significantly reduced in a case of SLE class III lupus nephritis. (Immunoperoxidase, 400X).

DISCUSSION

While it is widely accepted that SLE patients exhibit greater propensity to develop CV complications (5, 72), the mechanisms leading to this enhanced risk remain unclear. Our group and others have proposed that IFN-α may play a prominent role in vascular damage in SLE, at least in part, by interfering in the balance between blood vessel damage and repair leading to a dysfunctional endothelium(11, 14). Indeed, abnormalities in vasculogenesis have been proposed to play an important role in the development of atherosclerosis and organ damage (including renal failure) in various diseases (31, 73-76).

While IFN-α is a potent antiangiogenic factor (77-79), the exact pathways by which this molecule regulates EPC/CAC gene expression had not been characterized(39, 80). In this study, we demonstrate that IFN-α exerts its antiangiogenic effects on EPCs/CACs through the modulation of molecules relevant to IL-1 function and signaling and downregulation of VEGF-A. Further, these IFN-induced pathways are operational in vivo in serum and tissue of SLE patients.. A previous study reported high levels of IL-1RN in SLE serum and correlation with lupus disease activity(81). Lupus PBMCs are also less capable of releasing IL-1β in response to stimulation(82). However, a link between these abnormalities and IFN-α in SLE and their potential role in vascular damage in this disease had not been reported. Our results also support studies performed in other populations, where exogenous type I IFNs induce IL-1 downregulation and IL-1RN increases in patients with viral hepatitis C, as well as previous reports of the effect of IFN-α on healthy control myeloid cells in vitro, where IL-1RN is upregulated without inducing concomitant increases in IL-1(83-85). Supporting the hypothesis that other type I IFNs may promote similar modulation, IFN-β also has effects on IL-1 family members; an indirect effect by inhibiting the ability of T cells to stimulate monocytes to synthesize IL-1β and a direct effect by upregulating IL-1RN synthesis in the latter(86, 87).

Previous studies on hepatocytes report that exposure to type I IFNs leads to STAT activation and stimulation of the formation of STAT2: STAT6 complexes, subsequently leading to enhanced IL-1RN synthesis (88). In this study, we confirmed that IFN-α led to STAT 2 and 6 phosphorylation in control and lupus EPCs/CACs and that inhibition of JAK-STAT signaling leads to abrogation of the antiangiogenic responses. The effects of JAK inhibition on IL-1 and VEGF-A mRNA expression are similar to what has been seen in other angiogenesis models(89) and in different cell types including macrophages(90).

In contrast to the changes observed in peripheral blood EPC/CAC cultures, bone marrow EPCs responded to IFN-α primarily by significantly upregulating IL1-RN expression, but displayed no significant changes in other IL-1 related molecules or in VEGF-A, although a similar trend was observed. These results indicate that part of the changes seen in peripheral blood were secondary to direct alterations in the myeloid CAC compartment and/or that EPCs present in peripheral blood represent a distinct subset responding in different ways to exogenous signals than those EPCs that have not left the bone marrow. Future studies should investigate this possibility.

While IL1-α and IL-1β are considered predominantly proinflammatory molecules produced in response to infection or cellular damage, they also have strong proangiogenic properties and effects on murine and human EPCs(58, 91, 92) (93). Further, recombinant IL-1RN (anakinra) has been used as antiangiogenic therapy in various diseases. Part of the effects of IL-1β on angiogenesis are mediated through the regulation of various proangiogenic factors, including VEGF-A (94, 95). IL-1 can induce increased VEGF-A synthesis (96) and genes in a cluster of IL 1-related molecules, including IL-1RN, may regulate VEGF-induced angiogenesis (97). This may explain, at least in part, why the strongest responses in improvement of EPC/CAC function in the SLE cultures were observed with addition of IL-1β, as this molecule may have promoted upregulattion of endogenous VEGF-A expression. IL-1β also plays a key role in ischemia-induced neovascularization by mobilizing endothelial precursor cells in a VEGF-dependent manner as well as by upregulating VEGF and VEGFR-2 expression on endothelial cells (98).

The role of proinflammatory stimuli on the viability and function of EPCs is conflicting. Some studies have reported enhanced EPC death in inflammatory conditions (99) while others report increased endothelial differentiation (59). Our study does not address the question of the effects of IL-1β modulation in a proinflammatory setting but, rather, in a proangiogenic setting in the absence of exogenous proinflammatory mediators. In this case, IL-1β (but not other proinflammatory cytokines) can induce substantial improvements in EPC/CAC differentiation and survival (55). In addition, IL-1β inhibited the skewing of myeloid cells present in the proangiogenic culture from CACs to mature DCs(11, 100), a phenomenon previously reported to be induced by IFN-α. While this may seem counterintuitive as IL-1β is widely used to promote DC maturation, it is generally used in combination with other proinflammatory cytokines (101-104). Previous studies have reported that IL-1β can impair maturation in DCs treated with rapamycin(105) but,to our knowledge, no other study had examined the isolated effects of IL-1β treatment on DCs or differentiating monocytes.

While decreased IL-1 and increased IL-1RN may theoretically support a phenotype that protects the vasculature, given the anti-inflammatory effects, it is reasonable to speculate that in a disease like SLE where there is ongoing endothelial cell damage and death(12), a cytokine profile that promotes an antiangiogenic response would be overall deleterious and could accelerate atherosclerosis development. A variety of vascular insults in SLE that lead to disease flares(106, 107) (108) in conjunction with increased levels of type I IFNs, may lead to periods of endothelial damage followed by aberrant repair due to decreased IL-1 and VEGF-A and increased IL-1RN. This could allow the initiation and expansion of vascular lesions during these flares.

We have also addressed if these findings are operational in vivo in SLE, potentially due to increased circulating and/or tissue IFN-α levels as well as increased sensitivity to this molecule by SLE cells (53, 109). Kidneys from SLE patients display repression of VEGF-A and induction of IL-1RN and this correlates with decreased renal vascular density and vascular rarefaction. These results are consistent with our hypothesis of impaired renal vasculogenesis in SLE and may help to increase our understanding on the mechanisms leading to loss of renal function in this disease. Progressive renal disease is characterized by glomerulosclerosis and interstitial fibrosis and although many studies have focused on the mechanisms underlying excessive deposition of extracellular matrix, there is increasing evidence of a vascular component as a key driver in the pathogenesis of renal scarring (110, 111). Loss of glomerular and peritubular capillaries is strongly associated with the progression of chronic kidney disease to end-stage renal disease in other populations and animal models(112). VEGF-A is constitutively expressed in human and rodent kidneys(113, 114) and rodent models of progressive renal failure show loss in glomerular and tubular VEGF-A coinciding with loss of capillaries and the subsequent development of glomerulosclerosis and interstitial fibrosis (115). In humans, pharmacologic inhibition of VEGF-A during cancer treatment can lead to proteinuria and worsening kidney function (116). Glomerular and peritubular capillary rarefaction is an important feature of disease progression in other conditions including diabetic nephropathy (117). Recently, a role for VEGF-A contributing to the maintenance of glomerular or peritubular capillaries and renal tissue survival has been proposed in diabetes(118). Interestingly, other renal diseases including minimal change disease or idiopathic membranous nephritis have not shown the reduction of tubulointerstitial VEGF-A, indicating that renal damage or proteinuria per se do not lead to downregulation of this molecule(118). This was confirmed in our study, where patients with ANCA-positive vasculitis and renal damage did not display downregulation of VEGF-A or renal capillary drop-out. Supporting our data, a recent study shows that decreased VEGF in SLE renal tissue may predict short-term loss of kidney function(119).

Various studies have also linked progression of renal disease to certain IL-1RN polymorphisms which are associated to higher IL-1RN levels in the general population (120, 121). However, the exact mechanisms linking renal failure progression to high IL-1RN had not been characterized. Given the observations made in our study, it is possible that the antiangiogenic effect of increased IL-1RN further promotes a reduction in renal capillaries and hampers renal perfusion. Therefore, reduced VEGF-A and IL-1RN (potentially secondary to enhanced exposure to type I IFN levels in the kidney) may contribute to decreased endothelial survival and angiogenesis and to progression of kidney damage. This hypothesis is supported by recent observations that type I IFNs produced by resident renal cells promote end-organ disease in autoantibody-mediated glomerulonephritis(122) and that systemic administration of IFN-α to lupus murine models worsens nephritis(123). While these effects are likely multifactorial, the potential role of this cytokine in leading to microvascular rarefaction and progression of renal failure may play an important role.

IL-1RN has been used in various inflammatory conditions including rheumatoid arthritis and various autoinflammatory diseases(124, 125). The exact role that exogenous IL-1RN could play in overall CV risk in these diseases is unclear. While small numbers of SLE patients have been treated with recombinant IL-1RN without major adverse effects(126, 127), no studies have addressed the effect of this compound on vascular risk in SLE or other systemic autoimmune diseases. Our study adds a note of caution that should be further investigated with regards to the role of IL-1 blockade in conditions where vascular repair is already impaired, including SLE and RA(11, 74).

Overall, we have identified the mechanisms by which IFN-α interferes with EPC/CAC function and may lead to abnormal vascular repair, atherosclerosis progression and loss of renal function in SLE. Future studies should also investigate if similar abnormalities are also present in individuals with other autoimmune diseases associated to increased type I IFN signatures (128, 129), (130, 131), (132, 133).

Abbreviations used

ANCA

Anti-neutrophil cytoplasmic antibody

CACs

Circulating angiogenic cells

CV

Cardiovascular

DCs

dendritic cells

diI

1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate

EPCs

endothelial progenitor cells

IL-1R1

IL-1 receptor 1

IL-1RN

IL-1R antagonist

RA

Rheumatoid Arthritis

SLE

systemic lupus erythematosus

UEA-1

Ulex europaeus agglutinin-1

VEGF-A

vascular endothelial growth factor A