Statins Induce Regulatory Recruitment via a CCL1 Dependent Pathway Emilia Mira, Beatriz León, Domingo F. Barber, Sonia Jiménez-Baranda, Iñigo Goya, Luis Almonacid, Gabriel This information is current as Márquez, Angel Zaballos, Carlos Martínez-A., Jens V. Stein, of September 26, 2021. Carlos Ardavín and Santos Mañes J Immunol 2008; 181:3524-3534; ; doi: 10.4049/jimmunol.181.5.3524

http://www.jimmunol.org/content/181/5/3524 Downloaded from

References This article cites 52 articles, 27 of which you can access for free at: http://www.jimmunol.org/content/181/5/3524.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 26, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2008 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Statins Induce Recruitment via a CCL1 Dependent Pathway1

Emilia Mira,* Beatriz Leo´n,* Domingo F. Barber,* Sonia Jime´nez-Baranda,* In˜igo Goya,* Luis Almonacid,* Gabriel Ma´rquez,† Angel Zaballos,* Carlos Martı´nez-A.,* Jens V. Stein,‡ Carlos Ardavı´n,* and Santos Man˜es2*

The statins, a group of inhibitors of the 3-hydroxy-3-methylglutaryl coenzyme A reductase, are reported to influence a variety of immune system activities through 3-hydroxy-3-methylglutaryl coenzyme A reductase-dependent and -independent mechanisms. How statin treatment regulates immune system function in vivo nonetheless remains to be fully defined. We analyzed the immu- nomodulatory effects of lovastatin in a Candida albicans-induced delayed-type hypersensitivity reaction in mice. In this model, lovastatin administration reduced the acute inflammatory response elicited by C. albicans challenge. This anti-inflammatory activity of lovastatin was associated with a shift from a Th1 to a Th2 immune response, as well as an increase in the percentage Downloaded from of regulatory T cells at the inflammation site and in the regional draining lymph node. The lovastatin-induced increase in regulatory T cells in the inflamed skin was dependent on expression of CCL1, a that is locally up-regulated by statin administration. The anti-inflammatory effect of lovastatin was abrogated in CCL1-deficient mice. These results suggest that local regulation of chemokine expression may be an important process in statin-induced modulation of the immune system. The Journal of Immunology, 2008, 181: 3524–3534. http://www.jimmunol.org/

he statins are used extensively in medical practice be- mokines, master for T cell polarization, metallopro- cause of their ability to reduce cardiovascular mortality teases, and costimulatory and MHC molecules (reviewed in Refs. T and stroke (1). Although this protective activity was ini- 16 and 17). At high in vitro concentrations, statins might interfere tially ascribed to inhibition of cholesterol biosynthesis, it is now with function by altering signaling in cholesterol-en- evident that statins are pleiotropic drugs with immunomodulatory riched lipid rafts (18, 19). and anti-inflammatory properties. Studies in animal models, as Statins are also reported to induce immunomodulation by 3-hy- well as epidemiological and small-scale clinical trials, suggest that droxy-3-methylglutaryl coenzyme A reductase-independent mech- ␤ statin pleiotropism might be clinically relevant in other patholo- anisms. Several statins bind to the 2 integrin LFA-1, preventing by guest on September 26, 2021 gies, including infection and sepsis (2–5), autoimmunity (6–9), its interaction with the ICAM-1 (20). Statin blockade of the LFA- and organ transplantation (10, 11). 1/ICAM-1 interaction has functional implications in T cell activa- The immunomodulatory effect of statins is thought to involve tion, leukocyte diapedesis, and HIV-1 infection (20–22). several mechanisms that act on different cell types (6, 12, 13). The extent to which each of these pleiotropic activities of statins Statins not only prevent de novo synthesis of cholesterol, they also contributes to immunomodulation in vivo is not well understood. decrease the pool of intermediate isoprenoid metabolites, such as Moreover, the reported immunomodulatory activities of statins farnesyl pyrophosphate or geranylgeranyl pyrophosphate, thus af- would be expected to attenuate the immune response. Paradoxi- fecting GTP-binding signaling proteins, including Ras and Rho- cally, the capacity to mount protective immune responses to intra- GTPases (2, 14, 15). Impairment of GTPase function allows statins cellular pathogens is not diminished in statin-treated patients. In- to affect actin-based cytoskeletal remodeling, cell motility, and the deed, regular statin use has been associated with reduced risk of activity of transcription factors that control the expression of che- bacteremia and sepsis (4, 5), as well as lower incidence and pro- gression of specific cancer types (23). We analyzed the in vivo effect of lovastatin in a murine model *Department of Immunology and Oncology, Centro Nacional de Biotecnologı´a/Con- of skin delayed-type hypersensitivity (DTH)3 response to heat- † sejo Superior de Investigaciones Cientificas (CSIC), Madrid, Spain; Genetrix, Ma- inactivated Candida albicans. Skin DTH responses involve Ag drid, Spain; and ‡Theodor Kocher Institute, University of Bern, Bern, Switzerland uptake, processing, and transport by APC to tissue-draining lymph Received for publication February 29, 2008. Accepted for publication June 20, 2008. nodes, where Ag-specific T cells are triggered to initiate a primary The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance immune response. Secondary exposure to the Ag results in a robust with 18 U.S.C. Section 1734 solely to indicate this fact. local inflammatory response. The coordinated cell and 1 This work was supported in part by the Spanish Ministry of Education and Science responses required to induce DTH make this experimental system (SAF2005-00241 and NAN2004-08805-C04/04 to S.M.; SAF2006-07609 to C.A.), ideal for in vivo analysis of the immune response. Fundacio´n para la Investigacio´n y Prevencio´n del SIDA en Espan˜a (36550/06 to S.M.), La Marato TV3 Foundation (to S.M.), the Comunidad de Madrid (DIFHE- Our data show that lovastatin reduced the inflammatory re- MAT-CM S-SAL-0304-2006 to S.M. and C.A.), and the European Union FP6 sponse to secondary Ag exposure, and that this effect was not a (INNOCHEM, FP6 LSHB-CT-2005-518167 to S.M. and C.M.A.). The Department of Immunology and Oncology was founded and is supported by the Spanish National Research Council and by Pfizer. 3 Abbreviations used in this paper: DTH, delayed-type hypersensitivity; Ct, cycle 2 Address correspondence and reprint requests to Dr. Santos Man˜es, Department of threshold; DC, ; PO-LN, popliteal lymph node; SLO, secondary lym- Immunology and Oncology, Centro Nacional de Biotecnologı´a/Spanish National Re- phoid organs; Treg, regulatory T; WT, wild type. search Council, Darwin 3, Madrid E-28049 Spain. E-mail address: smanes@ cnb.csic.es Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00 www.jimmunol.org The Journal of Immunology 3525

Table I. Changes in mRNA levels in the whole footpad after A 4.4 Kb 4.4 Kb C. albicans challenge of lovastatin- and vehicle-treated mice 6.4 Kb SacI SacI BamHI SphI NcoI NcoI BamHI SacI Lovastatin Lovastatin/Vehicle CCL1 WT SphI Molecule (2Ϫ⌬⌬Ct)a Vehicle (2Ϫ⌬⌬Ct)a Ratiob P E1 I1 E2 I2 E3 Probe A Probe B IL-1␣ 2.8 Ϯ 0.4 4.4 Ϯ 1.1 Ϫ1.6 IL-1␤ 25.7 Ϯ 2.3 146 Ϯ 37 Ϫ5.7 IL-2 1.8 Ϯ 0.4 1.7 Ϯ 0.7 ϩ1.1 IL-3 not detected not detected —– IL-4 4.4 Ϯ 1.4 0.42 Ϯ 0.8 ϩ10.5 NotI TaqI BamHI BamHI NotI IL-5 2.2 Ϯ 0.7 0.06 Ϯ 1.5 ϩ36.7 Targeting Neor E3E2I2 HSV-TK Vector IL-6 25.3 Ϯ 1.9 47.5 Ϯ 10.1 Ϫ1.9 Ϯ Ϯ ϩ SphI IL-7 1.2 0.2 0.5 1.0 2.4 IL-9 not detected not detected —– IL-10 13.5 Ϯ 8.8 2.9 Ϯ 1.0 ϩ4.7 IL-12␣ 3.4 Ϯ 0.6 12.8 Ϯ 5.2 Ϫ4.8 8.8 Kb IL-12␤ not detected not detected —– 4.4 Kb TGF-␤ 3.8 Ϯ 1.2 2.8 Ϯ 0.7 ϩ1.4 SacI BamHI TNF-␣ 17.7 Ϯ 6.3 25.7 Ϯ 7.1 Ϫ1.5 SphI NcoI TaqI BamHI SacI CCL1 KO IFN-␥ 25.0 Ϯ 7.1 100.5 Ϯ 64 Ϫ3.2 r gene Neo E3E2I2 CCL1 14.5 Ϯ 6.5 1.7 Ϯ 2.1 ϩ8.5 Downloaded from Probe A SphI Probe B CCL2 13.7 Ϯ 5.8 9.9 Ϯ 3.3 ϩ1.4 CCL3 51.3 Ϯ 8.2 53.5 Ϯ 19.2 Ϫ1.0 BC CCL4 33.0 Ϯ 9.5 79.0 Ϯ 5.7 Ϫ2.8 5’ Analysis 3’ Analysis b.p. CCL5 8.3 Ϯ 2.4 41.8 Ϯ 14 Ϫ5.0 -/--/- +/+ +/+ +/- +/- 390403 444 Kb 403 404 406 Kb 2036 CCL6 5.6 Ϯ 0.7 3.7 Ϯ 0.7 ϩ1.5 1636 Ϯ Ϯ ϩ 9 9 1018 CCL7 25.7 5.2 14.5 3.3 1.8 8 517 Ϯ Ϯ ϩ 7 8 KO allele CCL9/10 8.2 0.9 4.2 0.9 1.9 7 506 Ϯ Ϯ ϩ

6 http://www.jimmunol.org/ 6 396 CCL11 1.5 0.5 0.7 1.3 2.4 5 344 421 bp Ϯ Ϯ ϩ 4 5 CCL17 3.3 0.5 1.5 0.2 2.2 4 298 3 CCL20 3.7 Ϯ 0.3 4.1 Ϯ 0.6 Ϫ1.1 3 CCL22 3.4 Ϯ 1.8 2.3 Ϯ 1.0 ϩ1.5 2 CCL27 0.5 Ϯ 0.5 0.16 Ϯ 1.9 ϩ3.1 2 2036 1636 CXCL1 6.0 Ϯ 2.0 20.2 Ϯ 1.1 Ϫ3.4 1018 517 WT allele CXCL2 32.2 Ϯ 5.5 186 Ϯ 77.2 Ϫ5.8 1 506 Ϯ Ϯ Ϫ 1 396 604 bp CXCL5 33.5 2.5 274 80.8 8.2 344 CXCL9 14.0 Ϯ 2.0 19.6 Ϯ 1.6 Ϫ1.4 SphI Digestion SacI Digestion 298 CXCL10 28.8 Ϯ 7.2 65.2 Ϯ 4.9 Ϫ2.3 Probe A Probe B CXCL11 26.4 Ϯ 1.1 44.8 Ϯ 12.4 ϩ1.7 FIGURE 1. Generation of the CCL1 knockout mouse. A, The WT allele CXCL12 2.5 Ϯ 0.9 2.3 Ϯ 0.3 ϩ1.1 by guest on September 26, 2021 of the CCL1 gene, a fragment of the targeting construct, and the mutant CXCL13 4.6 Ϯ 1.8 4.6 Ϯ 1.4 ϩ1.0 allele are shown. The 5Ј and 3Ј probes used for Southern hybridization are CXCL14 3.9 Ϯ 0.5 4.5 Ϯ 0.7 ϩ1.2 Ϯ Ϯ ϩ indicated. B, Southern blot analysis of SphI- or SacI-restricted genomic CXCL15 3.0 1.6 0.4 1.3 7.5 Ϯ Ϯ Ϫ DNA prepared from targeted embryonic stem cell clones indicates the pres- CXCL16 3.9 0.1 4.5 0.7 1.1 CCR1 13.4 Ϯ 2.5 19.8 Ϯ 8.1 Ϫ1.5 ence of mutant alleles (arrow). Blots of SphI- and SacI-digested DNA were Ϯ Ϯ ϩ 32 CCR7 5.2 1.6 9.6 2.7 1.9 probed with P-labeled probes A (NcoI-NcoI) and B (BamHI-SacI), re- CXCR3 3.5 Ϯ 1.8 2.4 Ϯ 0.7 ϩ1.5 spectively. C, PCR analysis of genomic DNA isolated from CCL1ϩ/ϩ, a CCL1ϩ/Ϫ, and CCL1Ϫ/Ϫ mouse tails was amplified by two separate PCR ⌬⌬Ct coefficient was calculated using ⌬Ct values for the PBS-injected footpad Ϫ⌬⌬Ct Ͼ using specific primers (see Materials and Methods). Polynucleotides of 604 as reference, and the relative expression value was calculated as 2 . Values 1 indicate increase in mRNA expression for the corresponding gene in the C. albicans- and 421 bp were amplified for the WT and targeted alleles, respectively. injected footpad compared with the PBS-injected control footpad; values Ͻ1 indicate down-modulation of the mRNA. Data are mean Ϯ SEM (n ϭ 6). b Quotient between lovastatin and vehicle 2Ϫ⌬⌬Ct values; ϩ, indicates lovastatin/ vehicle ratio; Ϫ, indicates vehicle/lovastatin ratio. consequence of a defective primary response, vascular permeabil- ity, or diapedesis of immune cells. Lovastatin changed the expres- sion of specific cytokines, promoting a shift from a Th1 to a Th2 response. In the Ag-challenged footpad, lovastatin also increased and 3.2-kpb BamHI-BamHI DNA fragments from the 5Ј and 3Ј regions, expression of CCL1, a chemokine that triggers of Th2 respectively. These DNA fragments were then subcloned at either end of a and a subset of T regulatory (Treg) cells (24–29). Lovastatin treat- neomycin resistance gene, under the control of the phosphoglycerate kinase ment increased the percentage of Treg cells at inflammatory sites promoter. The herpes simplex thymidine kinase gene was also fused at the and in regional tissue-draining lymph nodes, compared with vehi- 5Ј end of the cloned CCL1 sequences in the replacement targeting con- cle-treated animals. In contrast, the percentage of Treg cells was struct (Fig. 1A). The resulting plasmid was linearized with NotI and elec- troporated into the 129 SvJ embryonic stem cell line. Gancyclovir- and similar in the inflamed footpads of vehicle- and lovastatin-treated G418-resistant clones were selected, and 15 ␮g of genomic DNA from CCL1-deficient mice, suggesting that statin-induced CCL1 up-reg- each clone was SphIorSacI digested, electrophoresed, and blotted onto ulation affects Treg cell recruitment to the inflammation site. Our Hybond-N membranes (Amersham Biosciences). High stringency hybrid- 32 data highlight control of chemokine expression as a major mech- izations with CCL1-specific P-labeled probes were performed in Rapid- Hyb buffer (Amersham Biosciences) (Fig. 1B). Targeted embryonic cell anism for the immunomodulation induced by the statins in vivo. clones were injected individually into CD1 morulae, which were trans- ferred to pseudopregnant CD1 females, as described (30). Chimeric males Materials and Methods were bred to C57BL/6 females; the offspring was genotyped by Southern Ϫ/Ϫ Generation of CCL1-deficient mice blot (data not shown) and PCR with specific primers (Fig. 1C). CCL1 mice were backcrossed onto the C57BL/6 genetic background for 10 gen- The CCL1 (also known as Scya1) gene was amplified by PCR from 129 erations. The genotype of each animal included in the study was verified by SvJ mouse strain genomic DNA and used to subclone 3.7-kpb NcoI-NcoI PCR using tail DNA. 3526 LOVASTATIN CONTROLS TREG CELL TRAFFICKING

+ + + + + A Lovastatin (10 mg/Kg, i.p.) or Vehicle CD11b CD11b Gr1 CD11b F4/80 12 10 0.15 10 8 0.10 8 6 -7 01234 6 -14 4 4 0.05 2 2

Swelling ) 6 0 0 0 Sensitization Challenge 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 + + T cells CD4 T cells CD8 T cells BC 4 0.1 0.1 ) 1.8 6 20 3 Vehicle 2 0.05 0.05 15 Lovastatin 1.2 1

10 0 0 0 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0.6 - - NK cells Cell number increase (x 10 Cell number CD4 CD8 T cells NKT cells B cells Vehicle 5 4 1 6 0.2 Lovastatin 3 0.8 0 0 4

Swelling Increase (mm) 0.6 0 1 2 3 4 0 1 2 3 4 2 0.1 Days after challenge increase (x 10 Cell number Days after challenge 0.4 2 1 0.2 Downloaded from Vehicle-treated Lovastatin 0 0 0 0 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 D y = 9.31 x – 3.18 y = 11.16 x – 1.83 Days after challenge ρ = 0.56, p < 0.001 ρ = 0.78, p < 0.001 FIGURE 3. Effect of lovastatin on dermis infiltration by specific cell ) 6 30 30 types. Cell infiltrate recovered from the footpad was analyzed by flow 20 20 cytometry after staining with appropriate Abs. The increase in absolute cell number was calculated by subtracting cells from the PBS-injected footpad 10 10 http://www.jimmunol.org/ from cells in the C. albicans-challenged footpad. Results are mean Ϯ SEM 0 0 of five mice per group in a representative experiment (n ϭ 3). Cell number Cell number increase (x 10 012012 Swelling increase (mm) FIGURE 2. Lovastatin reduces dermis swelling and cell infiltration. A, Flow chart illustrating C. albicans-elicited DTH. B and C, Time course of and AF6-120.1), anti-CD86 (GL1), anti-Ly-6C (AL-21), anti-NK1.1 footpad swelling (B) and cell infiltration (C)inC. albicans-elicited DTH (PK136), anti-CD25 (7D4), anti-CD69 (H1.2F3), anti-CD44 (IM7), and response. Swelling increase was calculated for each mouse by subtracting anti-CD62L (MEL-14). To detect intracellular IL-12, cells were first sur- the swelling in the PBS-injected footpad (control) from the swelling in the face stained and subsequently fixed (4% paraformaldehyde in PBS, 4°C, 5 C. albicans-injected footpad. Cell infiltrate was recovered from footpads min), permeabilized (0.02% saponin), and stained with anti-IL-12 p40/p70 by guest on September 26, 2021 after collagenase digestion. The results are mean Ϯ SEM of five mice/ (clone C15.6). Treg cells were stained with the FoxP3 staining set (eBio- ␥ group in one representative experiment of five. D, Linear regression of science). In all cases, FcRs were blocked with anti-Fc RII/III (clone 2.4G2). Flow cytometry was performed in an EPICS XL flow cytometer swelling increase vs the increase in cell number. A statistically significant (Coulter Electronics). correlation coefficient was found for both groups (n ϭ 25). Vascular leakage assay Mice sensitized and challenged as above were injected with FITC-BSA Induction of DTH response to C. albicans (Sigma-Aldrich; 600 ␮gin200␮l of PBS, i.v.); after 90 min, swelling was measured and FITC-BSA accumulation in footpad extracts was quantified All experiments involving the use of animals were supervised by the Cen- (Cytofluor; Millipore). Serum FITC-BSA was measured to normalize the tro Nacional de Biotecnologı´a Ethics Committee, and performed according fluorescence injected into each mouse. Vascular permeability was calcu- to European Union guidelines. Twelve-week-old C57BL/6 female mice lated as the difference between fluorescence in C. albicans- and PBS-in- (The Jackson Laboratory) were rested for 1 wk in an air-filtered enclosure jected footpads. before use; CCL1Ϫ/Ϫ mice were bred in our facility. Mice were treated daily with lovastatin (Sigma-Aldrich; 200 ␮g/mouse in absolute ethanol; T cell chemotaxis 10 mg/kg or 1 mg/kg, i.p.) or vehicle during the experimental period. One week after the start of the treatment, mice were sensitized with heat-inac- The effect of lovastatin on in vitro chemotaxis of T cells toward CCL21 and tivated C. albicans (strain 3153, serotype A; National Collection Patho- CCL1 was analyzed using Transwell assays. Peripheral and mesenteric lymph genic Fungi, Bristol, U.K.; 1 ϫ 108 cells in 200 ␮l of PBS, i.p.), and nodes from 7- to 12-wk-old female BALB/c mice (for CCL21-induced che- challenged 1 wk later in the subplantar tissue of the left hind foot with this motaxis) or PO-LN from C. albicans-challenged mice (for CCL1-induced che- immunogen (1 ϫ 107 cells in 50 ␮l of PBS, s.c.); control right footpads of motaxis) were isolated, and single-cell suspensions were prepared with 5 ϫ each mouse received 50 ␮l of PBS. Duplicate measurements of footpad 106 cells/ml in RPMI 1640/standard supplements containing 1 or 10% FCS, swelling were made using a dial thickness gauge (Mitutoyo) on the 4 con- anti-CD3/CD28 (both at 0.5 ␮g/ml), and indicated lovastatin and/or meval- secutive days postchallenge; the thickness increase was calculated as the onate concentrations. After overnight incubation, cell migration to medium or difference between left and right footpad measurements. 100 nM CCL21 or 10 nM CCL1 (PeproTech) was assessed after2hinTrans- well assays (5 ␮m pore size; Costar). CCR7 and CCR8 levels, or viability, as Flow cytometry analysis assessed by forward/side light scatter characteristics, were unaffected by lova- statin or mevalonate treatment (data not shown). Cell suspensions were prepared from the footpad injection site, draining popliteal lymph node (PO-LN), and . Organs were minced and di- Homing assays gested with collagenase P (Boehringer Mannheim; 2 mg/ml in PBS, 500 ␮l, 37°C) for 10 min (PO-LN and ) or 2 h (footpads) with contin- Chemotaxis in Transwell and homing assays were performed, as described uous agitation. Cell number was estimated in a cell counter (CASY; (31). Eight-week-old BALB/c mice from our animal facility were vehicle or Scha¨rfe System) and stained with the following FITC-, PE-, or SpectralRed lovastatin treated daily for 1 wk before the start of the experiment (see DTH (SPRD)-conjugated Abs (BD Pharmingen, unless indicated): anti-CD3 reaction section above). Single-cell suspensions were obtained from peripheral (145-2C11), anti-CD4 (H129.19), anti-CD8 (53-6.7), anti-B220 (RA3- and mesenteric lymph nodes and labeled with Cell Tracker Green (C7025; 0.3 6B2), anti-CD11b (M1/70), anti-GR1 (RB6-8C5), F/480 (Serotec; CI:A3- mM) or Orange (C2927; 1.5 mM, 45 min 37°C; Molecular Probes). After 1), anti-CD11c (HL3), anti-DEC-205 (NLDC-145), anti-MHCII (M5/114 washing and mixing at a 1:1 ratio, 2.5 ϫ 107 green- and red-labeled cells were The Journal of Immunology 3527

A + AD TotalB cells CD4 T cells Vehicle ) 15 9 2 Lovastatin 6 1.0 Vehicle Lovastatin + Lovastatin Basal 10 6 Mevalonate migration Mevalonate )

6 1 0.4 5 3 + * F4/80 CCL21- (CD11b+) 0 0 0 induced 0 CD11b+ Gr1+ 23 23 23 migration CD8+ T cells CD11b+F4/80+ 0 5 10 15 20 25 - 0.4 3 0.08 0.1 Migrating cells (x 103) increase (x 10 Cell number 0.06 0.08 2 0.06 B Vehicle-treated host C Lovastatin-treated host 0.04 1.5 1.5 0.04

Cell number increase (x 10 Cell number 1 0.02 0.02 1.0 1.0 0 0 0 23 23 23 Days after challenge

Homing ratio 0.5 0.5 Homing ratio B pDCs Dermal DCs Inflammatory DCs 3 10 10 Downloaded from 0 0 spleenPPMLNPLNblood 8 8 spleenPPMLNPLNblood ) 4 2 6 6 FIGURE 5. Effect of lovastatin on leukocyte migration in vitro and in 4 4 vivo. A, T cells isolated from lymph nodes preincubated with vehicle, 1 2 2 lovastatin, mevalonate, or lovastatin plus mevalonate were assayed for che- motaxis toward basal medium or medium containing CCL21 in Transwell 0 0 0 23 23 23 assays. Migrating cells were calculated by FACS analysis; data are mean Ϯ http://www.jimmunol.org/ Ͻ ء + - 1.0 CD8 DCs 0.5 CD8 DCs SD of duplicates in one representative experiment of two ( , p 0.001, two-tailed t test). B and C, Lovastatin does not affect homeostatic T cell 0.8 0.4 homing. T cells isolated from vehicle- or lovastatin-treated mice were dif- 0.6 0.3

Cell number increase (x 10 Cell number ferentially labeled, mixed at a 1:1 ratio, and reinjected into age- and sex- Vehicle 0.4 0.2 matched vehicle-treated (B) or lovastatin-treated (C) mice. Red- and green- Lovastatin 0.2 0.1 labeled adoptively transferred cells in the indicate organs were counted by FACS, and the homing ratio was calculated (see Materials and Methods). 0 0 23 23 Data are representative of one experiment of three (n ϭ 4). D, Lovastatin Days after challenge does not affect leukocyte migration to the peritoneum. Peritoneal exudates by guest on September 26, 2021 FIGURE 4. Effect of lovastatin on draining PO-LN infiltration by spe- were induced by casein injection; cells in exudates were analyzed by FACS cific cell types. A and B, PO-LN at days 2 and 3 after C. albicans challenge with the indicated markers. The graph shows the difference in the number were disaggregated, and cells were analyzed by flow cytometry after ap- of cells recovered in exudates of casein- and PBS-injected mice; data are propriate staining. The increase in absolute cell number was calculated as mean Ϯ SEM (n ϭ 5) in one experiment of three. above. Results are mean Ϯ SEM of four mice per group in a representative experiment (n ϭ 3). TaqMan low-density arrays and quantitative RT-PCR Total RNA from C. albicans- and PBS-injected footpads was purified (TRI-Reagent; Sigma-Aldrich) and cDNA synthesized (high-capacity injected into the tail vein of age- and sex-matched untreated or lovastatin- cDNA archive kit; Applied Biosystems). Quantitative real-time PCR was treated mice. After 2 h, mice were sacrificed, and percentages of red- and performed with TaqMan low-density arrays in a 7900HT sequence detec- green-labeled adoptively transferred cells in peripheral and mesenteric lymph tion system (Applied Biosystems). The complete list of the analyzed nodes, Peyer’s patches, blood, and spleen were calculated by FACS. The ratio is given in Table I; each gene was assayed in duplicate in the same card. of the input population was also determined and used as a correction factor. A Differences in cycle threshold (⌬Ct) for each gene in C. albicans- and homing ratio was calculated for each organ as follows: (percentage of lovas- PBS-injected footpads were calculated using 18S rRNA as an internal con- tatin-treated adoptively transferred cells/percentage of vehicle-treated adop- trol. The ⌬⌬Ct coefficient was then calculated using ⌬Ct values of the ϫ tively transferred cells) (correction factor input). Switching dyes did not PBS-injected footpad as reference, and increase/decrease in mRNA ex- influence homing behavior. pression was calculated as 2Ϫ⌬⌬Ct. Variations in mRNA expression from at least six mice per group were averaged, and the lovastatin/vehicle or ve- Casein-induced peritonitis model hicle/lovastatin quotient was calculated. Peritoneal exudates were induced by casein injection, basically as reported T cell stimulation assays (32). BALB/c mice were injected with a casein sodium solution (Sigma- Aldrich; 0.2% w/v in PBS, 2 ml, i.p.). After 3 h, cells from peritoneal CD4ϩ T cells were isolated from cell suspensions of PO-LN from vehicle- exudates were recovered by washing with 5 ml of PBS and analyzed and lovastatin-treated mice by magnetic cell sorting (MACS; Miltenyi Bio- by FACS. tec) after incubation with biotin anti-CD4 mAb and streptavidin-conju- gated MACS microbeads. Purity of CD4ϩ T cell preparations was Ͼ95%. Adoptive transfer of sensitized T cells CD4ϩ T cells were stimulated for 48 h with anti-CD3␧ (145-2C11) and anti-CD28 (37.51; BD Pharmingen) Abs, and IFN-␥ and IL-4 were mea- C. albicans Lovastatin- and vehicle-treated mice were sensitized with as sured by ELISA (BD Pharmingen). for DTH responses; 1 wk after sensitization, T cells were isolated from splenocytes of these donor mice by negative selection and injected into Immunohistochemistry naive recipient mice (15 ϫ 106 cells, 200 ␮l of RPMI 1640 plus 1% FCS, i.v.). After 24 h, the mice were challenged with the same Ag in the left Footpads were snap frozen in tissue-freezing medium (Jung), and 4-␮m footpad, and the swelling increase was calculated. Unsensitized control cryosections were acetone fixed and stained with anti-CCL1 (R&D Sys- donor mice were included in each experiment. tems), anti-CD31 (MEC13.3; BD Pharmingen), biotinylated anti-CD3␧ 3528 LOVASTATIN CONTROLS TREG CELL TRAFFICKING

ACA B 2 B220+ CD11c+ IL-1β Vehicle Lovastatin IL-4 1.5 40 2400 IL-5 IL-6 30 1800 1 IL10 IL-12α 20 1200 γ 0.5 Vehicle IFN β 10 600 Lovastatin TGF- TNF-α 0 0 0 Concentration (pg/ml) IFNγ Swelling increase (mm) 0 1 2 3 4 -8 -4 0 4 8 12 32 36 IL-4 MHC class II Days after challenge x-Fold mRNA induction (Lovastatin vs. Vehicle)

+ + C B 60 CD4 60 CD8 ) 3 6 Dermal DC 12 Inflammatory DC 3

Vehicle Lovastatin 2 4 8 Vehicle 40 40 Lovastatin cells (x 10 + 1 2 4 Increase in 20 20 IL-12 0 0 0 Downloaded from

% positive cells 23 23 23 0 0 Days after challenge + + + + + + low low FIGURE 7. Lovastatin promotes a Th1-to-Th2 switch in the DTH re- CD69 CD25 CD44 CD69 CD25 CD44 sponse. A, Lovastatin-mediated induction or repression of mRNA for the CD62L CD62L indicated cytokines. Equal amounts of mRNA, isolated from PBS- or C. FIGURE 6. Lovastatin does not affect T cell activation in the sensitization ϩ ϩ albicans-injected footpads from vehicle- and lovastatin-treated mice at day phase. A, MHCII staining of spleen B220 and CD11c cells 1 wk after 2 postchallenge, were analyzed by quantitative RT-PCR using specific priming with C. albicans. MHCII staining of cells from lovastatin-treated probes. The ⌬⌬Ct coefficient was calculated using ⌬Ct values for the PBS- http://www.jimmunol.org/ (dashed line) and vehicle-treated (solid line) mice is shown. B, Activation and ϩ ϩ injected footpad as reference, and mRNA expression was determined as memory markers in CD4 and CD8 T cells isolated from the spleen of 2Ϫ⌬⌬Ct. Values are the relative level of expression between the lovastatin- vehicle- and lovastatin-treated mice 1 wk after C. albicans priming. Data are and vehicle-treated mice (see Materials and Methods), and represent the Ϯ mean SEM of the percentage of double-positive cells for the indicated mean in mRNA expression variation from at least six mice/group. Dotted ϭ ϩ markers (n 5 mice/group in a representative experiment of three). C, Sen- lines mark Ϯ2-fold change. B, CD4 T cells isolated from the draining sitized T cells from vehicle- or lovastatin-treated mice were adoptively trans- PO-LN of vehicle- or lovastatin-treated mice were activated ex vivo, and ferred to untreated, naive recipient mice, which were challenged with the same IL-4 and IFN-␥ protein levels were analyzed by ELISA (mean Ϯ SEM; Ϯ ϭ Ag in the footpad, and swelling was recorded. Data are mean SEM (n 4 n ϭ 4). C, IL-12 expression was determined by intracellular staining of mice/group) in one representative experiment of three. cells isolated from PO-LN of vehicle- and lovastatin-treated mice at the by guest on September 26, 2021 indicated days postchallenge. Data are expressed as the percentage of IL- 12ϩ cells as determined by FACS (mean Ϯ SEM; n ϭ 4). (145-2C11), anti-CCR8 (ab1663; Abcam), and/or FITC anti-FoxP3 (FJK- 16s; eBioscience), as indicated. Specific binding was developed using ap- propriate Alexa 488-conjugated secondary Abs or Cy3 streptavidin. Sections statin- or vehicle-treated mice were primed i.p. with Ag, and chal- were mounted with 4Ј,6Ј-diamidino-2-phenylindole-containing Vectashield (Vector Laboratories), and examined in a Leica microscope at ϫ200 (CCL1/ lenged (s.c.) 1 wk later in the left hind footpad; as an internal CD31 staining) or ϫ400 (FoxP3/CD3 staining) magnification. All images control, the right hind footpad received the same volume of PBS. were acquired using the same ␥ settings and analyzed blind to treatment con- Swelling was measured daily in both footpads for 4 days (Fig. 2B). ditions. CCL1/CD31 area and intensity parameters were measured with NIH- Maximum swelling occurred on days 1 and 2 postchallenge; as ImageJ software, as described (33). In brief, all images collected in each chan- nel were transformed to 8-byte grayscale, and the same black and white predicted, lovastatin consistently reduced, but did not prevent Ag- threshold was applied (upper threshold level 50; lower threshold level 15); the induced inflammation (Fig. 2B). The lovastatin dosage used in threshold level was chosen to eliminate the anti-CCL1 background using sec- these experiments (10 mg/kg) yields plasma levels in mice of the Ϫ Ϫ tions from CCL1 / mice. Particle analysis was performed automatically to active metabolite lovastatin hydroxyacid comparable to those obtain the area stained by each Ab and the intensity value for each spot. found in humans treated with 40 mg of lovastatin (34). Nonethe- FoxP3/CD3-stained cells were scored in six random fields of two different sections for a given mouse; fields with less than three CD3ϩ cells scored were less, a slight reduction in footpad inflammation was also observed not considered for statistical analyses. in mice treated with lovastatin at 1 mg/kg (data not shown). The effect of statins on vascular permeability is debated (17, Statistical analyses 35, 36). To address a possible effect on the vasculature, we The correlation coefficient (r) was calculated between the increments in injected vehicle- and lovastatin-treated mice with FITC-BSA swelling and cell number in the inflamed footpads. The percentage of ϩ and analyzed vascular leakage. Vascular permeability increased FoxP3 cells in the inflamed footpad, the ratio of CCL1/CD31 area, and moderately and transiently at day 2 postchallenge in lovastatin- the CCL1 intensity data were compared using a two-tailed Student’s t test. Data from wild-type (WT) and CCL1Ϫ/Ϫ mice were analyzed by ANOVA, treated compared with vehicle-treated mice; permeability was sub- followed by the nonparametric Student-Newman-Keuls test for multiple com- sequently recovered (data not shown). We next determined cell parisons. In all cases, percentages were square-root transformed before infiltration in swollen footpads after collagenase digestion of analysis. minced dermis. To estimate the Ag-induced increase in cell num- ber, we subtracted the number of cells recovered from PBS-in- Results jected controls from that of C. albicans-challenged footpads. In Effects of lovastatin on inflammation in the DTH reaction vehicle-treated mice, cell numbers increased linearly and peaked at to C. albicans Ags day 3 postchallenge; lovastatin reduced the number of infiltrating To study the immunomodulatory role of lovastatin, we established cells in the footpad from day 2 postchallenge (Fig. 2C). Positive a DTH response to heat-inactivated C. albicans (Fig. 2A). Lova- correlation of swelling and cell number increase was significant for The Journal of Immunology 3529

AB CCL1 CCL1 + CD31

CCL1 FIGURE 8. Lovastatin up-regu- lates CCL1 expression in endothelial CCL3 -/- cells of inflamed footpads. A, CCL4

CCL5 CCL1 Changes in chemokine expression CCL17 were determined by quantitative RT- CCL22 PCR, using the samples in Fig. 7. B, CCL27 CCL1 (red) and CD31 (green) expres- CXCL1 CXCL2 sion was analyzed in sections of CXCL5 inflamed footpads from vehicle- and CXCL15 lovastatin-treated mice. A C. albi- -8 -6 -4 -2 0 2 4 6 8 10 cans-challenged footpad from a x-Fold mRNA induction (Lovastatin vs. Vehicle) CCL1Ϫ/Ϫ mouse was used to deter- mine nonspecific staining by the anti- C D p = 0.02 CCL1 Ab. Representative images (red p = 0.02 20

and merged) are shown (magnifica- Vehicle-treated area + tion, ϫ200). C and D, Quantification 0.2 ) of the normalized CCL1-stained area 3

(C) and mean CCL1 staining intensity 10 Downloaded from (D) of the images in B. Data are 0.1 mean Ϯ SEM of the values obtained x 10 (AU for three different sections from three Mean CCL1 Intensity 0 0 mice for each treatment. Data were Ratio of CCL1/CD31 compared by two-tailed Student’s t test; the probability value is indicated. E 1 Values in C were corrected by the http://www.jimmunol.org/ square-root method before statistical 0.8 analysis. E, ELISA determination of 0.6 Vehicle CCL1 levels in total cell extracts from Lovastatin 0.4 inflamed footpads from vehicle- and Lovastatin-treated lovastatin-treated mice (n ϭ 4). 0.2

CCL1 (pg/20 µ g lysate) 0 by guest on September 26, 2021 both mouse groups (Fig. 2D; p ϭ 0.56 and p ϭ 0.78 for vehicle Lovastatin reduces leukocyte numbers in the regional draining and lovastatin treatment, respectively; p Ͻ 0.001). Because lova- lymph node statin did not affect vascular leakage in our system, these results We analyzed the number and cell types infiltrating the draining suggest that the anti-inflammatory effects of lovastatin are largely PO-LN of Ag-injected footpads from vehicle- and lovastatin- mediated by reducing cell recruitment. treated mice. Analyses were performed at days 2 and 3 postchal- lenge, because maximum swelling and the most acute anti- inflammatory effects of lovastatin were observed at this time. As Lovastatin decreases footpad infiltration of specific leukocyte for the inflamed dermis, lovastatin decreased total cell numbers in types PO-LN (Fig. 4A), but did not alter the relative percentage of each We characterized the leukocyte infiltrate in the footpad by flow cell type (data not shown). Lovastatin reduced the numbers of B ϩ ϩ cytometry after staining cells with the appropriate Abs. The most and T cells (CD4 and CD8 ) on days 2 and 3, whereas a clear abundant dermis-infiltrating cells were myeloid cells (CD11bϩ), reduction was observed in macrophages and monocytes on day 3 specifically and monocytes (CD11bϩGr1ϩ) (Fig. 3). postchallenge. Lovastatin also affected the dendritic cell (DC) sub- CD11bϩGr1ϩ cell number increased with time, whereas macro- types found in the PO-LN (37); the numbers of plasmacytoid DC, Ϫ phages (CD11bϩF4/80ϩ) infiltrated the footpad only transiently at inflammatory, and CD8 DC were reduced only at day 2, whereas ϩ day 3 postchallenge. Most of the CD3ϩTCRϩ T cells infiltrating dermal and CD8 DC numbers were diminished only at day 3 the footpad were CD4ϪCD8Ϫ, and few CD4ϩ and CD8ϩ T cells postchallenge (Fig. 4B). were found; in all cases, T cell infiltration was delayed compared with neutrophils. A similar pattern was observed for NKT cells Lovastatin effects on leukocyte mobilization and T cell activation (CD3ϩNK1.1ϩ); in contrast, NK cell (CD3ϪNK1.1ϩ) infiltration in vivo showed a bell-shaped curve with a maximum at day 2 postchal- Because statins could influence the immune system at different lenge. Finally, B cells were comparable to macrophages in number levels, including leukocyte trafficking, Ag presentation, and T cell and infiltration time course. Lovastatin treatment did not alter the activation, we analyzed the effect of lovastatin treatment on these relative percentage of any cell type in the inflamed footpad (data processes during the DTH reaction. Subtoxic lovastatin doses in- not shown), but drastically reduced absolute numbers of infiltrating hibited T cell chemotaxis toward the chemokine CCL21 in Trans- ϩ Ϫ Ϫ neutrophils, macrophages, CD3 CD4 CD8 T cells, NKT cells, well assays in an 3-hydroxy-3-methylglutaryl coenzyme A reduc- and NK cells (Fig. 3); CD4ϩ, CD8ϩ, and B cells were only min- tase-dependent manner (Fig. 5A). We nonetheless found that T imally affected or unaffected by lovastatin treatment. cells from lovastatin- or vehicle-treated mice showed similar in 3530 LOVASTATIN CONTROLS TREG CELL TRAFFICKING vivo homing to secondary lymphoid organs (SLO) when adop- A B tively transferred into vehicle-treated (Fig. 5B) or lovastatin- treated (Fig. 5C) hosts. Moreover, there was no difference in the 12 1 Day 1 Day 2 number of cells recovered from SLO of untreated or lovastatin- Vehicle 0.8 6 6 Lovastatin V V Lov Lov V V Lov Lov treated hosts (88 Ϯ 13 ϫ 10 vs 91 Ϯ 6 ϫ 10 cells, respec- 8 0.6 tively). In a casein-induced peritonitis assay, we also found that I.B.: FoxP3 /actin ratio + 0.4 lovastatin treatment did not affect mobilization of myeloid cells 4 to the peritoneal cavity compared with vehicle-treated mice I.B.: Actin 0.2 (Fig. 5D). These results suggest that, at this dose, lovastatin FoxP3 0 0 does not affect in vivo mobilization of myeloid or T cells in 12 1 2 Days after challenge physiological conditions. mRNA induction FoxP3 x -fold Days after challenge We analyzed the effects of lovastatin on the APC and/or T cell C D Vehicle Lovastatin compartments. Concurring with previous data (38), lovastatin re- p = 0.03 ϩ 1.6 ) duced MHC class II expression in spleen CD11c cells 1 wk after + priming with C. albicans (Fig. 6A), although the reduction was 1.2 slight at the dose used. Indeed, we found no apparent differences ϩ in the levels (data not shown) or percentages of spleen CD4 and 0.8 cells (CD4 CD8ϩ cells expressing activation (CD69, CD25) and memory +

(CD44, CD62Llow) markers (Fig. 6B), nor were changes observed 0.4 Downloaded from ϩ Ϯ Ϯ ϩ in the percentage of CD4 (12 0.6; 15.4 1.6%) or CD8 cells % FoxP3 0 (8.8 Ϯ 0.4; 11.15 Ϯ 0.1%) in vehicle- and lovastatin-treated mouse CD3 (red), FoxP3 (green) spleens. Vehicle Lovastatin To further address whether lovastatin affected the priming phase of the DTH response, T cells were isolated from vehicle- or lov- E C. albicans-injected PBS-injected footpad footpad astatin-treated mice 1 wk after priming with C. albicans, and adop- 20 http://www.jimmunol.org/ ) tively transferred to naive, untreated mice; 1 day after cell transfer, + mice were injected with C. albicans, and the increment in swelling 15 was measured as a readout of the DTH response. We detected a 10 cells (CD4 similar increment in footpad swelling after the challenge, indepen- + dently of the origin of transferred T cells (Fig. 6C). These results 5 suggest that, in our conditions, lovastatin did not markedly affect primary T cell activation. % FoxP3 0 1 2 2 Days after challenge

Lovastatin induces changes in cytokine and chemokine by guest on September 26, 2021 ϩ ϩ expression FIGURE 9. Lovastatin increases the percentage of CD3 FoxP3 Treg cells in the inflamed footpad and draining lymph node. A, x-fold induction We used real-time quantitative RT-PCR to analyze changes in of FoxP3 mRNA levels in a fraction enriched in lymphoid cells isolated mRNA levels of representative cytokines, , adhesion from the inflamed footpad of vehicle- and lovastatin-treated mice at the molecules, and proteases isolated from total cell lysates from in- indicated days postchallenge. Data are mean Ϯ SEM in one representative flamed footpads of vehicle- and lovastatin-treated mice. Compared experiment of three (n ϭ 3). B, Total cell extracts from the inflamed foot- with the PBS-injected control, C. albicans challenge induced a pads of vehicle (V)- and lovastatin (Lov)-treated mice were resolved by Ͼ2-fold up- or down-modulation in 40 (80%) of the 50 genes SDS-PAGE and immunoblotted sequentially with anti-FoxP3 and anti- analyzed (Table I); 15 (37.5%) of these 40 genes were differen- actin Abs. Left panels, Show representative immunoblots; right panel, shows the mean Ϯ SEM of the FoxP3/actin ratio of the densitometric tially regulated by lovastatin. Among the most important values obtained from four mice/group in two independent experiments. C, changes observed, lovastatin increased mRNA levels for the FACS detection of CD4ϩFoxP3ϩ T cells in the inflamed footpad at day 2 Th2 cytokines IL-4 and IL-5, and reduced levels for the Th1- postchallenge. Data are mean Ϯ SEM of the percentage of FoxP3ϩ cells in associated cytokines IL-12 and IFN-␥ (Fig. 7A). Lovastatin ad- the CD4ϩ population in one representative experiment of two. Value of p ministration thus appears to promote a shift from a Th1- to a was calculated by two-tailed Student’s t test using the square root of the Th2-type immune response, as reported in other in vivo models data. D, Footpad sections from vehicle- or lovastatin-treated mice at day 2 (6, 9, 15, 39). We also found that lovastatin induced down- postchallenge were costained with anti-CD3 (red) and anti-FoxP3 (green) ϩ ϩ ϩ regulation of the proinflammatory cytokine IL-1␤ and up-reg- Abs. Open and solid arrowheads indicate CD3 and CD3 FoxP3 cells, ϩ ϩ ulation of the anti-inflammatory cytokine IL-10 (Fig. 7A). respectively. Magnification, ϫ400. E, CD4 FoxP3 T cells in draining The effect of lovastatin on Th1/Th2 cytokine expression was PO-LN of C. albicans- and PBS-injected footpads were analyzed by FACS ϩ at the indicated days postchallenge. Data show the mean Ϯ SEM of the also analyzed at the protein level in CD4 T cells isolated from ϩ ϩ percentage of CD4 FoxP3 T cells in PO-LN of four mice/group in a the draining PO-LN of control and lovastatin-treated mice after representative experiment (n ϭ 3). in vitro restimulation with anti-CD3 plus anti-CD28 Abs. ELISA analysis of cell culture supernatants showed that T cells from lovastatin-treated mice had higher IL-4 and lower IFN-␥ whereas CCR3, CCR4, and CCR8 are expressed predominantly in levels than those from vehicle-treated animals (Fig. 7B). The Th2 cells (40). Lovastatin treatment reduced the expression of che- percentage of the main IL-12-producing cell types was de- mokines associated with Th1 responses (CCL4, CCL5) and up- creased in lovastatin-treated compared with vehicle-treated regulated Th2-associated chemokines (CCL11, CCL1) (Fig. 8A mouse PO-LN (Fig. 7C). and Table I). Lovastatin also up-regulated CCL27, whose Lovastatin also induced changes in mRNA levels for various (CCR10) is expressed on both Th1 and Th2 cells (41), and inflammatory chemokines in the inflamed footpad. Th1 cells usu- CXCL15, whose receptor remains to be identified. Lovastatin re- ally express the CXCR3, CXCR6, and CCR5 chemokine receptors, duced mRNA levels for the CXCR2 ligands CXCL1, CXCL2, and The Journal of Immunology 3531

ABWT CCL1-/- Vehicle Lovastatin FIGURE 10. The anti-inflammatory effect of lova- a 3 ) 50 b statin requires CCL1 expression. A, Increase in C. al- a a + bicans-induced footpad swelling of vehicle- and lovas- b Vehicle 40 Ϫ/Ϫ tatin-treated WT and CCL1 mice. Mean (bar) and 2 each data point from two independent experiments are 30 cells (CD3

ϩ ϩ + a given (n ϭ 9). B, CD3 FoxP3 cells were quantified in 20 a 1 a footpad sections stained as in Fig. 9D from vehicle- or Ϫ/Ϫ 10

lovastatin-treated WT and CCL1 mice at day 2 post- % FoxP3 Swelling increase (mm) Ϯ 0 Lovastatin 0 challenge. Representative images and the mean SEM -/- -/- ϩ ϩ ϩ WT CCL1 WT CCL1 of the CD3 FoxP3 /CD3 cell ratio are shown (four CD3 (red), FoxP3 (green) sections/mouse from three mice in two independent ex- periments). C, Representative images of footpad sec- WT CCL1-/- ϩ C D tions (day 2 postchallenge) stained for FoxP3 and CCR8ϩ. (For B and C, magnification, ϫ400.) D, Anal- ysis of CD4ϩFoxP3ϩ T cells in the PO-LN draining the 20

) bbb C. albicans-injected footpads in control- and statin- Vehicle + treated WT and CCL1Ϫ/Ϫ mice, as in Fig. 9E. Data are 15 mean Ϯ SEM of nine mice/condition from two inde- a Downloaded from pendent experiments. A–D, Values for bars with the 10 cells (CD4 same lower case letter do not differ significantly from + one another (Student-Newman-Keuls test, p Յ 0.05). 5 Lovastatin

% FoxP3 0 -/- CCR8+ (red), FoxP3+ (green) WT CCL1 http://www.jimmunol.org/

CXCL5, which are chemoattractants for Gr1ϩ cells; this could coinciding with the initiation of CD4ϩ T cell infiltration in the explain the reduced number of Gr1ϩ cells in the inflamed footpads footpad. Accordingly, FACS analysis indicated that lovastatin of lovastatin-treated mice. Other chemokines were only mini- treatment significantly increased the percentage of CD4ϩFoxP3ϩ mally affected by lovastatin (Table I). We found no lovastatin- cells in inflamed footpads (Fig. 9C). Immunohistochemistry induced differences in expression of the adhesion molecules showed a larger number of infiltrating CD3ϩFoxP3ϩ cells in the ICAM-1, ICAM-2, E-selectin, or VCAM-1, or the metallopro- Ag-challenged footpad of lovastatin-treated mice compared with teases matrix metalloproteases-2, -9, -13, or -14 (data not controls (Fig. 9D). ϩ ϩ shown). Collectively, the results indicate that lovastatin pro- Lovastatin also increased the percentage of CD4 FoxP3 Treg by guest on September 26, 2021 duces a shift from Th1- to Th2-associated cytokines, as well as cells in the PO-LN draining the C. albicans-injected compared changes in the expression pattern of chemokines that can affect with vehicle-treated footpads at day 2 postchallenge (Fig. 9E). No- the trafficking of different leukocyte subsets. tably, lovastatin treatment did not augment CD4ϩFoxP3ϩ cell per- centages in naive mouse spleen (data not shown) or in the PO-LN Lovastatin increases the percentage of Treg cells in draining the PBS-injected footpad (Fig. 9E, right). We detected no the inflamed footpad and draining lymph nodes increase in the number or the percentage of CD4ϩFoxP3ϩ cells in Of the chemokines whose expression was altered by lovastatin the peritoneum of lovastatin-treated compared with control mice treatment, we focused on CCL1 because its receptor, CCR8, is (data not shown). These results suggest that lovastatin might spe- associated with Th2 and Treg cell mobilization (26, 29, 42). We cifically affect Treg cell recruitment to inflamed areas and to the analyzed CCL1 protein levels by immunohistochemistry to vali- SLO that drain these sites. date the lovastatin-induced CCL1 mRNA up-regulation described above, using C. albicans-injected footpads of CCL1-deficient mice CCL1 is a key element in the immunomodulatory activity of as a staining control (Fig. 8B). CCL1 staining was weaker in the lovastatin inflamed footpads from Ag-injected, vehicle-treated mice than in To study the functional relationship between lovastatin-induced those from lovastatin-treated animals; in both cases, CCL1 staining CCL1 up-regulation and its anti-inflammatory effect, we analyzed was largely confined to CD31ϩ cells (Fig. 8B). Quantification of the C. albicans-induced DTH response in vehicle- and lovastatin- the images indicated that the CCL1-stained area relative to that of treated CCL1-deficient mice. These mice did not exhibit an overt CD31 (Fig. 8C), as well as CCL1 staining intensity (Fig. 8D) were phenotype; when maintained in barrier isolation, they were healthy significantly higher in inflamed footpads from lovastatin-treated and bred according to Mendelian inheritance patterns. FACS anal- than from vehicle-treated mice. At day 2 postchallenge, ELISA ysis of CD4ϩ, CD8ϩ, and B220ϩ lymphocyte subpopulations measurement of CCL1 levels in footpad cell lysates confirmed the from thymus, lymph nodes, spleen, and peripheral blood indicated lovastatin-induced increase in CCL1 expression in inflamed skin no differences between CCL1-deficient and WT mice (our unpub- (Fig. 8D). We found no changes in CCL1 levels in other organs, lished results). such as noninflamed skin and lungs, as a consequence of lovastatin In contrast to a significant lovastatin-induced reduction of in- treatment (data not shown). flammation in WT mice, we found that DTH responses to C. al- We next measured the levels of forkhead box P3 (FoxP3) tran- bicans were comparable in vehicle- and lovastatin-treated scription factor, a specific marker for the Treg cell lineage. FoxP3 CCL1Ϫ/Ϫ animals (Fig. 10A). Compared with vehicle treatment, mRNA (Fig. 9A) and protein levels (Fig. 9B) were augmented in lovastatin did not increase the percentage of CD3ϩFoxP3ϩ cells the inflamed footpad of lovastatin-treated compared with vehicle- infiltrating the inflamed footpads of CCL1Ϫ/Ϫ mice (Fig. 10B); treated mice. The FoxP3 increase occurred at day 2 postchallenge, this contrasted with the significant lovastatin-triggered increase in 3532 LOVASTATIN CONTROLS TREG CELL TRAFFICKING

CD3ϩFoxP3ϩ cell infiltration of the WT mouse footpad (Fig. tantly, CCR8 is expressed in a large proportion of murine FoxP3ϩ 10B). Moreover, we observed a reduction in the number of T cells, and Treg cells are hyperresponsive to CCL1 (26, 42). ϩ ϩ Ϫ Ϫ CCR8 FoxP3 cells in the inflamed skin of CCL1 / compared Several studies propose that the receptors for inflammatory che- with WT mice, independently of the treatment (Fig. 10C). These mokines (CCR2, CCR4, CCR5, and CCR8) on regulatory cell sub- results suggest that CCL1 deficiency impairs the immunomodula- sets would enable these cells to enter inflamed tissues (24, 26, 47, tory effect of lovastatin by affecting the infiltration of specific cell 48). Our observations suggest that statin-induced up-regulation of types into the inflamed footpad. CCL1 reduces the inflammatory reaction by controlling the traf- We also found that CCL1 deficiency affected Treg cell traf- ficking of specific CCR8ϩ T cell subsets. ficking in the draining PO-LN. Whereas lovastatin administra- In normal skin, CCL1 is undetectable (46) or is barely de- tion induced a significant increase in the percentage of Treg tected in CD31ϩ dermal microvessels (27). A recent report cells in draining PO-LN in WT animals, the percentage of in- showed that C. albicans triggers expression of proinflammatory filtrating Treg cells in draining PO-LN was unaffected by treat- Ϫ/Ϫ genes, including a wide array of chemokines, in primary human ment in CCL1 mice (Fig. 10D). Nonetheless, the percentage endothelial cells (49); notably, C. albicans did not induce CCL1 of Treg cells in the draining PO-LN of vehicle-treated expression in endothelial cells in this study. Accordingly, we CCL1Ϫ/Ϫ mice was significantly higher than that in vehicle- found no increase (Ͻ2-fold) in CCL1 mRNA expression in ve- treated WT animals (Fig. 10D). This suggests that CCL1 defi- hicle-treated C. albicans-injected footpads compared with PBS- ciency could affect Treg cell recruitment into or egression from injected controls, whereas CCL1 expression increased markedly the PO-LN in inflammatory conditions, independently of statin in CD31ϩ cells from lovastatin-treated, yeast-injected footpads treatment. Collectively, our results indicate a major role for Downloaded from CCL1 in the in vivo immunomodulatory and anti-inflammatory (Table I). The selective transcriptional regulation of specific activities of lovastatin in this system. chemokines by statin treatment could explain the CCL1 require- ment for lovastatin activity, despite the redundancy of the che- Discussion mokine system. Indeed, of the receptors for inflammatory che- Several studies have shown that statins are not only lipid-lowering mokines reported in Treg cells (CCR2, CCR4, CCR5, and

drugs, but also are potent modulators of the immune system (16). CCR8), CCL1 was the only ligand clearly up-regulated by in http://www.jimmunol.org/ Given the broad administration of these drugs in the population, it vivo lovastatin treatment; in contrast, CCR4 ligands (CCL17 is extremely important to elucidate how in vivo statin treatment and CCL22) were only minimally affected, and CCR5 agonists affects immune system function. (CCL3, CCL4, and CCL5) were down-regulated in the footpad ϩ To define the mechanisms underlying the immunomodulatory of lovastatin-treated mice (Table I). We observed FoxP3 cells activity of lovastatin, we used a C. albicans-induced DTH model in the inflamed footpad of CCL1-deficient mice, indicating that in mice, which recapitulates the coordinated cell interactions that other chemokine-receptor pairs also contribute to Treg cell in- take place during Th1 immune reactions. In our system, lovastatin filtration independently of statin treatment. attenuated the acute inflammatory response triggered by C. albi- CCR8-expressing FoxP3ϩ cells have been found not only in cans challenge. Concurring with studies in animal models of Th1- inflamed tissues, but also in peripheral lymph nodes (42). In our by guest on September 26, 2021 mediated autoimmune diseases (6, 9, 15), we found that lovastatin model, lovastatin treatment increased Treg cells both in the drain- treatment triggered a decrease in production of proinflammatory ing PO-LN and in the footpad of WT mice. The percentage of Treg cytokines and promoted a Th2-biased DTH response. We also de- cells in the inflamed footpad did not increase following statin treat- ϩ tected a slight reduction in MHCII expression in CD11c cells in ment in CCL1-deficient mice, suggesting a role for the CCL1- lovastatin-treated mice (Fig. 6A). This decrease appears not to af- CCR8 axis. In contrast, the Treg cell increase in statin-treated fect the inductive phase of the DTH response, however, because mouse lymph nodes is probably CCL1-CCR8 independent. In- splenocytes isolated from lovastatin- or vehicle-treated mice and deed, CCL1 deficiency enhanced the percentage of Treg cells in adoptively transferred to naive recipients responded equally well the draining PO-LN, independently of lovastatin administra- to antigenic challenge (Fig. 6C). The lovastatin dose used neither tion. These results suggest that Treg cell homing to PO-LN altered homeostatic T cell homing to SLO, in contrast to other involves other chemokine/receptor pairs (50, 51). Although fur- studies (22), nor affected in vivo mobilization of myeloid cells to ther research is needed to understand the mechanism underlying the peritoneum (Fig. 5). Nonetheless, lovastatin treatment inhibited these differences, CCL1 deficiency might impair Treg cell traf- CCL21- and CCL1-induced chemotaxis of T cells in vitro (Fig. 5A, ficking between lymph nodes and inflamed tissue, as reported and data not shown). for CCR7 (52), or CCL1 may have a role in Treg cell emigra- In addition to the reported effects of statin administration, our tion from lymphoid organs. model showed that lovastatin promoted an increase in the percent- ϩ ϩ In a recent report, Mausner-Fainberg et al. (53) suggested that age of CD3 FoxP3 cells at the inflammation site and in the certain statins enhance differentiation of human CD4ϩCD25Ϫ draining PO-LN. The lovastatin-induced increase in Treg cells at Ϫ ϩ ϩ ϩ the inflamed area was dependent on expression of CCL1, a che- FoxP3 T cells to CD4 CD25 FoxP3 Treg cells. This raised the mokine locally up-regulated by statin administration. Strikingly, question as to whether the lovastatin-induced increase in Treg cells the anti-inflammatory effect of lovastatin was abrogated in CCL1- in PO-LN and inflamed tissue in our DTH model is a consequence deficient mice, suggesting that the CCL1-mediated increase in not only of differential trafficking, but also of a direct effect of Treg cells in the inflamed footpad might be critical for the anti- statin treatment on Treg cell expansion. Nonetheless, these authors ϩ Ϫ Ϫ inflammatory activity of lovastatin. did not find statin-induced conversion of CD4 CD25 FoxP3 to To date, CCL1 is the only known mammalian ligand for CCR8 Treg cells when T cells were isolated from C57BL/6 mice (53), (43), a receptor preferentially expressed on Th2 compared with suggesting that lovastatin would not induce Treg cell expansion/ Th1 effector cells (29, 42). Although the role of CCR8 in mediat- differentiation in our system. Consistent with this idea, we found ing Th2 cell recruitment is debated (43–45), two reports assign a no changes in natural Treg cell numbers in the spleens of naive major role to CCL1 in CCR8ϩ Th2 cell trafficking to the skin in mice or in the draining PO-LN of PBS-injected footpads of homeostatic and inflammatory conditions (27, 46). More impor- lovastatin-treated mice. The Journal of Immunology 3533

In summary, statins would thus modulate the immune system virion-associated host intercellular adhesion molecule 1 and its natural cell sur- through coordinated activity in various arms of the immune reac- face ligand LFA-1. J. Virol. 78: 12062–12065. 22. Schramm, R., M. D. Menger, Y. Harder, R. Schmits, O. Adam, tion, including terminal differentiation of immune cells, as well as G. Weitz-Schmidt, and H. J. Schafers. 2007. Statins inhibit lymphocyte homing the regulation of the molecules that steer these cells to the sites at to peripheral lymph nodes. Immunology 120: 315–324. which they exert their immunomodulatory activity. 23. Demierre, M. F., P. D. Higgins, S. B. Gruber, E. Hawk, and S. M. Lippman. 2005. Statins and cancer prevention. Nat. Rev. Cancer 5: 930–942. 24. Annunziato, F., L. Cosmi, F. Liotta, E. Lazzeri, R. Manetti, V. Vanini, Acknowledgments P. Romagnani, E. Maggi, and S. Romagnani. 2002. Phenotype, localization, and We are grateful to A. Viola and R. Varona for critical reading of the mechanism of suppression of CD4ϩCD25ϩ human thymocytes. J. Exp. Med. manuscript and helpful comments. We thank F. Ortego for statistical anal- 196: 379–387. yses; L. Go´mez, C. M. Martı´nez, and C. Contreras for technical assistance; 25. Colantonio, L., H. Recalde, F. Sinigaglia, and D. D’Ambrosio. 2002. Modulation of expression and chemotactic responsiveness during differ- M. C. Moreno for FACS analyses; and C. Mark for editorial assistance. entiation of human naive T cells into Th1 or Th2 cells. Eur. J. Immunol. 32: 1264–1273. Disclosures 26. Iellem, A., M. Mariani, R. Lang, H. Recalde, P. Panina-Bordignon, F. Sinigaglia, The authors have no financial conflict of interest. and D. D’Ambrosio. 2001. Unique chemotactic response profile and specific expression of chemokine receptors CCR4 and CCR8 by CD4ϩCD25ϩ regulatory T cells. J. Exp. Med. 194: 847–853. References 27. Schaerli, P., L. Ebert, K. Willimann, A. Blaser, R. S. Roos, P. Loetscher, and 1. Tobert, J. A. 2003. Lovastatin and beyond: the history of the HMG-CoA reduc- B. Moser. 2004. A skin-selective homing mechanism for human immune sur- tase inhibitors. Nat. Rev. Drug Discov. 2: 517–526. veillance T cells. J. Exp. Med. 199: 1265–1275. 2. Del Real, G., S. Jime´nez-Baranda, E. Mira, R. A. Lacalle, P. Lucas, 28. Sebastiani, S., P. Allavena, C. Albanesi, F. Nasorri, G. Bianchi, C. Traidl, C. Go´mez-Mouto´n, M. Alegret, J. M. Pen˜a, M. Rodrı´guez-Zapata, S. Sozzani, G. Girolomoni, and A. Cavani. 2001. Chemokine receptor expression M. Alvarez-Mon, et al. 2004. Statins inhibit HIV-1 infection by down-regulating and function in CD4ϩ T with regulatory activity. J. Immunol. 166: Downloaded from Rho activity. J. Exp. Med. 200: 541–547. 996–1002. 3. Erkkila, L., M. Jauhiainen, K. Laitinen, K. Haasio, T. Tiirola, P. Saikku, and 29. Zingoni, A., H. Soto, J. A. Hedrick, A. Stoppacciaro, C. T. Storlazzi, M. Leinonen. 2005. Effect of simvastatin, an established lipid-lowering drug, on F. Sinigaglia, D. D’Ambrosio, A. O’Garra, D. Robinson, M. Rocchi, et al. 1998. pulmonary Chlamydia pneumoniae infection in mice. Antimicrob. Agents Che- The chemokine receptor CCR8 is preferentially expressed in Th2 but not Th1 mother. 49: 3959–3962. cells. J. Immunol. 161: 547–551. 4. Gupta, R., L. C. Plantinga, N. E. Fink, M. L. Melamed, J. Coresh, C. S. Fox, N. W. Levin, and N. R. Powe. 2007. Statin use and hospitalization for sepsis in 30. Torres, M. 1998. The use of embryonic stem cells for the genetic manipulation of patients with chronic kidney disease. J. Am. Med. Assoc. 297: 1455–1464. the mouse. Curr. Top. Dev. Biol. 36: 99–114. http://www.jimmunol.org/ 5. Hackam, D. G., M. Mamdani, P. Li, and D. A. Redelmeier. 2006. Statins and 31. Nombela-Arrieta, C., R. A. Lacalle, M. C. Montoya, Y. Kunisaki, D. Megias, sepsis in patients with cardiovascular disease: a population-based cohort analysis. M. Marques, A. C. Carrera, S. Man˜es, Y. Fukui, C. Martı´nez-A., and J. V. Stein. Lancet 367: 413–418. 2004. Differential requirements for DOCK2 and phosphoinositide-3-kinase ␥ dur- 6. Aktas, O., S. Waiczies, A. Smorodchenko, J. Dorr, B. Seeger, T. Prozorovski, ing T and B lymphocyte homing. Immunity 21: 429–441. S. Sallach, M. Endres, S. Brocke, R. Nitsch, and F. Zipp. 2003. Treatment of 32. Metcalf, D., L. Robb, A. R. Dunn, S. Mifsud, and L. Di Rago. 1996. Role of relapsing paralysis in experimental encephalomyelitis by targeting Th1 cells granulocyte- colony-stimulating factor and granulocyte colony-stim- through atorvastatin. J. Exp. Med. 197: 725–733. ulating factor in the development of an acute inflammatory response in 7. Lawman, S., C. Mauri, E. C. Jury, H. T. Cook, and M. R. Ehrenstein. 2004. mice. Blood 88: 3755–3764. Atorvastatin inhibits autoreactive activation and delays lupus development 33. Mira, E., R. A. Lacalle, J. M. Buesa, G. G. de Buitrago, S. Jime´nez-Baranda, in New Zealand Black/White F1 mice. J. Immunol. 173: 7641–7646. C. Go´mez-Mouto´n, C. Martı´nez-A., and S. Man˜es. 2004. Secreted MMP9 pro- 8. McCarey, D. W., I. B. McInnes, R. Madhok, R. Hampson, O. Scherbakov, motes more efficiently than constitutive active MMP9 bound to the

I. Ford, H. A. Capell, and N. Sattar. 2004. Trial of atorvastatin in rheumatoid tumor cell surface. J. Cell Sci. 117: 1847–1857. by guest on September 26, 2021 arthritis (TARA): double-blind, randomized placebo-controlled trial. Lancet 363: 34. Gegg, M. E., R. Harry, D. Hankey, H. Zambarakji, G. Pryce, D. Baker, 2015–2021. P. Adamson, V. Calder, and J. Greenwood. 2005. Suppression of autoimmune 9. Youssef, S., O. Stuve, J. C. Patarroyo, P. J. Ruiz, J. L. Radosevich, E. M. Hur, retinal disease by lovastatin does not require Th2 cytokine induction. J. Immunol. M. Bravo, D. J. Mitchell, R. A. Sobel, L. Steinman, and S. S. Zamvil. 2002. The 174: 2327–2335. HMG-CoA reductase inhibitor, atorvastatin, promotes a Th2 bias and reverses 35. Van Nieuw Amerongen, G. P., M. A. Vermeer, P. Negre-Aminou, J. Lankelma, paralysis in central nervous system autoimmune disease. Nature 420: 78–84. J. J. Emeis, and V. W. van Hinsbergh. 2000. Simvastatin improves disturbed 10. Kobashigawa, J. A., J. D. Moriguchi, H. Laks, L. Wener, A. Hage, endothelial barrier function. Circulation 102: 2803–2809. M. A. Hamilton, G. Cogert, A. Marquez, M. E. Vassilakis, J. Patel, and L. Yeatman. 2005. Ten-year follow-up of a randomized trial of pravastatin in 36. Wei, C. Y., K. C. Huang, Y. H. Chou, P. F. Hsieh, K. H. Lin, and W. W. Lin. heart transplant patients. J. Heart Lung Transplant. 24: 1736–1740. 2006. The role of Rho-associated kinase in differential regulation by statins of ␤ ␬ 11. Masterson, R., T. Hewitson, M. Leikis, R. Walker, S. Cohney, and G. Becker. -1 - and lipopolysaccharide-mediated nuclear factor B activation and 2005. Impact of statin treatment on 1-year functional and histologic renal allo- inducible nitric-oxide synthase gene expression in vascular smooth muscle cells. graft outcome. Transplantation 80: 332–338. Mol. Pharmacol. 69: 960–967. 12. Dunn, S. E., S. Youssef, M. J. Goldstein, T. Prod’homme, M. S. Weber, 37. Ardavı´n, C. 2007. Subpopulations and differentiation of mouse dendritic cells. In S. S. Zamvil, and L. Steinman. 2006. Isoprenoids determine Th1/Th2 fate in Advances in Molecular and Cellular Microbiology: Dendritic Cell Interactions pathogenic T cells, providing a mechanism of modulation of autoimmunity by with Bacteria. M. Rescigno, ed. Cambridge University Press, Cambridge, pp. atorvastatin. J. Exp. Med. 203: 401–412. 3–26. 13. Yilmaz, A., C. Reiss, A. Weng, I. Cicha, C. Stumpf, A. Steinkasserer, 38. Kwak, B., F. Mulhaupt, S. Myit, and F. Mach. 2000. Statins as a newly recog- W. G. Daniel, and C. D. Garlichs. 2006. Differential effects of statins on relevant nized type of immunomodulator. Nat. Med. 6: 1399–1402. functions of human -derived dendritic cells. J. Leukocyte Biol. 79: 39. Arora, M., L. Chen, M. Paglia, I. Gallagher, J. E. Allen, Y. M. Vyas, A. Ray, and 529–538. P. Ray. 2006. Simvastatin promotes Th2-type responses through the induction of 14. Cordle, A., J. Koenigsknecht-Talboo, B. Wilkinson, A. Limpert, and G. Landreth. the chitinase family member Ym1 in dendritic cells. Proc. Natl. Acad. Sci. USA 2005. Mechanisms of statin-mediated inhibition of small G-protein function. 103: 7777–7782. J. Biol. Chem. 280: 34202–34209. 40. Moser, B., M. Wolf, A. Walz, and P. Loetscher. 2004. Chemokines: multiple 15. Greenwood, J., L. Steinman, and S. S. Zamvil. 2006. Statin therapy and autoim- levels of leukocyte migration control. Trends Immunol. 25: 75–84. mune disease: from protein prenylation to immunomodulation. Nat. Rev. Immu- 41. Vestergaard, C., M. Deleuran, B. Gesser, and C. Gronhoj Larsen. 2003. Expres- nol. 6: 358–370. sion of the T-helper 2-specific chemokine receptor CCR4 on CCR10-positive 16. Blanco-Colio, L. M., J. Tunon, J. L. Martin-Ventura, and J. Egido. 2003. Anti- lymphocytes in atopic dermatitis skin but not in psoriasis skin. Br. J. Dermatol. inflammatory and immunomodulatory effects of statins. Kidney Int. 63: 12–23. 149: 457–463. 17. Greenwood, J., and J. C. Mason. 2007. Statins and the vascular endothelial in- flammatory response. Trends Immunol. 28: 88–98. 42. Soler, D., T. R. Chapman, L. R. Poisson, L. Wang, J. Cote-Sierra, M. Ryan, A. McDonald, S. Badola, E. Fedyk, A. J. Coyle, et al. 2006. CCR8 expression 18. Ehrenstein, M. R., E. C. Jury, and C. Mauri. 2005. Statins for atherosclerosis: as ϩ good as it gets? N. Engl. J. Med. 352: 73–75. identifies CD4 memory T cells enriched for FOXP3 regulatory and Th2 effector 19. Jury, E. C., D. A. Isenberg, C. Mauri, and M. R. Ehrenstein. 2006. Atorvastatin lymphocytes. J. Immunol. 177: 6940–6951. restores Lck expression and lipid raft-associated signaling in T cells from patients 43. Goya, I., J. Gutierrez, R. Varona, L. Kremer, A. Zaballos, and G. Marquez. 1998. with systemic lupus erythematosus. J. Immunol. 177: 7416–7422. Identification of CCR8 as the specific receptor for the human ␤-chemokine I-309: 20. Weitz-Schmidt, G., K. Welzenbach, V. Brinkmann, T. Kamata, J. Kallen, cloning and molecular characterization of murine CCR8 as the receptor for C. Bruns, S. Cottens, Y. Takada, and U. Hommel. 2001. Statins selectively inhibit TCA-3. J. Immunol. 160: 1975–1981. leukocyte function antigen-1 by binding to a novel regulatory integrin site. Nat. 44. Chensue, S. W., N. W. Lukacs, T. Y. Yang, X. Shang, K. A. Frait, S. L. Kunkel, Med. 7: 687–692. T. Kung, M. T. Wiekowski, J. A. Hedrick, D. N. Cook, et al. 2001. Aberrant in 21. Giguere, J. F., and M. J. Tremblay. 2004. Statin compounds reduce human im- vivo T helper type 2 cell response and impaired eosinophil recruitment in CC munodeficiency virus type 1 replication by preventing the interaction between chemokine receptor 8 knockout mice. J. Exp. Med. 193: 573–584. 3534 LOVASTATIN CONTROLS TREG CELL TRAFFICKING

45. Gonzalo, J. A., Y. Qiu, J. M. Lora, A. Al-Garawi, J. L. Villeval, J. A. Boyce, signaling pathways to establish a proinflammatory gene expression program in A. C. Martinez, G. Marquez, I. Goya, Q. Hamid, et al. 2007. Coordinated in- primary human endothelial cells. J. Immunol. 179: 8435–8445. volvement of mast cells and T cells in allergic mucosal inflammation: critical role 50. Huehn, J., K. Siegmund, J. C. Lehmann, C. Siewert, U. Haubold, M. Feuerer, of the CC chemokine ligand 1:CCR8 axis. J. Immunol. 179: 1740–1750. G. F. Debes, J. Lauber, O. Frey, G. K. Przybylski, et al. 2004. Developmental 46. Gombert, M., M. C. Dieu-Nosjean, F. Winterberg, E. Bunemann, R. C. Kubitza, stage, phenotype, and migration distinguish naive- and effector/memory-like L. Da Cunha, A. Haahtela, S. Lehtimaki, A. Muller, J. Rieker, et al. 2005. CCL1- CD4ϩ regulatory T cells. J. Exp. Med. 199: 303–313. CCR8 interactions: an axis mediating the recruitment of T cells and Langerhans-type 51. Schneider, M. A., J. G. Meingassner, M. Lipp, H. D. Moore, and A. Rot. 2007. dendritic cells to sites of atopic skin inflammation. J. Immunol. 174: 5082–5091. CCR7 is required for the in vivo function of CD4ϩ CD25ϩ regulatory T cells. 47. Bystry, R. S., V. Aluvihare, K. A. Welch, M. Kallikourdis, and A. G. Betz. 2001. J. Exp. Med. 204: 735–745. B cells and professional APCs recruit regulatory T cells via CCL4. Nat. Immunol. 52. Menning, A., U. E. Hopken, K. Siegmund, M. Lipp, A. Hamann, and J. Huehn. 2: 1126–1132. 2007. Distinctive role of CCR7 in migration and functional activity of naive- and 48. Iellem, A., L. Colantonio, and D. D’Ambrosio. 2003. Skin-versus gut-skewed effector/memory-like Treg subsets. Eur. J. Immunol. 37: 1575–1583. homing receptor expression and intrinsic CCR4 expression on human peripheral 53. Mausner-Fainberg, K., G. Luboshits, A. Mor, S. Maysel-Auslender, blood CD4ϩCD25ϩ suppressor T cells. Eur. J. Immunol. 33: 1488–1496. A. Rubinstein, G. Keren, and J. George. 2007. The effect of HMG-CoA reductase 49. Mu¨ller, V., D. Viemann, M. Schmidt, N. Endres, S. Ludwig, M. Leverkus, inhibitors on naturally occurring CD4ϩCD25ϩ T cells. Atherosclerosis 197: J. Roth, and M. Goebeler. 2007. Candida albicans triggers activation of distinct 829–839. Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021