Loss of Id3 Increases VCAM-1 Expression, Macrophage Accumulation, and Atherogenesis in Ldlr–/– Mice
Objective—Inhibitor of differention-3 (Id3) promotes B cells homing to the aorta and atheroprotection in Apoe−/− mice. We sought to determine the impact of loss of Id3 in the Ldlr−/− mouse model of diet-induced atherosclerosis and identify novel Id3 targets in the vessel wall.
Methods and Results—Ex vivo optical imaging confirmed that Id3−/− Ldlr−/− mice have significantly fewer aortic B cells than Id3+/+ Ldlr−/− mice. After 8 and 16 weeks of Western diet, Id3−/− Ldlr−/− mice developed significantly more atherosclerosis than Id3+/+ Ldlr−/− mice, with Id3+/− Ldlr−/− mice demonstrating an intermediate phenotype. There were no differences in serum lipid levels between genotypes. Immunostaining demonstrated that aortas from Id3−/− Ldlr−/− mice had greater intimal macrophage density and C-C chemokine ligand 20 and vascular cell adhesion molecule 1 (VCAM-1) expression compared with Id3+/+ Ldlr−/− mice. Real-time polymerase chain reaction demonstrated increased VCAM-1 mRNA levels in the aortas of Id3−/− Ldlr−/− mice. Primary vascular smooth muscle cells from Id3−/− mice expressed greater amounts of VCAM-1 protein compared with control. Gain and loss of function studies in primary vascular smooth muscle cells identified a role for Id3 in repressing VCAM-1 promoter activation. Chromatin immunoprecipitation demonstrated interaction of E12 with the VCAM-1 promoter, which is inhibited by Id3.
Conclusion—Id3 is an atheroprotective transcription regulator with targets in both B cells and vessel wall cells leading to reduced macrophage accumulation and reduced atherosclerosis formation.
Atherosclerosis remains the leading cause of morbidity and mortality in the United States,1 and further understanding of pathways that protect from the development of atherosclerosis may elucidate potential new therapies.2 Inhibitor of differention-3 (Id3), a member of the helix–loop–helix family of transcription factors, has recently been implicated as an atheroprotective factor.3,4 Id3 is a dominant negative regulator of gene pathways controlled by basic helix–loop–helix proteins, such as the E-proteins E47 and E12.5 These broadly expressed factors have been implicated in regulating growth and differentiation in several cell types, most notably B cells.6 Recent data provide evidence that loss of Id3 leads to reduced aortic B cells and increased development of atherosclerosis in Apoe−/− mice.4
In humans, a single nucleotide polymorphism in Id3 at rs11574 is associated with increased carotid intima media thickness (cIMT) in the Diabetes Heart Study.4 This single nucleotide polymorphism results in a change from an alanine to a threonine in the Id3 C terminus, a region of the protein that has previously been shown to be essential for the dominant negative function of Id3.7 Biochemical studies demonstrated that the Id3 protein encoded by the risk allele had a marked attenuation in dimerization with E12 and antagonisms of E12 function.4 Given the potential implications for Id3 in human atherosclerosis, demonstrating the role of Id3 in atheroprotection in a second model of diet-induced atherosclerosis with the Apoe gene intact, and identifying novel mechanisms whereby Id3 may regulate atherogenesis is of clear significance.
The increased number of macrophages in the aorta of Id3−/− Apoe−/− mice has been linked to the reduced number of aortic B cells also seen in these mice, as adoptive transfer of B cells to B cell–deficient μMT mice led to a reduced number of aortic macrophages.3 Yet, Id3 may repress proinflammatory genes in other cell types. In support of this hypothesis, bone marrow transplantation experiments demonstrated that Id3-mediated atheroprotection was not solely a result of bone marrow–derived cells.3 Thus, further characterization of the atherosclerotic plaque may help elucidate additional mechanisms of Id3-mediated atheroprotection.
Results of the present study demonstrate a step-wise increase in Western diet-induced atherosclerosis in Ldlr−/− mice heterozygous and homozygous for Id3 gene deletion, consistent with the step-wise increase in cIMT in humans heterozygous and homozygous for the Id3 single nucleotide polymorphism at rs11574.4 Similar to the B cell phenotype seen in the Apoe−/− mouse null for Id3, Id3−/− Ldlr−/− mice had decreased aortic B cells. In addition, plaque characterization revealed increased macrophage density deep into the intima in Id3−/− Ldlr−/− mice. Consistent with this finding, C-C chemokine ligand 20 (CCL20) and vascular cell adhesion molecule 1 (VCAM-1) expression were significantly increased in lesions from mice null for Id3. Moreover, loss of Id3 resulted in increased VCAM-1 mRNA and protein expression in vascular smooth muscle cells (VSMC). E12 activated the human VCAM-1 promoter, whereas Id3 antagonized this effect. VCAM-1 promoter activation was significantly increased in VSMCs from mice null for Id3. Finally, we demonstrated E12 binding to the VCAM-1 promoter, although this was inhibited by Id3. Taken together, the present results provide novel evidence for a role for Id3 in suppressing specific chemokine and adhesion molecule expression in vessel wall cells, further implicating Id3 as an important factor regulating atheroprotective pathways.
Materials and Methods
Detailed methods can be found in the online-only Data Supplement. A brief description of each method is provided below.
All animal protocols were approved by the Animal Care and Use Committee at the University of Virginia. Id3−/− mice were bred to the Ldlr−/− background to obtain an atherogenic Id3−/− Ldlr−/− strain.
Determination of Serum Cholesterol Levels
Analysis of Atherosclerosis
For en face analysis, aortas from the heart to the iliac bifurcation were opened longitudinally, pinned, and stained using Sudan IV as previously described.10 For immunohistochemistry, the heart and aortic arch to the left subclavian artery was embedded in paraffin and 5-μm-thick serial sections were generated. Intimal cellularity was assessed using the MOVAT method.11
Optical Imaging of Aortic B Lymphocytes
To detect endogenous B lymphocytes with near-infrared fluorescent imaging, aortas were harvested from 5 chow-fed 8- to 10-week-old Id3+/+ Ldlr−/− mice, 5 Id3−/− Ldlr−/− mice, and 5 μMT mice, and incubated overnight at 4°C in Cy5.5-labeled anti-CD19 solution. Ex vivo fluorescence-mediated tomography quantitative imaging (fluorescence-mediated tomography 2500, VisEn Medical) was performed to determine mean fluorescence for each aorta.
Immunohistochemical Analysis of Atherosclerosis
Paraffin-embedded sections were stained for macrophages and T cells and assessed by fluorescent microscopy. Macrophage area/intima area was measured as the macrophage area within the intima divided by the total area of the intima. We stained for C-X-C chemokine ligand-12 CCL20, VCAM-1, and smooth muscle α-actin using 3,3′-diaminobenzidine staining. Microscopy was performed at the University of Virginia Advanced Microscopy Core with 4× and 10× objectives. 3,3′-diaminobenzidine staining within the intima was measured independently by 2 investigators using ImagePro Plus 7.0 software. Terminal deoxynucleotidyl transferase dUTP nick end labeling was performed to assess cellular apoptosis.
Adoptive Transfer of B Cells and Intimal VCAM-1 Staining
To assess whether B cells impact VCAM-1 expression in atherosclerotic plaque, 45×106 Apoe−/− B cells or vehicle control was adoptively transferred via tail vein injection to B cell–deficient µMT Apoe−/− mice fed a Western diet for 8 weeks as previously described.3 After adoptive transfer, the mice continued Western diet for an additional 8 weeks, were sacrificed, and aortic sections were stained for VCAM-1.
VCAM mRNA Isolation and Polymerase Chain Reaction
Total RNA was isolated from aorta of five 8-week-old chow-fed Id3+/+ Ldlr−/− and 4 Id3−/− Ldlr−/− mice, and cDNA was synthesized and detected using primers for VCAM-1 and cyclophilin.
VCAM Western Method
Primary smooth muscle cells isolated from C57BL/6 and Id3−/− aortas were prepared, and Western blotting was carried out using an antibody to Id3, β-tubulin, and VCAM-1.
VCAM-1 Promoter-Reporter Assay
Primary smooth muscle cells harvested from C57BL/6 murine aortas were transfected with human VCAM-1-promoter-luciferase reporter together with human E12, human Id3, and/or empty vector. Cells were harvested 48 hours after transfection, and luciferase values were normalized to protein.
Chromatin Immunoprecipitation was performed as previously described9 using primary VSMCs from aortas of E47−/− mice with anti-E2A antibody, anti-RNA polymerase II, or immunoglobulin G control sera. The anti-E2A antibody binds to both E47 and E12, the gene products of E2A. Immunoprecipitated DNA fragments were analyzed by real-time polymerase chain reaction (RT-PCR) to assess for recovery of the mouse VCAM promoter.
Data are presented as the mean±standard error of the mean. A P value <0.05 was considered statistically significant. All statistical analyses were performed using Number Crunching Statistical Software 2007 (Kaysville, Utah) and GraphPad Prism5 (La Jolla, CA).
Loss of Id3 Increased Atherosclerosis Development in Ldlr−/− Mice
We previously demonstrated that Id3 is atheroprotective in an Apoe−/− murine model of atherosclerosis.3,4 To determine whether Id3 plays a role in other animal models of atherosclerosis, Id3+/+ Ldlr−/− and Id3−/− Ldlr−/− mice were fed either a Western diet or a chow diet for 8 weeks then harvested for en face analysis of lesion area (Figure 1A). Although there was no significant difference in the development of atherosclerosis after 8 weeks of chow diet between the 2 genotypes, Id3−/− Ldlr−/− mice had significantly more atherosclerosis after 8 weeks of Western diet feeding than Id3+/+ Ldlr−/− mice (4.9±0.4% versus 2.7±0.1%, P=0.002). To determine whether heterozygosity for Id3 had any impact on the development of atherosclerosis, Id3+/+ Ldlr−/−, Id3+/− Ldlr−/−, and Id3−/− Ldlr−/− mice were placed on Western diet for 16 weeks, and aortas were stained for en face analysis. As seen in Figure 1B, there were significant differences between all 3 genotypes (6.86±0.1% versus 8.81±0.1% versus 13.92±0.1%, respectively, ANOVA P=0.001), suggesting that even heterozygosity for Id3 can predispose to increased development of atherosclerosis. As demonstrated in the Table, there are no significant differences in serum lipid levels between the different groups that might account for the difference in atherosclerosis.
Optical Imaging Demonstrates Fewer B Cells in Aorta of Id3−/− Ldlr−/− Mice
We previously demonstrated that Apoe−/− mice null for Id3 have significantly decreased numbers of aortic B cells.3 We sought to validate this in Id3−/− Ldlr−/− mice with optical imaging using a Cy5.5-labeled anti-CD19 antibody. As a control, we used μMT mice, which lack peripheral B cells because of deletion of genomic DNA sequences that encode the transmembrane domain of the B cell receptor µ heavy chain.12 Aortas from B cell–deficient μMT mice were found to have a minimal amount of background signal, which likely represents the nonspecific retention of the Cy5.5-labeled anti-CD19 antibody. Although a significant increase in fluorescence over the control μMT aortas was observed in both Id3+/+ Ldlr−/− (P<0.002) and Id3−/− Ldlr−/− aortas (P<0.03; Figure 2A and 2B), the increase in fluorescence over controls in the aortas of Id3−/− Ldlr−/− mice was significantly lower than that in the aortas of Id3+/+ Ldlr−/− mice (15±4% versus 28±5%; P=0.047; Figure 2A and 2B), suggesting that Id3−/− Ldlr−/− mice have fewer aortic B cells.
Greater Macrophage Burden Is Present in the Atherosclerotic Plaque of Id3−/− Ldlr−/− Mice
To determine whether loss of Id3 alters plaque macrophage and T cell content, immunohistochemical staining of plaques from Id3+/+ Ldlr−/− and Id3−/− Ldlr−/− aortas was performed after 16 weeks of Western diet. As seen in the representative aortic cross-sections stained for macrophages (red) and T cells (green) in Figure 2C and corresponding quantitation (Figure 2D), macrophage content was greater in the aortic plaque of Id3−/− Ldlr−/− mice than in Id3+/+ Ldlr−/− mice. Although there was a trend toward an increased number of T cells in both the adventitia and intima in Id3−/− Ldlr−/− mice after 16 weeks of Western diet, this difference was not statistically significant (Figure 2C and 2E). The number of intimal T cells was not significantly different between Id3−/− Ldlr−/− mice and Id3+/+ Ldlr−/− mice after adjusting for the number of 4′,6-diamidino-2-phenylindole stained intimal cells (4.1±0.7% versus 3.0±0.4%, P=0.32).
CCL20 and VCAM-1 Expression Are Increased in the Aortic Atherosclerotic Plaques of Id3−/− Ldlr−/− Mice
Immunohistochemical staining for CCL20 in aortic cross-sections from Id3−/− Ldlr−/− and Id3+/+ Ldlr−/− mice after 16 weeks of Western diet revealed significantly greater intimal CCL20 staining in the atherosclerotic plaque of Id3−/− Ldlr−/− mice compared with Id3+/+ Ldlr−/− mice (Figure 3A and 3B). In contast, there was no significant difference in immunostaining for the ligand for CXCR4, C-X-C chemokine ligand-12 (Figure 3C and 3D). VCAM-1 is a member of the immunoglobulin-like superfamily of adhesion molecules and is known to play a critical role in leukocyte recruitment in atherosclerosis.13 Immunohistochemical staining of VCAM-1 in aortic cross-sections after 16 weeks of Western diet revealed significantly greater expression of intimal VCAM-1 in the atherosclerotic plaque of Id3−/− Ldlr−/− mice compared with Id3+/+ Ldlr−/− mice (Figure 3E and 3F).
Id3-Mediated Differences in Macrophage Content and VCAM-1 Expression Are Not a Result of Increased Cellularity or Vascular Smooth Muscle Content
We performed Movat staining to assess whether the increase in macrophage content or VCAM-1 staining may be a result of increased cellularity and found no difference in intimal cellularity between Id3−/− Ldlr−/− and Id3+/+ Ldlr−/− aortic plaque (Figure 3G). Terminal deoxynucleotidyl transferase dUTP nick end labeling staining was performed to assess whether differences in intimal macrophage content may be secondary to differences in apoptosis and found no significant difference between intimal terminal deoxynucleotidyl transferase dUTP nick end labeling staining in Id3−/− Ldlr−/− and Id3+/+ Ldlr−/− aortic plaque (Figure 3H). Finally, we found no significant difference in intimal smooth muscle α-actin staining between Id3−/− Ldlr−/− and Id3+/+ Ldlr−/− aortic plaque (Figure 3I), suggesting the increased intimal VCAM-1 staining is not a result of increased vascular smooth muscle content.
Increased Intimal VCAM-1 Content Is Not Mediated by B Cells
To assess whether the increase in intimal VCAM-1 staining was a result of the decreased aortic B cell content, we performed adoptive transfer of 45 x 106 Apoe−/− B cells (n=8) or vehicle control (n=9) into B cell–deficient µMT Apoe−/− mice after 8 weeks of Western diet and then fed the mice an additional 8 weeks of Western diet. We found no significant difference in intimal VCAM-1 staining and a trend toward increased VCAM-1 staining in the B cell recipients compared with vehicle control recipients (0.135±0.023 versus 0.091±0.02, P=0.17), suggesting that the decrease of B cells in the aortic wall in the Id3−/− mice should not account for the increase in VCAM-1.
Loss of Id3 Leads to Increased Transcription and Expression of VCAM-1 in VSMCs
To determine whether loss of Id3 resulted in increased VCAM-1 mRNA expression in the aorta, we performed RT-PCR on RNA obtained from the aortas of Id3+/+ Apoe−/− or Id3−/− Apoe−/− mice. RT-PCR confirmed the earlier VCAM-1 staining, identifying a significant 59% increase in VCAM-1 mRNA levels in the aortas of Id3−/− Ldlr−/− mice compared with Id3+/+ Ldlr−/− mice (Figure 4A). Western blot analysis for VCAM-1 in VSMCs harvested from C57BL/6 and Id3−/− mice revealed significantly more VCAM-1 protein in VSMCs from Id3−/− mice compared with wild-type controls (Figure 4B and 4C). Id3 regulates target gene expression through dimerization with E-proteins, such as E12, antagonizing E12 DNA binding and transcription regulatory effects. Because VCAM-1 contains multiple E-box sites (Figure 4D), we performed promoter-reporter studies to assess whether E12 activates the VCAM-1 promoter and if Id3 can inhibit this activation. Cotransfection of VSMCs from C57BL/6 mice with a human VCAM-1 promoter-luciferase reporter plasmid (VCAM-1Luc) and an E12 expression vector resulted in activation of the VCAM-1 promoter, an effect which was antagonized by cotransfection with the Id3 expression plasmid (Figure 4E). Consistent with VCAM-1 promoter repression with Id3 overexpression, analysis of primary VSMCs from C57BL/6 and Id3−/− mice transfected with VCAM-1Luc revealed significantly more VCAM-1 promoter activation in VSMCs null for Id3 (Figure 4F). Finally, we performed chromatin immunoprecipitation in primary VSMC isolated from E47−/− murine aortas that were infected with adenovirus expressing Id3 or empty vector and assessed for occupancy of E12 at the VCAM promoter using an anti-E2A antibody, which binds both E47 and E12. Our results demonstrate that E12 associates with the VCAM-1 promoter in E47−/− smooth muscle cells and this association is inhibited by Id3.
Atherosclerosis, a complex polygenic disorder, is the underlying pathology that leads to cardiovascular morbidity and mortality.14 Treatment of lipids has been a mainstay of prevention, yet despite the success of statin drugs, cardiovascular disease remains the leading cause of death in Western countries. Inflammation has emerged as a potentially new therapeutic frontier,15 underscoring the need for a better understanding of factors and pathways that limit vessel wall inflammation. Previous studies have implicated Id3 (a dominant negative regulator of gene expression) as an atheroprotective factor in Apoe−/− mice.3,4 Id3 is broadly expressed and regulates gene pathways involved in growth, differentiation, and survival in many cells types, including immune cells.6,16,17 As such, it is intriguing to hypothesize that Id3 could be an important upstream regulator of genes involved in vessel wall inflammation in many cell types. That a functionally significant polymorphism in the human ID3 gene is associated with cIMT in humans4 further supports this hypothesis. The current study confirms the atheroprotective role of Id3 in a second murine model of atherogenesis and is the first to demonstrate that, in addition to regulation of B cell–mediated atheroprotection, Id3 also represses the expression of inflammatory genes in the vessel wall.
The chemokine receptor 6 (CCR6) has been identified as a downstream target of Id3 in B cells and found to be important for aortic homing of B cells and B cell–mediated atheroprotection.3 The ligand for CCR6 is CCL20, a chemokine induced by low-density lipoprotein cholesterol and found in greater abundance in diseased versus nondiseased human coronary arteries,18 suggesting that CCL20/CCR6-mediated homing of immune cells to sites prone to or with existing atherosclerosis may be an important mechanism regulating atherogenesis. Indeed, B cells home to these sites, and this homing is associated with a significant reduction in aortic macrophages.3 CCR6 is also expressed on monocytes, and the CCL20/CCR6 system has been implicated in macrophage accumulation in plaques and atherogenesis.19 Results of the present study demonstrate that, in addition to promoting CCR6 expression in B cells, Id3 also repressed the expression of CCL20 in the vessel wall. The increased CCL20 expression in vessels from Id3−/− Ldlr−/− mice could be a second mechanism whereby loss of Id3 promotes macrophage accumulation in lesions.
VCAM-1 has been shown to play a critical role in the recruitment of monocytes and other leukocytes to atherosclerotic plaque.13 α4β1 (VLA-4) integrin is expressed on monocytes and binds to VCAM-1 to help facilitate transmigration across the activated endothelium.20 The ability of monocytes to adhere to the early atherosclerotic endothelium can be significantly reduced through treatment with antibodies that block VCAM-1.21 Although complete knockout of VCAM-1 is embryonically lethal, the importance of VCAM-1 in the development of atherosclerosis was demonstrated using Vcam1D4D/D4D Ldlr−/− mice, disrupting the fourth Ig domain of VCAM-1.22 Vcam1D4D/D4D Ldlr−/− mice developed significantly less atherosclerosis than Ldlr−/− mice.22 Furthermore, less significant reductions in expression of VCAM-1 in Vcam1D4D/+ Apoe−/− mice still resulted in reduced monocyte adhesion and fatty streak formation.23 Results of the present study demonstrate increased VCAM-1 staining in aortic plaques and increased VCAM-1 mRNA in whole aorta in Id3−/− Ldlr−/− mice. In addition to monocyte recruitment through endothelial cell expression of VCAM-1, VSMCs in the developing plaque may also express VCAM-1, contributing to macrophage retention in the vessel wall. Consistent with our results demonstrating increased macrophages deep within the plaque in Id3−/− Ldlr−/− mice, we demonstrate that VSMC express VCAM-1, and the amount of VCAM-1 expressed is significantly greater in VSMC from Id3−/− mice. However, we demonstrated that the increased intimal VCAM-1 staining was a result of loss of Id3 and not of increased VSMC content based on staining for smooth muscle α-actin. Additionally, we demonstrated in a B cell-deficient Apoe−/− model of atherosclerosis that adoptive transfer of B cells did not significantly alter intimal VCAM-1 expression. In fact, there was a trend toward increased VCAM-1 expression after adoptive transfer of B cells, suggesting that the increased intimal VCAM-1 expression in Id3−/− Ldlr−/− murine atherosclerosis is not because of reduced numbers of aortic B cells.
Id3 functions through dominant negative inhibition of E-proteins, such as E12, and indeed, E12 activates the VCAM-1 promoter, an effect antagonized by cotransfection with Id3. Furthermore, luciferase expression driven by the VCAM-1 promoter was significantly greater in VSMC from mice null for Id3. Finally, chromatin immunoprecipitation experiments demonstrated that E12 binds to the VCAM-1 promoter and this interaction is inhibited by Id3. Thus, VCAM-1 expression is upregulated by E12 and inhibited by Id3.
In summary, results confirm an atheroprotective role for Id3 in a second murine model of diet-induced atherosclerosis, the Ldlr−/− mouse. In addition, consistent with data in humans demonstrating that those heterozygous for the single nucleotide polymorphism in Id3 at rs11574 had an incremental increase in cIMT relative to homozygotes, Id3+/− Ldlr−/− also had an intermediate phenotype. Optical imaging confirmed Id3 as important for constitutive homing of B cells to the aorta, raising the possibility that the increase in atherosclerosis in the Id3−/− Ldlr−/− compared with Id3+/+ Ldlr−/− controls was also at least in part a result of reduced aortic B cells. Notably, results identify new Id3 targets in the vessel wall whose repression, especially VCAM-1, may lead to reduced macrophage accumulation and atheroprotection. Taken together, results suggest Id3 as an upstream regulator of important atheroprotective pathways in both B cells and vessel wall cells.
The authors gratefully acknowledge Melissa Bevard of the Cardiovascular Research Center at the University of Virginia Health System for her assistance with immunohistochemical staining.
Sources of Funding
This work was supported by National Institutes of Health (NIH) grants P01 HL55798 (to C.A.M.), NIH RO1 HL62522 (to C.A.M.), NIH R01 HL096447 (to C.A.M.), NIH Training Grant 5-T32 HL007355-29 (to M.J.L.), and American Heart Association Mid-Atlantic Affiliate postdoctoral fellowship award 10POST3560000 (to M.J.L.).
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.112.300352/-/DC1.
- Received March 30, 2012.
- Accepted September 7, 2012.
- © 2012 American Heart Association, Inc.
- Roger VL,
- Go AS,
- Lloyd-Jones DM,
- et al
- Doran AC,
- Lipinski MJ,
- Oldham SN,
- et al
- Doran AC,
- Lehtinen AB,
- Meller N,
- Lipinski MJ,
- Slayton RP,
- Oldham SN,
- Skaflen MD,
- Yeboah J,
- Rich SS,
- Bowden DW,
- McNamara CA
- Langlands K,
- Yin X,
- Anand G,
- Prochownik EV
- Chen B,
- Han BH,
- Sun XH,
- Lim RW
- Allain CC,
- Poon LS,
- Chan CS,
- Richmond W,
- Fu PC
- Galkina E,
- Ley K
- Klingenberg R,
- Hansson GK
- Calvayrac O,
- Rodríguez-Calvo R,
- Alonso J,
- Orbe J,
- Martín-Ventura JL,
- Guadall A,
- Gentile M,
- Juan-Babot O,
- Egido J,
- Beloqui O,
- Paramo JA,
- Rodríguez C,
- Martínez-González J
- Wan W,
- Lim JK,
- Lionakis MS,
- Rivollier A,
- McDermott DH,
- Kelsall BL,
- Farber JM,
- Murphy PM
- Huo Y,
- Hafezi-Moghadam A,
- Ley K
- Dansky HM,
- Barlow CB,
- Lominska C,
- Sikes JL,
- Kao C,
- Weinsaft J,
- Cybulsky MI,
- Smith JD