Impaired Regulatory T-Cell Response and EnhancedAtherosclerosis in the Absence of Inducible
Costimulatory MoleculeIsrael Gotsman, MD; Nir Grabie, PhD; Rajat Gupta, BS; Rosa Dacosta, AB;
Malcolm MacConmara, MB, BCh; James Lederer, PhD; Galina Sukhova, PhD;Joseph L. Witztum, MD; Arlene H. Sharpe, MD, PhD; Andrew H. Lichtman, MD, PhD
Background—T-cell–mediated immunity contributes to the pathogenesis of atherosclerosis, but little is known about howthese responses are regulated. We explored the influence of the inducible costimulatory molecule (ICOS) onatherosclerosis and associated immune responses.
Methods and Results—Bone morrow chimeras were generated by transplanting ICOS-deficient or wild-type bone marrowinto irradiated atherosclerosis-prone, LDR receptor–deficient mice, and the chimeric mice were fed a high-cholesteroldiet for 8 weeks. Compared with controls, mice transplanted with ICOS-deficient marrow had a 43% increase in theatherosclerotic burden, and importantly, their lesions had a 3-fold increase in CD4� T cells, as well as increasedmacrophage, smooth muscle cell, and collagen content. CD4� T cells from ICOS-deficient chimeras proliferated moreand secreted more interferon-� and tumor necrosis factor-� than T cells from control mice, which suggests a lack ofregulation. FoxP3� regulatory T cells (Treg) were found to constitutively express high ICOS levels, which suggests arole for ICOS in Treg function. ICOS-deficient mice had decreased numbers of FoxP3� Treg and impaired in vitro Tregsuppressive function compared with control mice.
Conclusions—ICOS has a key role in regulation of atherosclerosis, through its effect on regulatory T-cell responses.(Circulation. 2006;114:2047-2055.)
Atherosclerosis is characterized by the accumulation oflipids and chronic inflammatory cells in arterial vessel
walls.1 Although there is ample evidence that immune re-sponses contribute to the pathogenesis of this commondisease, there are many unanswered questions about howthose immune responses are regulated and how they may betargeted for therapeutic intervention.2 CD4� T-helper (Th)cells are evident in the atherosclerotic plaque and have beenshown to be important in the development of atherosclero-sis.3,4 Immune responses mediated by CD4� Th1 cells spe-cific for plaque antigens such as oxidized LDL (oxLDL) orheat shock protein 60/65 may have an important role in thepropagation of the inflammatory process.5 CD4� T-cell re-sponses to antigen are modulated by costimulatory signals,delivered by B7 family molecules on antigen-presenting cellsthat bind to CD28 family molecules on the T cells. We haveshown that the absence of B7-1 and B7-2 significantlyreduces the progression of atherosclerosis.6 Inducible co-stimulatory molecule (ICOS) and its ligand (ICOS ligand) are
CD28 and B7 family members, respectively, that also have animportant role in modulating T-cell activation. ICOS isexpressed on activated T cells after antigen stimulation7 andmay be important in secondary activation of differentiated Tcells. ICOS appears to support T-cell proliferation and Th2differentiation.8–10 In addition, ICOS has been implicated inregulatory T-cell (Treg) function.11,12 ICOS-deficient miceare more susceptible to experimental autoimmune encephali-tis.7 ICOS has also been shown to be important in antibodyisotype switching.7,13,14
Clinical Perspective p 2055Recent evidence suggests that the local balance between
effector and natural regulatory T cells influences the devel-opment of atherosclerosis in animal models.15,16 Furthermore,Th1 proinflammatory cytokines such as interferon (IFN)-�,interleukin (IL)-18, and IL-12 may have a key role inpromoting atherogenesis,17,18 whereas T-cell responsivenessto transforming growth factor-� (TGF-�) has been shown to
Received April 10, 2006; revision received August 2, 2006; accepted September 8, 2006.From the Immunology and Vascular Research Divisions (I.G., N.G., R.G., R.D., A.H.S., A.H.L.), Department of Pathology, Donald W. Reynolds
Cardiovascular Clinical Research Center (G.S.), Department of Medicine, and Department of Surgery (M.M., J.L.), Brigham and Women’s Hospital, andHarvard Medical School, Boston, Mass; and Department of Medicine (J.L.W.), University of California, San Diego, La Jolla, Calif.
Guest Editor for this article was Roberto Bolli, MD.Correspondence to Andrew H. Lichtman, MD, PhD, Department of Pathology, Brigham and Women’s Hospital, 77 Avenue Louis Pasteur, NRB 752N,
limit atherosclerosis.19 In addition, natural antibodies specificfor oxidized phospholipids may have a protective effect onthe development of atherosclerotic lesions.20,21 Because ICOSis reported to influence effector T-cell, antibody, and regula-tory T-cell responses, we hypothesized that the ICOS path-way regulates immune responses to plaque antigens andthereby influences development of arterial disease. To testthis hypothesis, we used a bone marrow chimeric approach toexamine the influence of ICOS deficiency on atherosclerosisin LDL receptor (LDLR)–deficient mice. Our findings showthat ICOS deficiency in bone marrow–derived cells results inenhanced atherosclerosis and associated immune responsesdue to the requirement of ICOS for development and functionof atheroprotective natural Treg.
MethodsMiceLDLR�/� mice, backcrossed 10 times onto a C57BL/6 background,and wild-type C57BL/6 mice were purchased from Jackson Labora-tories (Bar Harbor, Me). ICOS�/� mice were derived as describedelsewhere14 and were backcrossed 10 generations onto a C57BL/6background. All mice were housed and bred in accordance with theinstitutional guidelines of Brigham and Women’s Hospital andHarvard Medical School.
Bone Marrow Transplantation ProtocolAt 8 weeks of age, LDLR�/� mice were lethally irradiated with acesium source (total dose 1300 rad, split into 2 doses 4 hours apart).Donor bone marrow cells were injected after the second irradiation.Eight-week-old ICOS�/� or wild-type donor mice were euthanizedby CO2 asphyxiation, and femurs, tibias, and humeri were harvestedby a clean technique. Bone marrow cells were extracted, and 20million cells were injected via tail vein into sex-matched irradiatedLDLR�/� recipient mice. Recipient mice received normal chow andwater containing sulfamethoxazole (4.14 mg/mL) and trimethoprim(0.83 mg/mL) for 4 weeks.
Induction of AtherosclerosisFour weeks after the transplantation, the mice were begun on asemipurified cholate-free diet (No. D12108; Research Diets Inc,New Brunswick, NJ) containing 40% kcal lipid, 1.25% cholesterol22
ad libitum (6 to 8 males and 6 to 8 females per group). After 8 weekson this diet, the mice were fasted overnight and killed by CO2
asphyxiation. Blood was collected by vena cava nicking, and thearterial tree was perfused with Dulbecco’s phosphate-buffered saline(Gibco BRL, Gaithersburg, Md). The heart and the whole perfusedaorta were dissected from the aortic valve to the iliac bifurcation.The heart base, including the aortic sinus with the proximal aorticarch and the aortic branches, was cut and separated from theremaining aorta, then rapidly frozen in optimal cutting temperatureembedding medium (OCT, ProSciTech, Thuringowa, Australia); theremaining thoracic and abdominal aorta (descending aorta) fromeach mouse was fixed in 10% buffered formalin.
Aortic Atherosclerotic Lesion AnalysisThe aortic sinus was analyzed as described previously.6 Alternate5-�m cryosections of �200 �m of the aortic sinus were prepared.From the area in which 3 aortic valve cusps are clearly seen, alternatesections were collected for quantification and stained with oil red Oaccording to the method of Paigen et al.23 Section images werecaptured digitally (SV Micro, Zeiss, Jena, Germany) and quantifiedwith Image-Oro Plus software (Media Cybernetics, Silver Spring,Md). Plaque lesion area and percent of the total cross-sectionalvessel wall area were quantified by an independent operator, blindedto the experimental protocol, and the results were expressed as theaverage of 6 sections per mouse.
ImmunohistochemistrySerial cryostat sections of aortic sinus adjacent to the oil redO–stained sections were stained with the respective molecule-specific antibodies, as described previously.24 Antibodies includedrat anti-mouse Mac-3 for macrophage identification (1:1000, BDPharmingen, San Diego, Calif) and anti-CD4� for T cells (1:120, BDPharmingen). For mouse smooth muscle cell actin staining, primaryantibody (fluorescein isothiocyanate [FITC]-conjugated �-actin,1:500, Sigma, St Louis, Mo) was applied, followed by anti-FITCbiotin-conjugated secondary antibody (1:400, Sigma). Proliferatingcells were stained by rabbit antinuclear antigen Ki67 antibody(Novocastra Laboratories, Newcastle upon Tyne, United Kingdom).For FoxP3 staining, we used a goat anti-mouse/rat FoxP3 antibodyFJK-16s (eBioscience, San Diego, Calif).
Quantitative analysis of lesional content of macrophages, smoothmuscle cells, and collagen was determined by computer-assistedimage analysis6 and expressed as percentage of intimal area tonormalize for overall differences between the study groups. Quan-tification of CD4, Ki67, and FoxP3 staining was done by countingindividual positively stained lesional cells, which are easily resolved,in the aortic sinus sections.
Polymerase Chain Reaction Analysis for EngraftmentTo assess engraftment, the presence of the LDLR DNA gene product(383 bp) was analyzed by polymerase chain reaction (PCR) inperipheral blood leukocytes (PBLs) sampled from the recipient mice4 weeks after bone marrow transplantation (before initiation ofhigh-cholesterol diet) and in CD4� splenocytes obtained when micewere euthanized as previously described6 with a GFX genomic bloodDNA purification kit (GE Healthcare Bio-Sciences, Piscataway, NJ).
Serum Cholesterol AnalysisOvernight fasting serum samples were collected after 8 weeks ofcholesterol-enriched diet. Total serum cholesterol levels, triglycer-ides, and plasma lipoproteins were analyzed by an online dual-enzymatic method for simultaneous quantification of cholesterol andtriglycerides by high-performance liquid chromatography at SkylightBiotech Inc (Akita, Japan) according to the procedure described byUsui et al25 and were expressed in milligrams per deciliter (mg/dL).
Serum Immunoglobulin AnalysisIgM, IgG1, IgG2c, and IgG3 antibodies specific for malondialde-hyde LDL and copper-oxidized (Cu-Ox) LDL were detected byisotype-specific ELISA as described previously.17 IgG2c was de-tected with an IgG2a/c-specific reagent that may underestimateIgG2c levels. We also determined titers in serum of IgM autoanti-bodies to oxLDL with the T15/EO6 idiotype as describedpreviously.26
Ex Vivo Assays of CD4� Proliferation andCytokine SecretionAfter 8 weeks of diet, splenic CD4� T cells were isolated byanti-CD4 microbeads (Miltenyi Biotec, Auburn, Calif). The cellswere stimulated in 96-well cultures (2.5�105/well) with plate-boundanti-CD3� (145-2C11, BD Pharmingen) or human copper-oxidizedLDL (Biomedical Technologies Inc, Stoughton, Mass) plus irradi-ated spleen cells (2.5�106/well). Ovalbumin (10 �g/mL) and me-dium alone were used as controls. Culture supernatants were re-moved at 48 hours and analyzed by flow cytometry–based cytokinebead assays (BD Pharmingen) of culture supernatants for IFN-�,IL-2, tumor necrosis factor (TNF)-�, IL-4, IL-5, and IL-10.27
Cultures were assayed for proliferation after 64 hours by uptake of[3H]thymidine (1 �Ci/well), added 16 hours before harvest. Data areexpressed as mean proliferation indices of triplicates calculated fromthe ratios of incorporated radioactive counts per minute in thepresence or absence of antigen.
2048 Circulation November 7, 2006
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
Real-Time PCRCD4� spleen cells stimulated with anti-CD3 were analyzed forcytokine production by TaqMan real-time PCR (RT-PCR). TotalRNA was isolated from 5�106 million purified T cells, reversetranscribed, and analyzed by quantitative RT-PCR with SYBR greenPCR mix (Bio-Rad Laboratories, Hercules, Calif).24,28 All real-timereactions were performed on the iCycler iQ real-time PCR detectionsystem (Bio-Rad), and analysis was done with the accompanyingsoftware. The presence of single amplicons resulting from RT-PCRwas verified by dissociation curve analysis. Levels of specific geneexpression in the samples are presented relative to endogenous levelsof �-actin housekeeping gene expression in the same sample.
Flow CytometryLymphocytes were stained with BD Biosciences phycoerythrin (PE)-or allophycocyanin-conjugated anti-CD25 (PC61), PE- or FITC- orPE-Cy5–conjugated anti-CD4 (L3T4), FITC-conjugated anti-CD3,and PE-conjugated anti-ICOS (7E.17G9). For intracellular stainingof FoxP3, a FoxP3-allophycocyanin antibody (FJK-16s) and Foxp3staining buffer set from eBioscience were used according to themanufacturer’s recommended protocol. All samples were analyzedusing a FACSCalibur flow cytometer with CellQuest software (BDBiosciences) as described previously.24
Regulatory T-Cell Isolation and Functional AssaysTregs were isolated from spleen cell suspensions by the CD4�CD25�
regulatory T cell isolation kit (Miltenyi Biotec), which depletessamples of non-CD4� T cells followed by positive selection ofCD4�CD25� cells. The purity of the isolated CD4�CD25� andCD4�CD25� populations was �95% by fluorescence-activated cell
sorter analysis. CD4�CD25� responder cells (5�104/well) werecultured in 96-well plates (0.2 mL) with soluble anti-CD3� (1�g/mL) plus irradiated spleen cells (5�104/well) and increasingnumbers of CD4�CD25� regulatory T cells (responder to suppressorratio of 1:16 to 1:1) for 72 hours. Cultures were pulsed with[3H]thymidine for the last 16 hours of culture. Culture supernatantswere removed at 48 hours and analyzed for cytokines, as describedabove. Anti-TGF-�–blocking antibody (R&D Systems Inc, Minne-apolis, Minn) was added to some suppression assays at 25 �g/mL.
Statistical AnalysisAll statistical analyses were performed with Prism software (Graph-Pad Software, Inc, San Diego, Calif). Differences between groups ofmice were analyzed by the Student t test or the Mann-Whitney test(for data that did not pass the Kolmogorov-Smirnov normality test),and data are expressed as mean�SEM.
The authors had full access to and take full responsibility for theintegrity of the data. All authors have read and agree to themanuscript as written.
ResultsBone Marrow Engraftment, Weights, andSerum LipidsTo confirm engraftment, mice were bled at 4 weeks aftertransplantation, and PBLs were counted. There was nodifference in total cell number between the groups(7.7�1.2�106/mL versus 9.1�0.8�106/mL in wild-type andICOS-null cell recipients, respectively, P�0.35). To furtherassess engraftment, we looked for the presence of wild-type
Figure 1. Engraftment and chimerism of donor hematopoietic cells.A, PCR amplification of the LDLR DNA gene product (383 bp) inPBLs 4 weeks after transplantation and in CD4� splenocytes attime of euthanasia. Wild-type (WT) and ICOS �/� refer to the donorcells transplanted. The mutant LDLR band represented �5% ofthe cells as confirmed by quantitative mixing controls. B, ICOSexpression on anti-CD3–stimulated PBLs and CD4� splenocytes.Expression of ICOS on stimulated CD4� T cells from untrans-planted LDLR �/� and ICOS �/�mice are located at the left of eachfluorescence-activated cell sorter analysis for comparison.
Figure 2. Effect of ICOS deficiency on atherosclerotic burden. Aand B, Aortic sinuses from female LDLR�/� mice transplantedwith wild-type (WT; A) vs ICOS�/� (B) bone marrow, stained withoil red O (magnification �10). C, Quantitative analysis of lesionburden. D, Quantitative analysis of fractional lesion area. Eachdata point represents mean value for each mouse; horizontalbars represent mean value for each group.
Gotsman et al ICOS and Regulatory T Cells in Atherosclerosis 2049
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
and mutant LDLR genes by PCR in PBLs and splenocytes ofthe recipient mice. The wild-type product was equally abun-dant in recipients of ICOS-null or wild-type marrow (Figure1A). Importantly, there was also no ICOS protein expressionon PBLs sampled at 4 weeks or on CD4� splenocytesobtained at the time of euthanasia from recipients of ICOS�/�
bone marrow (Figure 1B).There was no significant difference in male or female
weights between the groups (Table 1). The analysis of totalserum cholesterol levels, as determined by enzymatic assay,revealed no statistically significant differences between the 2groups (Table 2).
Quantitative and Phenotypic Differences inAtherosclerotic Lesions in ICOS-Deficient andControl MiceAfter 8 weeks of high-cholesterol diet, mice transplanted withICOS�/� cells had significantly more atherosclerosis in the aorticsinus than control mice transplanted with wild-type cells (Figure2). In mice transplanted with ICOS�/� cells, average lesion areawas increased from 309 000�42 000 �m2 (n�10) to474 000�37 000 �m2 (n�15; P�0.05; Figure 2C). Fractionallesion area was increased from 37.4�2.2% to 53.6�2.2%(P�0.0001), a 43% increase in the atherosclerotic burden(Figure 2D). The increase in atherosclerosis in recipients ofICOS�/� cells was significant when atherosclerosis was evalu-ated according to sex. For wild-type versus ICOS�/� males, therespective fractional areas were 34.5�1.8% (n�6) versus50.2�3.1% (n�7; P�0.005). For wild-type versus ICOS�/�
Figure 3. Effect of ICOS deficiency onatherosclerotic phenotype. Representa-tive sections of aortic sinuses stainedwith antibodies specific for CD4� T cells(A, B), Mac3 for macrophages (C, D),and smooth muscle actin for smoothmuscle cells (SMC; E, F). Collagen typesI and III were stained by Picrosirius red,and the sections were analyzed by polar-ization microscopy (G, H). Proliferatingcells were stained with anti-Ki67 mono-clonal antibody. Proliferating cells stainbrown, and nonproliferating cells staingreen (I, J). Quantitative analyses ofstained sections are shown for CD4 (K),macrophages (L), SMC (M), collagen (N),and proliferation (O). Each data pointrepresents the mean value obtained foreach mouse; horizontal bars representthe mean value for each group.
TABLE 1. Mean Weights of Mice According to Sex
Weight, g
Group12 Weeks
(Diet Initiation)20 Weeks
(Euthanasia)
WT3LDLR�/�
Male 23.2�0.3* 26.3�0.9
Female 20.4�0.7 21.8�0.8
ICOS�/�3LDLR�/�
Male 24.4�0.3 27.4�0.6
Female 18.9�0.2 20.9�0.6
*P�0.05 when comparing mice of the same sex between the 2 groups.
TABLE 2. Serum Lipids and Lipoproteins at Timeof Euthanasia
Experimental Groups
WT3LDLR�/� ICOS�/�3LDLR�/� P
Total cholesterol, mg/dL 787.0�26.1 862.5�77.0 0.4
LDL cholesterol, mg/dL 383.4�9.2 405.2�38.4 0.6
HDL cholesterol, mg/dL 88.6�4.9 80.2�6.1 0.3
Total triglycerides, mg/dL 115.2�17.7 119.5�22.8 0.9
2050 Circulation November 7, 2006
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
females, the respective fractional areas were 41.7�4.3% (n�4)versus 56.6�2.8% (n�8; P�0.01). We found minimal andunquantifiable atherosclerotic lesions in the aortic arch anddescending aorta, consistent with previous reports of irradiatedLDLR�/� bone marrow recipients.29
After 8 weeks of proatherogenic diet, there was a 3-foldincrease in CD4� T cells in the intima of mice transplantedwith ICOS�/� cells (Figures 3A, 3B, and 3K). In addition,there was increased smooth muscle cell content (Figures 3E,3F, and 3M) and collagen deposition (Figures 3G, 3H, and3N) in these mice. These differences between wild-type andICOS�/� groups were significant when both sexes wereevaluated together and when each sex was evaluated sepa-rately. There also were more macrophages in the ICOS�/�
group than the control group (Figures 3C, 3D, and 3L), butthe difference was not statistically significant. The presentimmunohistochemical data suggest that there is some hetero-geneity within the ICOS�/� group, most evident in the smoothmuscle cell content, with a distinct subgroup that does notshow an increase compared with the control group. We foundan increased presence of proliferating cells, stained by a
Ki67-specific antibody, in mice transplanted with ICOS�/�
cells compared with control mice (Figures 3I, 3J, and 3O).Taken together, these immunohistochemical data suggest thatmice transplanted with ICOS�/� cells have more activeatherosclerotic lesions with more T-cell infiltration; morelocal proliferation of T cells, macrophages, and/or smoothmuscle cells; and more synthesis of extracellular matrix.
Enhanced Immunologic Responses inICOS-Deficient MiceTo examine whether ICOS deficiency altered T-cell function, weisolated CD4� T cells from mice that received ICOS�/� orwild-type bone marrow and compared their responses to poly-clonal and antigen-specific stimuli. CD4� T cells isolated frommice transplanted with ICOS�/� cells had a significantly greaterproliferative response to anti-CD3 or oxLDL than control mice(Figure 4A). Although the magnitude of the oxLDL-specificresponse was low, the differences were consistent and statisti-cally significant when T cells were stimulated with 10 �g/mLoxLDL (Figure 4B).
CD4� T cells isolated from mice transplanted with ICOS�/�
bone marrow secreted more Th1 cytokines (IL-2, IFN-�, and
Figure 4. The effect of ICOS deficiency on immune responses. A and B, Proliferation indices of splenic CD4� T cells stimulated withanti-CD3 (A) or oxLDL (B); n�12 per group. C–E, Cytokine levels in supernatants of anti-CD3–stimulated CD4� spleen T cells describedin A and B; n�8 to 12 per group. F–H, Cytokine mRNA levels determined by quantitative RT-PCR of mRNA from the stimulated CD4�
T cells described in A–E; n�5 to 6 per group. I, Serum levels of proinflammatory cytokines at time of sacrifice; n�5 to 6 per group.Data shown are mean�SEM. conc indicates concentration; MCP-1, monocyte chemotactic protein-1.
Gotsman et al ICOS and Regulatory T Cells in Atherosclerosis 2051
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
TNF-�) and more Th2 cytokines (IL-10, IL-4, and IL-5) onstimulation with anti-CD3 (Figures 4C and 4D) than CD4� Tcells isolated from controls. In contrast, the CD4� T cellsfrom ICOS�/� bone marrow recipients produced significantlyless antiinflammatory cytokine TGF-� than CD4� T fromcontrol mice (Figure 4E). Quantitative RT-PCR analysis ofthe cytokine mRNA expression of the stimulated CD4� Tcells was concordant with the cytokine protein assays, withthe exception of IL-10, which did not differ significantlybetween the experimental groups (Figures 4F through 4H).
We found a significant increase in serum monocyte chemo-tactic protein-1 and a trend for increased TNF-� in micetransplanted with ICOS�/� cells compared with controls (Figure4I). There were no significant differences in IFN-�, IL-10,IL-12, and IL-6. These results are consistent with an enhancedsystemic inflammatory response in the mice with ICOS-deficient marrow.
Antibody responses to oxLDL may modulate atheroscle-rotic disease, and ICOS may influence production of theseantibodies.20,21 We therefore measured titers of serum anti-bodies specific for different forms of oxLDL (Figure 5).There were statistically significant reductions in anti-malondialdehyde-LDL IgG1 and anti-oxLDL IgG3 in theICOS-deficient group, which may reflect impaired isotypeswitching to these isotypes in these mice due to ICOSdeficiency.30 However, the data do not support a global defectin isotype switching in the mice transplanted with ICOS-deficient bone marrow.30 There were no differences in thephosphorylcholine-specific IgM natural antibody identifiedby the EO6 antibodies idiotype, which has been shown to beprotective in atherosclerosis.26 Overall, the data do notsupport a role for changes in antibody titers as an explanationfor the difference in atherosclerosis phenotype betweengroups.
Relationship Between ICOS and Regulatory T-CellNumbers and FunctionIn light of published data that suggest that ICOS may beimportant in Treg function and the present data indicatingenhanced immune responses in the absence of ICOS, weevaluated the relationship between ICOS and Treg. Using 2specific molecular markers of these cells, CD25 and intracel-lular FoxP3, we found that the frequency of ICOS-expressingcells was 3-fold higher in the CD4� Treg subset than in theCD4� non-Treg subset (Figure 6A). These data suggest thatICOS has a role in the development or function of this T-cellsubset. We therefore examined Treg counts and function inICOS�/� mice (unirradiated, LDLR�/�). We found a 30%reduction in Treg counts as evidenced by FoxP3 staining inthe spleen of ICOS�/� mice on regular or high-cholesterol dietcompared with control mice (Figure 6B). There was nosignificant difference in total spleen CD4� T-cell countsbetween the groups. We also quantified Treg in LDLR�/�
mice transplanted with ICOS�/� versus wild-type bone mar-row on high-cholesterol diet for 8 weeks. For this analysis,we included iliac lymph nodes, which drain the atheroscle-rotic descending aorta. Treg counts were reduced by 40% inthe iliac lymph nodes and the spleen in mice transplantedwith ICOS�/� cells compared with control mice (Figure 6C).These data support the hypothesis that ICOS positivelyinfluences the development or maintenance of Treg.
The presence of ICOS on Treg, together with the increasednumber of T cells and enhanced lesion formation in mice withICOS-deficient bone marrow, suggested that Treg may controlthe proatherogenic effector T cells and influence the inflamma-tory burden in the atherosclerotic plaque. We therefore lookedfor the presence of Tregs in the mouse atherosclerotic lesions byimmunohistochemical staining for FoxP3. We found that micetransplanted with either wild-type or ICOS�/� bone marrow had
Figure 5. Serum titers of oxLDL-specificantibodies. IgG1, IgG2c, IgG3, and IgMantibodies that bind malondialdehyde(MDA)-LDL (A, B), copper-oxidized (Cu-Ox) LDL (C, D), and phosphorylcholine-specific (EO6) antibodies (E) weredetected by isotype-specific ELISA. Datashown are mean�SEM; n�12 per group.RLU indicates relative light units.
2052 Circulation November 7, 2006
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
readily detectable Treg in the atherosclerotic lesions (Figure6D); however, we could not demonstrate significant quantitativedifferences between the groups. The total CD4� T-cell counts inthe lesions were more than 3-fold higher in the mice transplantedwith ICOS�/� cells compared with control mice (Figure 3),which suggests that the Tregs that are present in the lesions ofthese mice are insufficient in number or function to suppress thelocal effector T-cell response.
To evaluate whether regulatory T-cell function, in additionto numbers, is influenced by ICOS expression, we comparedthe suppressive function of the Treg isolated from ICOS�/�
and wild-type mice. Indeed, the ability of Treg from ICOS�/�
mice to suppress proliferation of wild-type or ICOS�/� cellresponders was reduced compared with Treg from wild-typemice (Figure 6E). Reduced suppression by ICOS�/� Treg wasalso clearly evident by analysis of interferon-� and TNF-� in
Figure 6. Expression of ICOS in Tregsand the effect of ICOS on Treg counts andfunction. A, Expression of ICOS in splenicTregs stained with CD4-FITC, ICOS-PE,and intracellular FoxP3-allophycocyanin(APC) or CD25-APC. Fluorescence-acti-vated cell sorter analysis was performedby gating on CD4� cells and determiningthe percent of ICOS� or FoxP3� cells. Atypical result of 3 experiments is shown.The numbers refer to percent of ICOS�
cells in FoxP3� and FoxP� subsets orCD25� or CD25� subsets. The axes repre-sent relative fluorescence. B, Treg inspleens of wild-type (WT) and ICOS�/�
mice. Staining of CD4� cells from ICOS�/�
and WT C57B1/6 mice after 10 weeks ofhigh-cholesterol diet (n�12) orno-cholesterol diet (n�7) using APC-labeled antibodies to FoxP3 or CD25. C,Treg counts by FoxP3-APC staining in iliaclymph nodes (LN) and spleens of LDLR�/�
mice transplanted with ICOS�/� (n�7) orWT (n�7) bone marrow after 8 weeks ofhigh-cholesterol diet. D, Immunohisto-chemical staining for FoxP3� cells in ath-erosclerotic lesions. Typical sections of 6stained for each group are shown. E, Sup-pression of T-cell proliferation by Tregsfrom ICOS�/� and WT C57Bl/6 mice. Cul-tures were set up as described in Methodswith WT Treg suppression of WT respond-ers (�), ICOS�/� Treg suppression of WTresponders (}), and ICOS�/� Treg sup-pression of ICOS�/� responders (�). Thedifferences between ICOS�/� Tregs com-pared with WT Tregs were significant(P�0.05) at every ratio. F and G, Suppres-sion of cytokine secretion by Tregs fromICOS�/� and WT C57Bl/6 mice. Superna-tants were harvested from the culturesdescribed in E at 48 hours, and concentra-tions of IFN-� (E) and TNF-� (F) weredetermined; WT Treg suppression of WTresponders (�) and ICOS�/� Treg suppres-sion of WT responders (}). Data shownare from 1 of 2 experiments with similarresults. H, TGF-�– and ICOS-dependentsuppression of T-cell proliferation by Tregsfrom hypercholesterolemic LDLR�/� micetransplanted with ICOS�/� or WT bonemarrow. Treg and responder cells wereisolated from mice after 8 weeks on high-cholesterol diet. Anti-TGF-�–blocking anti-body was added as indicated. The differ-ences between ICOS�/� and WT bonemarrow chimeric mice were significant(P�0.05) at all ratios. The differencebetween WT chimeric mice with or withoutanti-TGF-� was significant at all ratiosexcept 0. There was no significant differ-ence between ICOS�/� chimeric mice withor without TGF-�–blocking antibody.incorp. indicates incorporation.
Gotsman et al ICOS and Regulatory T Cells in Atherosclerosis 2053
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
the supernatants of these assay cultures (Figures 6F and 6G).To address whether TGF-� has a role in Treg function inatherosclerotic mice, we used a TGF-�–blocking antibody inthe suppression assay. We found that blocking TGF-� causedreduced suppression by Treg from LDLR�/� mice trans-planted with wild-type bone marrow (Figure 6H). Anti-TGF-� did not cause significant effects on the alreadyreduced suppression seen with Treg from LDLR�/� micetransplanted with wild-type ICOS�/� bone marrow.
DiscussionIn this study, we found that ICOS deficiency on bonemarrow– derived cells increases the atherosclerotic burdenin LDLR�/� mice. Importantly, the increase in lesions isassociated with increased infiltration of CD4� T cells,which indicates that plaque-based T-cell responses aremore active in the absence of ICOS. In addition, thelesions in the ICOS-deficient mice had more smoothmuscle cell and collagen deposition. The increased athero-sclerotic burden in ICOS�/� bone marrow–recipient micecorrelated with increased ex vivo T-cell responsiveness topolyclonal and antigen-specific stimuli and in monocytechemotactic protein serum levels. Taken together, thesedata imply that ICOS deficiency leads to an enhancedT-cell response to hypercholesterolemia and plaque anti-gens and increased atherosclerosis.
Similar to the present data, ICOS deficiency has been shownto enhance helper T-cell responses in models of autoimmunediseases, including experimental autoimmune encephalitis7 andinsulitis.11 Experimental evidence clearly show that ICOS is apositive costimulatory molecule for CD4� T cells.7 Therefore, itis paradoxical that ICOS deficiency increases immune responsesin vivo. A possible explanation is the role of ICOS in thedevelopment or function of Th2 cells, which could downregulateTh1 responses.8 However, recent studies show that ICOS is alsoinvolved in Th1 differentiation.31,32 Consistent with this, we sawan increase in both Th1 and Th2 cytokine production by CD4�
T cells isolated from ICOS-deficient hypercholesterolemic miceex vivo. Therefore, we do not think a change in Th1:Th2 balanceexplains the effect of ICOS deficiency in the present study. Thepresent data also do not support the idea that the effect of ICOSdeficiency is due to changes in antibody responses to oxidativelymodified LDL or phosphorylcholine, which are known tomodulate atherosclerosis.20,26
The best-characterized subset of regulatory T cells (Tregs) arethe “natural” CD4�CD25� Tregs, which constitute �10% ofperipheral CD4� T cells.33 Most natural Tregs express thetranscription factor FoxP3.34 We considered that impaired nat-ural Treg activity may be the mechanism underlying the in-creased immune response and atherosclerotic burden in thesetting of ICOS deficiency. Indeed, we found that ICOS plays animportant role in induction or maintenance, as well as function,of natural Treg. There was a 40% reduction in the number ofTregs in the absence of ICOS as determined by FoxP3 expres-sion. This difference is especially significant given the ability ofsmall numbers of Tregs to suppress T-cell responses. Further-more, we found a reduction in Treg suppressive function inICOS�/� mice assessed in vitro. This is consistent with recentstudies showing that ICOS is crucial to Treg-dependent self-
tolerance.11,12,35,36 Therefore, ICOS deficiency in the setting ofhypercholesterolemia would be predicted to enhance proathero-genic T-cell responses and exacerbate lesion development, as wesaw in the present study. A recent article reported that apoli-poprotein E–deficient mice immunized with an ICOS-Ig fusionprotein had increased early lesion development,37 but that studydid not clarify whether and how the immunization altered theICOS pathway in vivo, nor did it provide a mechanism by whichthe ICOS pathway might influence atherosclerosis. The presentdata are in accordance with a study demonstrating the impor-tance of Treg in inhibiting development of atherosclerosis inapolipoprotein E mice.16
Interestingly, the only cytokine that we analyzed whoseproduction was decreased in mice receiving ICOS�/� cells versuscontrol mice was TGF-�. In addition, an in vitro suppressiveassay from these mice suggests that TGF-� plays a role in Tregfunction. This cytokine is a potent regulator of effector T-celldifferentiation, and it inhibits the acquisition of specific Th1 orTh2 cell functions.38 Conversely, TGF-� promotes the develop-ment of FoxP3-expressing Tregs.39,40 Therefore, reduced TGF-�may contribute to the increased immune responses seen in theICOS�/� mice in the present study. Overall, the present datasupport the notion that Tregs inhibit atherosclerosis throughTGF-� secretion.16,19
In summary, our results indicate that ICOS is important forthe regulation of proatherogenic T-cell responses in hypercho-lesterolemic animals. Enhancement of Treg responses, throughmanipulation of the ICOS pathway, has potential as a therapeuticstrategy for atherosclerotic disease.
Sources of FundingThis work was supported by National Institutes of Health grantsP50HL56985 (Dr Lichtman), R01AI3831O (Dr Sharpe), andR01HL67249 (Dr Sukhova).
DisclosuresNone.
References1. Hansson GK. Immune mechanisms in atherosclerosis. Arterioscler
Thromb Vasc Biol. 2001;21:1876–1890.2. Hansson GK, Libby P. The immune response in atherosclerosis: a
opment and response to immunization reveal a complex role for CD4 inatherosclerosis. Circ Res. 2005;96:427–434.
4. Zhou X, Robertson AK, Hjerpe C, Hansson GK. Adoptive transfer of CD4�T cells reactive to modified low-density lipoprotein aggravates atherosclero-sis. Arterioscler Thromb Vasc Biol. 2006;26:864–870.
5. Buono C, Lichtman AH. Co-stimulation and plaque-antigen-specific T-cellresponses in atherosclerosis. Trends Cardiovasc Med. 2004;14:166–172.
6. Buono C, Pang H, Uchida Y, Libby P, Sharpe A, Lichtman AH. B7-1/B7-2co-stimulation regulates plaque antigen-specific T cell responses and athero-genesis in LDLR-deficient mice. Circulation. 2004;109:2009–2015.
8. Vieira PL, Wassink L, Smith LM, Nam S, Kingsbury GA, Gutierrez-RamosJC, Coyle AJ, Kapsenberg ML, Wierenga EA. ICOS-mediated signalingregulates cytokine production by human T cells and provides a unique signalto selectively control the clonal expansion of Th2 helper cells. EurJ Immunol. 2004;34:1282–1290.
9. Nurieva RI, Duong J, Kishikawa H, Dianzani U, Rojo JM, Ho I, Flavell RA,Dong C. Transcriptional regulation of Th2 differentiation by induciblecostimulator. Immunity. 2003;18:801–811.
2054 Circulation November 7, 2006
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
10. Lohning M, Hutloff A, Kallinich T, Mages HW, Bonhagen K, Radbruch A,Hamelmann E, Kroczek RA. Expression of ICOS in vivo defines CD4�effector T cells with high inflammatory potential and a strong bias forsecretion of interleukin 10. J Exp Med. 2003;197:181–193.
11. Herman AE, Freeman GJ, Mathis D, Benoist C. CD4�CD25� T regulatorycells dependent on ICOS promote regulation of effector cells in the predi-abetic lesion. J Exp Med. 2004;199:1479–1489.
12. Kohyama M, Sugahara D, Sugiyama S, Yagita H, Okumura K, Hozumi N.Inducible costimulator-dependent IL-10 production by regulatory T cellsspecific for self-antigen. Proc Natl Acad Sci U S A. 2004;101:4192–4197.
13. Tafuri A, Shahinian A, Bladt F, Yoshinaga SK, Jordana M, Wakeham A,Boucher LM, Bouchard D, Chan VS, Duncan G, Odermatt B, Ho A, Itie A,Horan T, Whoriskey JS, Pawson T, Penninger JM, Ohashi PS, Mak TW.ICOS is essential for effective T-helper-cell responses. Nature. 2001;409:105–109.
14. McAdam AJ, Greenwald RJ, Levin MA, Chernova T, Malenkovich N, LingV, Freeman GJ, Sharpe AH. ICOS is critical for CD40-mediated antibodyclass switching. Nature. 2001;409:102–105.
15. Heller EA, Liu E, Tager AM, Yuan Q, Lin AY, Ahluwalia N, Jones K, KoehnSL, Lok VM, Aikawa E, Moore KJ, Luster AD, Gerszten RE. ChemokineCXCL10 promotes atherogenesis by modulating the local balance of effectorand regulatory T cells. Circulation. 2006;113:2301–2312.
16. Ait-Oufella H, Salomon BL, Potteaux S, Robertson AK, Gourdy P, Zoll J,Merval R, Esposito B, Cohen JL, Fisson S, Flavell RA, Hansson GK,Klatzmann D, Tedgui A, Mallat Z. Natural regulatory T cells control thedevelopment of atherosclerosis in mice. Nat Med. 2006;12:178–180.
17. Buono C, Come CE, Stavrakis G, Maguire GF, Connelly PW, Lichtman AH.Influence of interferon-gamma on the extent and phenotype of diet-inducedatherosclerosis in the LDLR-deficient mouse. Arterioscler Thromb Vasc Biol.2003;23:454–460.
18. Lee TS, Yen HC, Pan CC, Chau LY. The role of interleukin 12 in thedevelopment of atherosclerosis in ApoE-deficient mice. Arterioscler ThrombVasc Biol. 1999;19:734–742.
19. Robertson AK, Rudling M, Zhou X, Gorelik L, Flavell RA, Hansson GK.Disruption of TGF-beta signaling in T cells accelerates atherosclerosis. J ClinInvest. 2003;112:1342–1350.
20. Binder CJ, Hartvigsen K, Chang M-K, Miller M, Broide D, Palinski W,Curtiss LK, Corr M, Witztum JL. IL-5 links adaptive and natural immunityspecific for epitopes of oxidized LDL and protects from atherosclerosis.J Clin Invest. 2004;114:427–437.
21. Horkko S, Bird DA, Miller E, Itabe H, Leitinger N, Subbanagounder G,Berliner JA, Friedman P, Dennis EA, Curtiss LK, Palinski W, Witztum JL.Monoclonal autoantibodies specific for oxidized phospholipids or oxidizedphospholipid-protein adducts inhibit macrophage uptake of oxidized low-density lipoproteins. J Clin Invest. 1999;103:117–128.
22. Lichtman AH, Clinton SK, Iiyama K, Connelly PW, Libby P, Cybulsky MI.Hyperlipidemia and atherosclerotic lesion development in LDL receptor-deficient mice fed defined semipurified diets with and without cholate.Arterioscler Thromb Vasc Biol. 1999;19:1938–1944.
23. Paigen B, Morrow A, Holmes PA, Mitchell D, Williams RA. Quantitativeassessment of atherosclerotic lesions in mice. Atherosclerosis. 1987;68:231–240.
24. Grabie N, Delfs MW, Westrich JR, Love VA, Stavrakis G, Ahmad F,Seidman CE, Seidman JG, Lichtman AH. IL-12 is required for differentiation
of pathogenic CD8� T cell effectors that cause myocarditis. J Clin Invest.2003;111:671–680.
25. Usui S, Hara Y, Hosaki S, Okazaki M. A new on-line dual enzymatic methodfor simultaneous quantification of cholesterol and triglycerides in lipoproteinsby HPLC. J Lipid Res. 2002;43:805–814.
27. Chen R, Lowe L, Wilson JD, Crowther E, Tzeggai K, Bishop JE, Varro R.Simultaneous quantification of six human cytokines in a single sample usingmicroparticle-based flow cytometric technology. Clin Chem. 1999;45:1693–1694.
28. Giulietti A, Overbergh L, Valckx D, Decallonne B, Bouillon R, Mathieu C.An overview of real-time quantitative PCR: applications to quantify cytokinegene expression. Methods. 2001;25:386–401.
29. Schiller NK, Kubo N, Boisvert WA, Curtiss LK. Effect of gamma-irradiationand bone marrow transplantation on atherosclerosis in LDL receptor-deficient mice. Arterioscler Thromb Vasc Biol. 2001;21:1674–1680.
30. McAdam AJ, Chang TT, Lumelsky AE, Greenfield EA, Boussiotis VA,Duke-Cohan JS, Chernova T, Malenkovich N, Jabs C, Kuchroo VK, Ling V,Collins M, Sharpe AH, Freeman GJ. Mouse inducible costimulatorymolecule (ICOS) expression is enhanced by CD28 costimulation and reg-ulates differentiation of CD4� T cells. J Immunol. 2000;165:5035–5040.
31. Bonhagen K, Liesenfeld O, Stadecker MJ, Hutloff A, Erb K, Coyle AJ, LippM, Kroczek RA, Kamradt T. ICOS� Th cells produce distinct cytokines indifferent mucosal immune responses. Eur J Immunol. 2003;33:392–401.
32. Wassink L, Vieira PL, Smits HH, Kingsbury GA, Coyle AJ, Kapsenberg ML,Wierenga EA. ICOS expression by activated human Th cells is enhanced byIL-12 and IL-23: increased ICOS expression enhances the effector functionof both Th1 and Th2 cells. J Immunol. 2004;173:1779–1786.
33. Sakaguchi S. Naturally arising CD4� regulatory T cells for immunologicself-tolerance and negative control of immune responses. Annu Rev Immunol.2004;22:531–562.
34. Fontenot JD, Rudensky AY. A well adapted regulatory contrivance: regu-latory T cell development and the forkhead family transcription factor Foxp3.Nat Immunol. 2005;6:331–337.
35. Akbari O, Freeman GJ, Meyer EH, Greenfield EA, Chang TT, Sharpe AH,Berry G, DeKruyff RH, Umetsu DT. Antigen-specific regulatory T cellsdevelop via the ICOS-ICOS-ligand pathway and inhibit allergen-inducedairway hyperreactivity. Nat Med. 2002;8:1024–1032.
36. Miyamoto K, Kingsley CI, Zhang X, Jabs C, Izikson L, Sobel RA, WeinerHL, Kuchroo VK, Sharpe AH. The ICOS molecule plays a crucial role in thedevelopment of mucosal tolerance. J Immunol. 2005;175:7341–7347.
37. Afek A, Harats D, Roth A, Keren G, George J. A functional role for induciblecostimulator (ICOS) in atherosclerosis. Atherosclerosis. 2005;183:57–63.
38. Li MO, Wan YY, Sanjabi S, Robertson AK, Flavell RA. Transforminggrowth factor-beta regulation of immune responses. Annu Rev Immunol.2006;24:99–146.
39. Chen W, Jin W, Hardegen N, Lei KJ, Li L, Marinos N, McGrady G, WahlSM. Conversion of peripheral CD4�CD25� naive T cells to CD4�CD25�regulatory T cells by TGF-beta induction of transcription factor Foxp3. J ExpMed. 2003;198:1875–1886.
40. Wan YY, Flavell RA. Identifying Foxp3-expressing suppressor T cells witha bicistronic reporter. Proc Natl Acad Sci U S A. 2005;102:5126–5131.
CLINICAL PERSPECTIVECD4� T-lymphocyte–mediated adaptive immune responses contribute significantly to the inflammatory process that occursin atherosclerotic lesions. Costimulatory pathways that involve B7 and CD28 family molecules are essential for theinitiation of most CD4� T-cell responses to antigens and therefore may promote arterial disease. However, recent studieshave shown that the B7-CD28 pathway is also essential for regulatory T-cell development and function and that regulatoryT-cell populations may serve to suppress proatherogenic T cells. Inducible costimulatory molecule (ICOS) is a CD28family member for which in vivo functions are incompletely understood. This study tested the hypothesis that ICOSmodulates proatherogenic T-cell responses. The data indicate that a deficiency in ICOS in mouse bone marrow–derivedcells results in a marked increase in T-cell responses to hypercholesterolemia, associated with more T-cell infiltrates inlesions and more lesion development. The effects of ICOS deficiency are related to a reduction in numbers and functionof regulatory T cells. These data suggest that pharmacological blockade of costimulatory pathways as a potentialtherapeutic strategy for atherosclerosis must be refined to avoid impairment of the beneficial effects of regulatory T cells.
Gotsman et al ICOS and Regulatory T Cells in Atherosclerosis 2055
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
Galina Sukhova, Joseph L. Witztum, Arlene H. Sharpe and Andrew H. LichtmanIsrael Gotsman, Nir Grabie, Rajat Gupta, Rosa Dacosta, Malcolm MacConmara, James Lederer,
Inducible Costimulatory MoleculeImpaired Regulatory T-Cell Response and Enhanced Atherosclerosis in the Absence of
is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/CIRCULATIONAHA.106.633263
2006;114:2047-2055; originally published online October 23, 2006;Circulation.
http://circ.ahajournals.org/content/114/19/2047World Wide Web at:
The online version of this article, along with updated information and services, is located on the
http://circ.ahajournals.org//subscriptions/
is online at: Circulation Information about subscribing to Subscriptions:
http://www.lww.com/reprints Information about reprints can be found online at: Reprints:
document. Permissions and Rights Question and Answer this process is available in the
click Request Permissions in the middle column of the Web page under Services. Further information aboutOffice. Once the online version of the published article for which permission is being requested is located,
can be obtained via RightsLink, a service of the Copyright Clearance Center, not the EditorialCirculationin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:
by guest on May 9, 2016http://circ.ahajournals.org/Downloaded from
