Evidence for Involvement of CCR8 by Pathogenic CD4 T Cells in Type 1

Home , CCL25, CCR8 (gene), Chemokine

Recruitment and Activation of Macrophages by Pathogenic CD4 T Cells in Type 1 Diabetes: Evidence for Involvement of CCR8 and CCL1 This information is current as of September 25, 2021. Joseph Cantor and Kathryn Haskins J Immunol 2007; 179:5760-5767; ; doi: 10.4049/jimmunol.179.9.5760 http://www.jimmunol.org/content/179/9/5760 Downloaded from

References This article cites 36 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/179/9/5760.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 25, 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 © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Recruitment and Activation of Macrophages by Pathogenic CD4 T Cells in Type 1 Diabetes: Evidence for Involvement of CCR8 and CCL11

Joseph Cantor and Kathryn Haskins2

Adoptive transfer of diabetogenic CD4 Th1 T cell clones into young NOD or NOD.scid recipients rapidly induces onset of diabetes and also provides a system for analysis of the pancreatic infiltrate. Although many reports have suggested a role for macrophages in the inflammatory response, there has been little direct characterization of macrophage activity in the pancreas. We showed previously that after migration to the pancreas, diabetogenic CD4 T cell clones produce a variety of inflammatory cytokines and chemokines, resulting in the recruitment of macrophages. In this study, we investigated mechanisms by which macrophages are recruited and activated by T cells. Analysis of infiltrating cells after adoptive transfer by the diabetogenic T cell clone BDC-2.5 Downloaded from indicates that large numbers of cells staining for both F4/80 and CD11b are recruited into the pancreas where they are activated to make IL-1␤, TNF-␣, and NO, and express the chemokine receptors CCR5, CXCR3, and CCR8. Diabetogenic CD4 T cell clones produce several inflammatory chemokines in vitro, but after adoptive transfer we found that the only chemokine that could be detected ex vivo was CCL1. These results provide the first evidence that CCR8/CCL1 interaction may play a role in type 1 diabetes through macrophage recruitment and activation. The Journal of Immunology, 2007, 179: 5760–5767. http://www.jimmunol.org/ doptive transfer of diabetogenic CD4 Th1 T cell clones contributions of donor T cells and recruited host cells through into young NOD or NOD.scid recipients rapidly induces analysis of cell surface phenotype and intracellular cytokine stain- A onset of diabetes and also provides a system for analysis ing. We have shown previously that diabetogenic CD4 T cell of the cellular components of the pancreatic inflammatory infiltrate clones produce a variety of inflammatory cytokines and chemo- (1, 2). These studies and others have established that there are kines after migration to the pancreas and, furthermore, that this ϩ ϩ large numbers of F4/80 and CD11b macrophages in the pan- activity results in the recruitment of large numbers of macrophages creatic infiltrates of NOD mice, both in spontaneous disease and in (6). These results have led us to hypothesize that recruitment and adoptive transfers of diabetogenic T cells (2, 3). The requirement activation of macrophages could be an important manifestation of 3 for macrophages in the pathogenesis of type 1 diabetes (T1D) has CD4 T cell effector function. We report here on the further char- by guest on September 25, 2021 been shown in studies in which disease was inhibited by their acterization of the macrophage component of the inflammatory depletion (4, 5). Although macrophages may be necessary for Ag infiltrate and on the analysis of the mechanisms by which macro- presentation to autoreactive T cells, other studies have suggested phages are recruited and activated by T cells. We show that mac- that they also function as effector cells in the destruction of islet ␤ rophages recruited and activated by pathogenic CD4 T cells ex- cells (5). There has been little evidence, however, to directly dem- press several important chemokine receptors involved in onstrate the effector function of macrophages in the diabetic pan- autoimmunity, including CCR8, which has recently been identified creas, and indeed, macrophage activity in the pancreatic infiltrate as a key molecule on activated microglia and macrophages in brain is almost completely uncharacterized at the single-cell level. lesions of patients with multiple sclerosis (7). Upon activation in Using an adoptive transfer system in which the activity of patho- the pancreas, macrophages are induced to make several inflamma- genic vs nonpathogenic T cell clones in young NOD or NOD.scid tory mediators, including chemokines that act as chemoattractants mice is compared, we have developed methods for the recovery for other immune cells. Our results include new and previously and ex vivo analysis of cells just before diabetes onset. These unreported findings on the properties of macrophages as effectors procedures allow us to investigate in an unequivocal manner the in ␤ cell destruction and establish these immune cells as key play- ers in pathogenesis of diabetes.

Department of Immunology, University of Colorado Health Sciences Center, Denver, CO 80206 Materials and Methods Received for publication December 29, 2006. Accepted for publication August 21, 2007. Mice The costs of publication of this article were defrayed in part by the payment of page NOD and NOD.scid breeding mice were initially acquired from The Jack- charges. This article must therefore be hereby marked advertisement in accordance son Laboratory or the Barbara Davis Center for Childhood Diabetes and with 18 U.S.C. Section 1734 solely to indicate this fact. were housed in speciﬁc pathogen-free conditions at the University of Col- 1 This work was supported by National Institutes of Health Grant R01DK50561. orado Health Sciences Center for Laboratory Animal Care. NOD.scid mice 2 Address correspondence and reprint requests to Dr. Kathryn Haskins, Department of were housed in sterile isolation cages. Mice in NOD.scid litters (6–10 days Immunology, University of Colorado Health Sciences Center, National Jewish Med- old) were used as recipients in adoptive transfer experiments. Breeding ical & Research Center (NJMRC), 1400 Jackson Street, Denver, CO 80206. E-mail mice and experimental animals were monitored for development of disease address: [email protected] by urine glucose. The 6.9 TCR transgenic (TCR-Tg) mouse was produced 3 Abbreviations used in this paper: T1D, type 1 diabetes; Tg, transgenic. using TCR genes from a diabetogenic T cell clone, BDC-6.9 (8). All procedures used were in accordance with institutional IACUC guidelines and Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 approved by the UCHSC Animal Care and Use Committee. www.jimmunol.org The Journal of Immunology 5761

Culture and expansion of T cell clones T cell clones were established from spleen and lymph nodes of diabetic NOD mice (9, 10) and were restimulated every 2 wk with a ␤ cell granule membrane fraction obtained from ␤ cell tumors as a source of Ag (11), irradiated NOD spleen cells as APCs, and EL-4 supernatant as a source of IL-2 in complete medium (CM). CM is DMEM supplemented with 44 mM sodium bicarbonate, 0.55 mM L-arginine, 0.27 mM L-asparagine, 1.5 mM L-glu- tamine, 1 mM sodium pyruvate, 50 mg/L gentamicin sulfate, 50 ␮M 2-ME, 10 mM HEPES, and 10% FCS. Cell numbers were expanded for transfer experiments by subculturing 3–6 ϫ 106 T cells 4 days after restimulation in a 5-fold volume of CM and additional IL-2. T cells were harvested, washed 3 times, and resuspended in HBSS for injection into young (Յ10 days of age) NOD.scid recipients. Assessment of cytokines and chemokines in vitro by intracellular staining Production of cytokines and chemokines by T cells in vitro was analyzed by intracellular cytokine staining, as described previously (6). In brief, T cells were stimulated in plates coated with 1 ␮g/ml anti-CD3 Ab for 24 h before Ab staining. The cells were surface stained in 50–100 ␮l of staining buffer (PBS, 0.5% BSA) containing rat anti-CD4 or isotype control Ab for 30–45 min, and then washed and ﬁxed in 2% formaldehyde. Cells were Downloaded from resuspended in permeabilization buffer (staining buffer plus 0.5% saponin), containing an isotype control or speciﬁc Ab mix for intracellular cytokines/ FIGURE 1. Macrophage recruitment to the pancreas by a pathogenic vs chemokines. Polyclonal intracellular staining Abs used were obtained from ϩ R&D Systems and included polyclonal goat IgG anti-CCL1 (TCA-3), anti- a nonpathogenic T cell clone. A, The percentage of CD11b cells (mac- CCL3 (MIP-1␣), anti-CCL4 (MIP-1␤), anti-CCL5 (RANTES), anti-CCL6 rophages) in the pancreas was assessed at 6 to 7 days after transfer of a (C10), anti-CCL9/10 (MIP-1␥), and anti-CCL21 (SLC) as primary Abs, pathogenic T cell clone, BDC-2.5, or a nonpathogenic control, the Th1 T

followed by FITC-rabbit anti-goat secondary Ab (Vector Laboratories). cell clone, BDC-2.4. Macrophage infiltration could be detected as early as http://www.jimmunol.org/ Monoclonal digoxigenin-MTAC-2 anti-lymphotactin and Cy5-anti- 2 days after T cell clone transfer, but was considered optimal at 6 days. digoxigenin secondary Abs were provided by B. Dorner (12). Because recipient mice start becoming diabetic at about 1 week after trans- Adoptive transfer of diabetes fer, we did not go out to further time points. After removal and digestion of pancreata, cells were stained for the CD11b macrophage marker and For disease transfer experiments, expanded cell cultures were harvested analyzed by flow cytometry. Cumulative data from two separate experi- ϫ 7) and T cells (1 10 were injected i.p. into age-matched 6- to 10-day-old ments was used for the bar graph, and error bars represent SEM from 4 to .scid p ϭ 0.00015 ϫ 2-tailed t test. B, F4/80 vs CD11b ,ء .NOD recipient mice. In some experiments, adoptive transfers were 5 mice for each clone performed with spleen cells from diabetic NOD or TCR-Tg donors. Onset surface staining on pancreatic single-cell suspensions after transfer of of diabetes was monitored by urine glucose screening and positive readings BDC-2.5. were confirmed by blood glucose measurement. Blood glucose concentra- by guest on September 25, 2021 tions Ͼ15 mM for more than 1 day were considered diagnostic of overt diabetes. Recovery of cells from pancreas Analysis of NO from macrophages stimulated by T cell clones Approximately 1 week after adoptive transfer with T cell clones or at onset Resting T cells (8 ϫ 104) were cultured for 48 h in the presence of freshly of diabetes (4–6 wk) following diabetic spleen cell transfers, recipient isolated NOD thioglycolate-elicited peritoneal exudate cells (5 ϫ 104)as NOD.scid mice were sacrificed. As per methods described previously (2, APC and 10 ␮g of a membrane preparation from ␤ tumor cells as Ag. As 6), spleens and pancreata were removed and digested in PBS containing an indirect assessment of NO, nitrite ion concentration in the harvested collagenase and GolgiPlug (BD Biosciences) in a 37°C water bath for supernatants was measured and averaged from triplicate wells using the 20–40 min. Single-cell suspensions of the digested pancreata and spleens one-step Griess reagent assay (13) and compared with nitrite standards. were prepared, flash-spun to remove tissue debris, and then centrifuged at The same medium used in the cell culture (CM) was the zero standard to 300 ϫ g for 10 min at 4°C to pellet. Cells were resuspended in CM/ account for nitrite background. GolgiPlug and incubated at 37°C for 3–5 h, and in the case of T cells, with or without PMA (100 ng/ml) and ionomycin (1 ␮g/ml). Results Ex vivo analysis of cytokines, chemokines, and chemokine Macrophages are recruited to the pancreas by pathogenic CD4 receptors T cells After culture ex vivo of cells isolated from the pancreas, cells were har- To investigate the activity of CD4 T cell clones in vivo following vested and washed before resuspension in staining buffer/GolgiPlug. Cells adoptive transfer, pancreatic tissue was removed at various time were stained in a 96-well round-bottom plate, 2 wells for each set of Ab points after transfer (but before onset of diabetes) and single cell combinations, one for specific Ab mix, and one for isotype control mix. suspensions of the infiltrating immune cells were analyzed by flow Abs for surface staining included anti-CD4, anti-CD11b, and anti-CCR5 (BD Biosciences); anti-F4/80 (Caltag); and anti-CXCR3 and anti-CCR8 cytometry. We found that in addition to producing inflammatory (R&D Systems). Surface and intracellular cytokine staining was performed cytokines in the pancreas, the pathogenic T cell clones recruited as described for in vitro intracellular cytokine/chemokine staining, with the substantial numbers of macrophages, detectable at 2, 4, and 6 days modification of adding GolgiPlug to all reagents until fixation and prein- after transfer (6). Similar data are illustrated in Fig. 1 in a exper- ␮ cubating in staining buffer containing 2 g/ml FcBlock (BD Biosciences) iment in which the diabetogenic T cell clone BDC-2.5, or a non- for 5 min on ice before addition of surface staining Abs. Polyclonal intracellular staining Abs used were from R&D Systems and included poly- pathogenic T cell clone BDC-2.4, were transferred to young NOD. clonal goat IgG, anti-CCL1 (TCA-3), anti-CCL3 (MIP-1␣), anti-CCL4 scid recipients and a week later, cells were isolated from the (MIP-1␤), anti-CCL5 (RANTES), anti-CCL6 (C10), anti-CCL9/10 (MIP- pancreas and analyzed ex vivo for the presence of macrophages 1␥), and anti-IL-1␤ as primary Abs, followed by FITC-rabbit anti-goat (Fig. 1A). Comparison of the percentages of CD11bϩ cells in pan- secondary Ab (Vector Laboratories) or Cy-5-conjugated donkey anti-goat secondary Ab (Jackson Immunoresearch). Staining for chemokine recep- creas after transfer of the two clones indicates that whereas sub- tors was conducted without ex vivo culture on unfixed single-cell suspen- stantial numbers of macrophages were observed in pancreas of ϩ sions from pancreas or spleen as indicated. mice receiving the BDC-2.5 T cell clone, the numbers of CD11b 5762 ACTIVATED MACROPHAGES IN T1D

FIGURE 2. Inﬂammatory cytokine production in macrophages recruited to the pancreas by a pathogenic T cell clone. Expanded cultures of the BDC-2.5 T cell clone or the nonpathogenic BDC-2.4 Th1 T cell clone were injected into young NOD.scid recipients. At 4–6 days after transfer, pancreatic single-cell suspensions were incubated in the presence of Brefeldin A and were stained for surface expression of CD11b and for intracellular cytokine protein. TNF-␣ (A) or IL-1␤ (B) production by these macrophages was analyzed by ﬂow cytometry. Histograms gated on CD11bϩ cells are included to indicate the

proportion of macrophages staining positive for the indicated cytokine protein. The data shown are representative of two independent experiments with Downloaded from similar results. macrophages in pancreas of BDC-2.4 recipients were quite low, of inﬂammatory chemokines. We found that whereas CD11bϩ similar to those found in uninjected controls (data not shown). cells isolated from the spleen produced very low levels of chemo-

These results suggest that macrophage recruitment is linked with kines, there were substantial amounts of three chemokines found in http://www.jimmunol.org/ the pathogenic T cell response in autoimmune diabetes. Because macrophages recruited to the pancreas. Fig. 3 shows intracellular CD11b is a marker for both macrophages and dendritic cells, we staining in CD11bϩ cells for CCL3 (MIP-1␣), CCL4 (MIP-1␤), performed two-color staining on cells recruited to the pancreas for and CCL6 (C10), 1 wk after transfer of the diabetogenic CD4 T CD11b and for F4/80. Although expressed at lower levels on ac- cell clone BDC-2.5. As also illustrated in this figure, no staining tivated macrophages, the F4/80 Ag is commonly considered to be was detected in macrophages for CCL1, CCL9/10, or CCL5. This the most macrophage-specific marker (14). We found that almost result demonstrates that the macrophages recruited by pathogenic all of the CD11bϩ cells recruited to the pancreas also stained for T cells are an important source of at least three inducible chemo- the F4/80 Ag (Fig. 1B), and therefore to identify macrophages in kines in the pancreas during disease. subsequent experiments, we used the brighter staining anti-CD11b by guest on September 25, 2021 Ab, which is a particularly valuable reagent in four-color staining NO is another product of macrophages activated by from the pancreas ex vivo. It is possible, however, that the diabetogenic CD4 T cells ϩ ϩ CD11b F4/80 cells may include some dendritic cells. NO can function either as an important mediator of positive cellular outcomes such as proliferation and activation, or as a signal- Macrophages recruited to the pancreas are activated to produce ing mediator leading to cell death (17). The deleterious effects of inflammatory cytokines NO include its functioning as an intermediate in the production of To test the hypothesis that macrophages are activated by diabeto- reactive oxygen and reactive nitrogen species (ROS and RNS). genic T cells upon recruitment to the pancreas, we analyzed the The exact role of NO in T1D appears to be complicated (17), but CD11b-staining population recovered from the pancreas by intra- its presence in combination with inflammatory cytokines, particu- cellular staining. Our results showed that after transfer of the di- larly IL-1␤, is a strong predictor of islet-cell cytotoxicity (18, 19). abetogenic T cell clone BDC-2.5, large numbers of the infiltrating Because macrophages are an important source of NO, we mea- macrophages were producing TNF-␣ (Fig. 2A). In contrast, adop- sured the ability of a panel of pathogenic CD4 T cell clones to tive transfer of the nonpathogenic clone BDC-2.4 resulted in few stimulate NO production by macrophages, using the concentration macrophages recruited and little to no stimulation of TNF-␣ pro- of nitrite ions as an indirect measure of NO. By stimulating T cell duction in the CD11bϩ cells that were present. A second cytokine clones with Ag, we are able to observe the effects on macrophages of interest in activated macrophages is IL-1␤, a macrophage-spe- cocultured with the T cells and thus exposed to T cell-derived cific cytokine that has been shown to be a very potent mediator of inflammatory cytokines. The data shown in Fig. 4 show that the islet cell cytotoxicity (15, 16). As with TNF-␣, we found that ability to stimulate macrophage production of this important in- activation of macrophages to produce IL-1␤ was associated with flammatory mediator is a property of diabetogenic CD4 Th1 T disease induced by pathogenic CD4 T cell clones, and that there cells. In contrast, a Th2 T cell clone, 2.5Tg/T2-X, used as a con- was a clear difference in macrophage IL-␤ levels in the pancreas trol, was unable to stimulate NO production by macrophages, cor- after transfers with BDC-2.5 vs the nonpathogenic clone, BDC-2.4 relating with its inability to transfer disease to the NOD.scid (Fig. 2B). These results suggest that production of inflammatory mouse. cytokines by macrophages is linked to the pathogenicity of CD4 T cells. Recruitment and activation of macrophages is a manifestation of T cell effector function in other diabetes induction and Activated macrophages in the pancreas also produce spontaneous disease models chemokines To determine whether macrophage recruitment and activation was We also wanted to know whether macrophages recruited to the a common feature of T1D pathogenesis, we analyzed macrophages pancreas after adoptive transfer of T cells were a significant source isolated from the pancreas in several adoptive transfer and two The Journal of Immunology 5763

FIGURE 4. NO production by macrophages stimulated by diabetogenic T cell clones. Resting T cells (8 ϫ 104) were cultured for 48 h in the presence of freshly isolated NOD thioglycolate-elicited peritoneal exudate cells (5 ϫ 104) as APC and 10 ␮g of a membrane preparation from ␤ tumor cells as Ag. As a measure of macrophage activation, nitrite ion concentration in the harvested supernatants was measured using the one-step Griess reagent assay, and compared with nitrite standards from Downloaded from duplicate wells. Results shown are obtained from cultures of six diabetogenic Th1 T cell clones and compared with data from a Th2 T cell clone (2.5 Tg/T2-X) used as a control. These data are representative of two identical experiments. http://www.jimmunol.org/

spontaneous models of T cell-mediated autoimmune diabetes (Fig. 5). Fig. 5A shows inflammatory cytokine production by macrophages in two commonly used adoptive transfer systems involving spleen cells or in transfers with defined T cell clones. Spleen cell transfers included either heterogeneous populations of spleen cells from diabetic NOD mice or a quasi-clonal population of spleen cells obtained from the widely used BDC-2.5 TCR-Tg mouse. These systems were compared with transfers with two T cell by guest on September 25, 2021 clones, either a clone from our panel of diabetogenic CD4 T cell clones, or the pathogenic insulin-reactive CD8 T cell clone G9 described by Wong et al. (20). These different transfer models exhibit varying kinetics of disease induction in young NOD.scid recipients, but all depend on a population of pathogenic T cells to transfer disease. As a control, a CD4 Th2 T cell clone, 2.5Tg/ T2-X, was used. The 2.5Tg/T2-X clone does not induce diabetes in NOD.scid recipients (21), nor does it lead to infiltration and cytokine production by macrophages (Fig. 5B). In Fig. 5C are results from analysis of macrophage inflammatory cytokine production in two spontaneous models of disease, either a spontaneously diabetic wild-type NOD mouse or a prediabetic 6.9 TCR-Tg/NOD.scid mouse (a TCR-Tg produced from a second diabetogenic CD4 T cell clone in our panel, BDC-6.9; see Ref. 8). Although mouse models for T1D differ in some respects and mechanisms of pathogenesis vary with the induction method, these results suggest that a common feature of disease mediated by pathogenic T cells, in both spontaneous and transfer models of diabetes, is a large num- ber of infiltrating macrophages, many of which are producing in- FIGURE 3. Chemokine production in macrophages recruited to the flammatory cytokines. The NOD spontaneous mouse model ex- pancreas by a pathogenic T cell clone. Six days following transfer of the hibits a longer time frame for disease onset, which likely explains BDC-2.5 clone into young NOD.scid recipients, single-cell suspensions the smaller numbers of macrophages in the pancreas at any of pancreas or spleen (pooled from 3 recipient mice) were incubated in given time point, even though a large percentage of them are the presence of Brefeldin A, stained for surface expression of CD11b producing inflammatory cytokine. Our findings lend support to and intracellular chemokine proteins, and analyzed by flow cytometry. the theory that the macrophage is a major effector cell for ␤ cell Percentages in the upper right quadrants indicate the proportion of macrophages staining positive for the indicated chemokine. Chemokine destruction through its cytotoxic activity (22). The fact that this staining in pancreatic macrophages in the left column is in contrast macrophage activity is dependent on pathogenic T cells is sug- to the relative lack of chemokine expression in control splenic gested by the lack of activated macrophages recruited to the macrophages. pancreas by either a Th2 control clone (Fig. 5B) or a nonpathogenic CD4 Th1 clone (Fig. 2). 5764 ACTIVATED MACROPHAGES IN T1D

FIGURE 5. Recruitment and activation of macrophages seen in a variety of T1D mouse models. Single-cell suspensions from pancreas of recipients of adoptive transfers of pathogenic cells (A), recipients of a Th2 control clone (B), and spontaneously diabetic mice (C) were incubated in the presence of Brefeldin A and analyzed by ﬂow cytometry for TNF-␣ or IL-1␤ production by CD11bϩ macrophages. Analyses were performed 6–7 days after transfer of T cell clones or 3–4 wk after transfer of TCR transgenic or diabetic NOD spleen cells, as well as on pancreata from spontaneously diabetic adult NOD mice and prediabetic TCR transgenic mice (time points were chosen to be as close to disease onset as possible). Percentages in the upper right quad- Downloaded from rants indicate the proportion of macrophages staining positive for TNF-␣ or IL-1␤. Each dot plot represents the staining from an individual mouse and represents similar data from 2 to 3 mice for each transfer experiment. http://www.jimmunol.org/

Macrophages recruited to the pancreas express chemokine Diabetogenic T cells secrete numerous macrophage receptors chemoattractants in vitro but only CCL1 in the pancreas As illustrated in Fig. 1, our results have demonstrated that adoptive Chemokines are potent chemoattractants and possible activators of transfer of diabetogenic CD4 T cell clones leads to the accumu- macrophages in inflammatory sites, and it has been reported that lation of large numbers of macrophages in the pancreas. For mac- mRNA for several inflammatory chemokines can be detected in rophages to be recruited to an inflammatory site in which T cells whole pancreas after disease transfer (26). We reported previously by guest on September 25, 2021 are producing cytokines and chemokines, they must express che- that a variety of chemokines can be detected both at the mRNA mokine receptors. To investigate further the mechanisms by which and at the protein level in a panel of seven CD4 diabetogenic T cell macrophages are recruited to the pancreas during T1D, we exam- clones (6). Table I is a list of all the chemokines tested in three of ined the expression of three chemokine receptors that have been implicated in autoimmune diseases. CCR5 has been shown previ- Table I. Chemokines produced by pathogenic CD4 Th1 T cell clones ously to be necessary for development of T1D (23) is a receptor for MIP-1␣, MIP-1␤, MIP-1␥, and RANTES, all of which are che- BDC-2.5 BDC-6.3 BDC-6.9 mokines produced at the protein level by the pathogenic CD4 T cell clones in response to TCR stimulation in vitro (Table I). Fig. CCL1 (TCA-3) ϩϩϩ ϪϪϪ 6 shows that this receptor is expressed by virtually all CD11bϩ CCL2 (MCP-1) CCL3 (MIP-1␣) ϩϩϩ macrophages recovered from the pancreas after transfer of BDC- CCL4 (MIP-1␤) ϩϩϩ 2.5, whereas splenic macrophages were negative for CCR5. We CCL5 (RANTES) ϩϩϩ also tested for expression of CXCR3 because this chemokine re- CCL6 (C10) ϩϩϩ ϪϪϪ ceptor was previously found to be required for infiltration into CCL7 (MCP-3) CCL8 (MCP-2) ϪϪϪ islets during progression to diabetes (24); however, the ligands for CCL9/10 (MIP-1␥) ϩϩϩ CXCR3 (CXCL9–11) were not expressed by the pathogenic T cell CCL11 (Eotaxin) ϪϪϪ clones (J. Cantor and K. Haskins, unpublished data). As shown in CCL12 (MCP-5) ϪϪϪ CCL17 (TARC) ϪϪϪ Fig. 6, CXCR3 is expressed at high levels on the surface of a ␣ ϪϪϪ ϩ CCL20 (MIP-3 ) substantial portion of CD11b macrophages recruited to the pan- CCL21 (SLC) ϪϪϪ creas by BDC-2.5 and is also expressed at much lower levels on a CCL22 (MDC) ϪϪϪ portion of splenic macrophages. Possibly the most interesting find- CCL24 (Eotaxin-2) ϪϪϪ CCL25 (TECK) ϪϪϪ ing from these experiments is that macrophages found in the CXCL2 (MIP-2) ϪϪϪ pancreas after adoptive transfer of pathogenic T cells also ex- CXCL5 (LIX) ϪϪϪ press CCR8 and at levels that are much higher (Ͼ5-fold) than CXCL9 (MIG) ϪϪϪ ϪϪϪ levels of CCR8 expression on splenic macrophages (Fig. 6, bot- CXCL10 (IP-10) CXCL11 (IP-9) ϪϪϪ tom panels). Expression of CCR8 by recruited macrophages is CXCL12 (SDF-1) ϪϪϪ of particular note because of the link between CCR8 and other CXCL13 (BLC) ϪϪϪ autoimmune disease states (25, 32), and because it has not pre- CX3CL1 (Fractalkine) ϪϪϪ ϩϩϪϩ viously been studied in T1D. XCL1 (Lymphotactin) / The Journal of Immunology 5765 Downloaded from

FIGURE 6. Chemokine receptors on CD11bϩ macrophages recruited to pancreas after transfer of a pathogenic CD4 T cell clone. Six days following transfer of the pathogenic CD4 Th1 T cell clone BDC-2.5 into young NOD.scid recipients, single-cell suspensions of pancreas or spleen were stained for surface expression of CD11b vs CCR5, CXCR3, or CCR8 and

analyzed by flow cytometry. Filled peaks represent specific staining com- http://www.jimmunol.org/ pared with isotype control Ab (unfilled peaks), gated on CD11bϩ cells. Each histogram represents data from an individual mouse. the CD4 Th1 T cell clones; those that tested positive at the mRNA level were confirmed by intracellular staining. These results indicate that at least seven chemokines are made by diabetogenic T cells, all of which are documented chemoattractants for macrophages (27–31). In vitro, these chemokines are produced only by guest on September 25, 2021 upon TCR-mediated activation, with the exception of RANTES, a potent T cell chemokine that is present constitutively in unstimu- lated T cell clones. Because macrophages and T cells migrating to the pancreas produce multiple inflammatory mediators, and as several chemoat- FIGURE 7. CCL1 (TCA-3) production ex vivo by the pathogenic clone BDC-2.5 from the pancreas after transfer. Six days after transfer of the T tractant receptors capable of binding multiple ligands are found on cell clone BDC-2.5 into 9-day-old NOD.scid mice, single-cell suspensions macrophages, it is not immediately apparent whether there is any of pancreas (pooled from 3 mice) were incubated in the presence of Brefel- one receptor-ligand pairing that is of particular significance in the din A, with or without PMA, and analyzed for intracellular staining by flow developing disease process. Our results indicate that macrophages cytometry. Percentages in the upper right quadrants indicate the proportion recruited to the pancreas express at least three chemokine receptors of T cells staining positive for the indicated chemokine. A similar pattern thought to be relevant to disease (Fig. 6), the most interesting of of chemokine expression was seen upon repetition of this experiment. which may be CCR8 because this chemokine receptor, unlike others, is known to bind only one ligand, the chemokine CCL1 (or TCA-3) (25). To further investigate the possible role of CCR8 on macrophages in T1D, we investigated whether diabetogenic CD4 Discussion T cell clones produce the ligand for this chemokine receptor in Our studies on the activity of macrophages in the inflammatory site vivo. We adoptively transferred the T cell clone BDC-2.5 to NOD. have indicated that recruitment and activation of macrophages is a scid mice and after removal spleen and pancreas 6 days later, we key component of pathogenic CD4 T cell effector function. First, analyzed the T cells by intracellular staining for production of comparison of the numbers of infiltrating macrophages after adop- chemokines. Although in vitro the T cells can be induced to tive transfer of a nonpathogenic vs a pathogenic CD4 T cell clone make several inflammatory chemokines, including CCL3 (MIP- indicates that macrophage recruitment is linked to pathogenicity. 1␣), CCL4 (MIP-1␤), CCL9/10 (MIP-1␥), CCL6 (C10), CCL5 Second, inflammatory cytokine production by macrophages, which (RANTES), and CCL1 (TCA-3) (6), the only chemokine that is stimulated by diabetogenic CD4 T cells, was investigated ex could be detected in CD4 T cells in the pancreas ex vivo was vivo from the pancreas. Our data demonstrate that TNF-␣ and CCL1 (Fig. 7). The few T cells that were found in the spleen IL-1␤ are made by macrophages in the pancreas after transfer of were negative for all chemokines tested (data not shown). The the T cell clone, BDC-2.5, but not with a nonpathogenic T cell production of CCL1 by pathogenic T cells in the pancreas, to- clone. These findings were extended and confirmed by our data on gether with the expression of CCR8 on macrophages recruited recruitment and activation of macrophages recruited to the pan- to the pancreas, provide the first evidence that CCR8/CCL1 creas in other T cell-mediated models of T1D (Fig. 5). A third interaction may play a role in T1D. important finding was that macrophages recruited and activated by 5766 ACTIVATED MACROPHAGES IN T1D diabetogenic CD4 T cells in the pancreas are a key source of inflammatory chemokines. The chemokines produced in the pancreas during disease progression are of the inducible category of chemoattractants, present in peripheral tissues during inflammation. As there has been little investigation of macrophage chemokines in autoimmunity, this observation now serves as a basis for future study into the role of macrophage-derived chemoattractants. For example, it may be that production of chemokines such as CCL3, CCL4, and CCL6 by macrophages in the disease site provides another mechanism by which inflammation is augmented, serving to further recruit and/or boost the activation of other inflammatory cells. Another intriguing question with regard to chemokines made by macrophages is whether different chemokines, or levels of cytokines/chemokines, represent heterogeneity among macrophages recruited to the pancreas, such as is suggested by the dot plots in Fig. 3 indicating that CD11bhigh vs CD11bdim cells may be secreting different chemokines. A fourth aspect of macro- FIGURE 8. Macrophage production of inflammatory mediators is a phage function in the pancreas is the NO production that is stim- manifestation of CD4 T cell effector function. In this hypothesized sce- ulated by pathogenic CD4 Th1 T cell clones. NO is produced by nario, a pathogenic CD4 T cell is activated upon recognition of autoantigen Downloaded from islet ␤ cells and is a mediator that combines with IL-1␤ and TNF-␣ to produce inflammatory cytokines and chemokines. Production of inflam- to cause ␤ cell death, at least in vitro (15, 19). Our work points to matory mediators by the T cells leads in turn to recruitment of macro- macrophages as another important source of NO in the pancreas phages and their activation to up-regulate chemokine receptors and to pro- during progression to disease. duce inflammatory cytokines, chemokines, and other cytotoxic mediators The fact that the infiltrating macrophages express CCR5, or signaling intermediates, such as NO.

CXCR3, and CCR8, three chemokine receptors implicated in in- http://www.jimmunol.org/ flammation, suggests that chemokines secreted by the CD4 T cells could be one mechanism whereby pathogenic CD4 T cells recruit, and perhaps help to activate, macrophages. Our data regarding and CCR8 in inflamed CNS tissue (7, 34). T cell chemokines such CCL1 (TCA-3) production by CD4 T cells and expression of as CCL1 attract macrophages to the pancreas through CCR8 and CCR8 on macrophages in the pancreas may be especially signif- lead to up-regulation of additional chemokine receptors (CCR5, icant in this regard. The role of this ligand/receptor pair in the CXCR3) on macrophages. Macrophages are in turn activated by T ␣ ␥ chemoattractant system is still being delineated and CCR8 has cell-derived cytokines such as TNF- and IFN- to become cyto- been studied more commonly in its role as a chemoattractant re- toxic effector cells that produce NO as well as inflammatory cy- ceptor on neutrophils, eosinophils, and Th2 T cells (27, 32, 33). tokines and chemokines, thereby contributing to the overall in- by guest on September 25, 2021 More recently, however, it has been reported that phagocytic mac- flammatory milieu and recruitment of other immune cells. rophages and activated microglial cells express CCR8 in CNS infiltrates in experimental allergic encephalitis, and that these Acknowledgments CCR8ϩ cells respond to CCL1 (7, 34). Our results raise the pos- We thank Brenda Bradley and Gene Barbour for technical assistance, sibility that production of CCL1 by pathogenic Th1 T cells in the Dr. Dirk Homann for Ab reagents and discussions, and Dr. Sue Wong for providing us with the NOD CD8 T cell clone, G9. We also thank pancreas is a means of recruiting CCR8-positive cytotoxic mac- Dr. Albert Zlotnik for critically reading and commenting on the rophages. Although substantially higher on macrophages in the manuscript. pancreas after adoptive transfer of pathogenic T cells, CCR8 is also expressed on macrophages in the spleen and it may be that Disclosures chemoattraction of these cells (or dendritic cells) to the pancreas The authors have no financial conflict of interest. depends initially on CCR8. The levels of CCR5, and possibly CXCR3, are increased on macrophages after traffic to the target References organ (35, 36) and may serve to “anchor” macrophages and other 1. Haskins, K., and M. McDuffie. 1990. Acceleration of diabetes in young NOD responding cell types in the pancreas. CCR5 and CXCR3 are both mice with a CD4ϩ islet-specific T cell clone. Science 249: 1433–1436. up-regulated by IFN-␥ and TNF-␣ in tissue-specific inflammation 2. Peterson, J. D., R. Berg, J. D. Piganelli, M. Poulin, and K. Haskins. 1998. Anal- ysis of leukocytes recruited to the pancreas by diabetogenic T cell clones. Cell. (35). The ligands for CCR5 (CCL3, CCL4, CCL5) are very com- Immunol. 189: 92–98. mon in inflammatory sites and although ligands for CXCR3 have 3. Rosmalen, J. G., T. Martin, C. Dobbs, J. S. Voerman, H. A. Drexhage, been shown by others to be expressed by inflamed islets (24), they K. Haskins, and P. J. Leenen. 2000. Subsets of macrophages and dendritic cells in nonobese diabetic mouse pancreatic inflammatory infiltrates: correlation with were not produced by the diabetogenic T cell clones. The non- the development of diabetes. Lab. Invest. 80: 23–30. specificity of CCR5 interactions, as well as the apparent lack of 4. Oschilewski, U., U. Kiesel, and H. Kolb. 1985. Administration of silica prevents diabetes in BB-rats. Diabetes 34: 197–199. CD4 T cell chemokines for CXCR3, underscore the potential im- 5. Yoon, J. W., H. S. Jun, and P. Santamaria. 1998. Cellular and molecular mech- portance of our data implicating the highly specific interaction that anisms for the initiation and progression of ␤ cell destruction resulting from the occurs between the T cell chemokine CCL1 and the macrophage collaboration between macrophages and T cells. Autoimmunity 27: 109–122. 6. Cantor, J., and K. Haskins. 2005. Effector function of diabetogenic CD4 Th1 T receptor CCR8. cell clones: a central role for TNF-␣. J. Immunol. 175: 7738–7745. In summary, these various findings would suggest that the re- 7. Trebst, C., S. M. Staugaitis, P. Kivisakk, D. Mahad, M. K. Cathcart, B. Tucky, cruitment of macrophages is directed first through the action of T. Wei, M. R. Rani, R. Horuk, K. D. Aldape, et al. 2003. CC chemokine receptor 8 in the central nervous system is associated with phagocytic macrophages. cytokines and chemokines produced by pathogenic T cells, as de- Am. J. Pathol. 162: 427–438. picted in Fig. 8. In particular, the inflammatory cytokine TNF-␣ is 8. Pauza, M. E., C. M. Dobbs, J. He, T. Patterson, S. Wagner, B. S. Anobile, B. J. Bradley, D. Lo, and K. Haskins. 2004. T-cell receptor transgenic response a central player in the activity of diabetogenic CD4 T cell clones to an endogenous polymorphic autoantigen determines susceptibility to diabetes. (6), and was found to be responsible for up-regulation of CCL1 Diabetes 53: 978–988. The Journal of Immunology 5767

9. Haskins, K., M. Portas, B. Bradley, D. Wegmann, and K. Lafferty. 1988. T- 24. Frigerio, S., T. Junt, B. Lu, C. Gerard, U. Zumsteg, G. A. Hollander, and L. Piali. lymphocyte clone specific for pancreatic islet antigen. Diabetes 37: 1444–1448. 2002. ␤ cells are responsible for CXCR3-mediated T-cell infiltration in insulitis. 10. Haskins, K., M. Portas, B. Bergman, K. Lafferty, and B. Bradley. 1989. Pancre- Nat. Med. 8: 1414–1420. atic islet-specific T-cell clones from nonobese diabetic mice. Proc. Natl. Acad. 25. Goya, I., J. Gutierrez, R. Varona, L. Kremer, A. Zaballos, and G. Marquez. 1998. Sci. USA 86: 8000–8004. Identification of CCR8 as the specific receptor for the human ␤-chemokine I-309: 11. Bergman, B., and K. Haskins. 1994. Islet-specific T-cell clones from the NOD cloning and molecular characterization of murine CCR8 as the receptor for mouse respond to ␤-granule antigen. Diabetes 43: 197–203. TCA-3. J. Immunol. 160: 1975–1981. 12. Dorner, B. G., S. Steinbach, M. B. Huser, R. A. Kroczek, and A. Scheffold. 2003. 26. Bradley, L. M., V. C. Asensio, L. K. Schioetz, J. Harbertson, T. Krahl, Single-cell analysis of the murine chemokines MIP-1␣, MIP-1␤, RANTES and G. Patstone, N. Woolf, I. L. Campbell, and N. Sarvetnick. 1999. Islet-specific ATAC/lymphotactin by flow cytometry. J. Immunol. Methods 274: 83–91. Th1, but not Th2, cells secrete multiple chemokines and promote rapid induction of autoimmune diabetes. J. Immunol. 162: 2511–2520. 13. Grisham, M. B., G. G. Johnson, and J. R. Lancaster, Jr. 1996. Quantitation of 27. Devi, S., J. Laning, Y. Luo, and M. E. Dorf. 1995. Biologic activities of the Methods Enzymol. nitrate and nitrite in extracellular fluids. 268: 237–246. ␤-chemokine TCA3 on neutrophils and macrophages. J. Immunol. 154: 14. Lin, H. H., D. E. Faunce, M. Stacey, A. Terajewicz, T. Nakamura, 5376–5383. J. Zhang-Hoover, M. Kerley, M. L. Mucenski, S. Gordon, and J. Stein-Streilein. 28. Asensio, V. C., S. Lassmann, A. Pagenstecher, S. C. Steffensen, S. J. Henriksen, 2005. The macrophage F4/80 receptor is required for the induction of antigen- and I. L. Campbell. 1999. C10 is a novel chemokine expressed in experimental specific efferent regulatory T cells in peripheral tolerance. J. Exp. Med. 201: inflammatory demyelinating disorders that promotes recruitment of macrophages 1615–1625. to the central nervous system. Am. J. Pathol. 154: 1181–1191. 15. Corbett, J. A., J. L. Wang, M. A. Sweetland, J. R. Lancaster, Jr., and 29. Dorner, B. G., A. Scheffold, M. S. Rolph, M. B. Huser, S. H. Kaufmann, M. L. McDaniel. 1992. Interleukin 1 ␤ induces the formation of nitric oxide by A. Radbruch, I. E. Flesch, and R. A. Kroczek. 2002. MIP-1␣, MIP-1␤, RANTES, ␤-cells purified from rodent islets of Langerhans: evidence for the ␤-cell as a and ATAC/lymphotactin function together with IFN-␥ as type 1 cytokines. Proc. source and site of action of nitric oxide. J. Clin. Invest. 90: 2384–2391. Natl. Acad. Sci. USA 99: 6181–6186. 16. Mandrup-Poulsen, T. 1996. The role of interleukin-1 in the pathogenesis of 30. Menten, P., A. Saccani, C. Dillen, A. Wuyts, S. Struyf, P. Proost, A. Mantovani, IDDM. Diabetologia 39: 1005–1029. J. M. Wang, and J. Van Damme. 2002. Role of the autocrine chemokines MIP-1␣ 17. Mandrup-Poulsen, T. 2003. ␤ cell death and protection. Ann. NY Acad. Sci. 1005: and MIP-1␤ in the metastatic behavior of murine T cell lymphoma. J. Leukocyte 32–42. Biol. 72: 780–789. Downloaded from 18. McDaniel, M. L., G. Kwon, J. R. Hill, C. A. Marshall, and J. A. Corbett. 1996. 31. Maurer, M., and E. von Stebut. 2004. Macrophage inflammatory protein-1. Int. Cytokines and nitric oxide in islet inflammation and diabetes. Proc. Soc. Exp. J. Biochem. Cell. Biol. 36: 1882–1886. Biol. Med. 211: 24–32. 32. Keller, M., Z. Spanou, P. Schaerli, M. Britschgi, N. Yawalkar, M. Seitz, 19. Thomas, H. E., R. Darwiche, J. A. Corbett, and T. W. Kay. 2002. Interleukin-1 P. M. Villiger, and W. J. Pichler. 2005. T cell-regulated neutrophilic inflamma- plus ␥-interferon-induced pancreatic ␤-cell dysfunction is mediated by ␤-cell tion in autoinflammatory diseases. J. Immunol. 175: 7678–7686. nitric oxide production. Diabetes 51: 311–316. 33. Chensue, S. W., N. W. Lukacs, T. Y. Yang, X. Shang, K. A. Frait, S. L. Kunkel, T. Kung, M. T. Wiekowski, J. A. Hedrick, D. N. Cook, et al. 2001. Aberrant in 20. Wong, F. S., I. Visintin, L. Wen, R. A. Flavell, and C. A. Janeway, Jr. 1996. CD8

vivo T helper type 2 cell response and impaired eosinophil recruitment in CC http://www.jimmunol.org/ T cell clones from young nonobese diabetic (NOD) islets can transfer rapid onset chemokine receptor 8 knockout mice. J. Exp. Med. 193: 573–584. J. Exp. Med. of diabetes in NOD mice in the absence of CD4 cells. 183: 67–76. 34. Murphy, C. A., R. M. Hoek, M. T. Wiekowski, S. A. Lira, and J. D. Sedgwick. 21. Poulin, M., and K. Haskins. 2000. Induction of diabetes in nonobese diabetic 2002. Interactions between hemopoietically derived TNF and central nervous mice by Th2 T cell clones from a TCR transgenic mouse. J. Immunol. 164: system-resident glial chemokines underlie initiation of autoimmune inﬂammation 3072–3078. in the brain. J. Immunol. 169: 7054–7062. 22. Suk, K., S. Kim, Y. H. Kim, K. A. Kim, I. Chang, H. Yagita, M. Shong, and 35. Croitoru-Lamoury, J., G. J. Guillemin, F. D. Boussin, B. Mognetti, L. I. Gigout, M. S. Lee. 2001. IFN-␥TNF-␣ synergism as the ﬁnal effector in autoimmune A. Cheret, B. Vaslin, R. Le Grand, B. J. Brew, and D. Dormont. 2003. Expression diabetes: a key role for STAT1/IFN regulatory factor-1 pathway in pancreatic ␤ of chemokines and their receptors in human and simian astrocytes: evidence for cell death. J. Immunol. 166: 4481–4489. a central role of TNF ␣ and IFN ␥ in CXCR4 and CCR5 modulation. Glia 41: 23. Carvalho-Pinto, C., M. I. Garcia, L. Gomez, A. Ballesteros, A. Zaballos, 354–370. J. M. Flores, M. Mellado, J. M. Rodriguez-Frade, D. Balomenos, and 36. Chen, J., B. P. Vistica, H. Takase, D. I. Ham, R. N. Fariss, E. F. Wawrousek,

A. C. Martinez. 2004. Leukocyte attraction through the CCR5 receptor controls C. C. Chan, J. A. DeMartino, J. M. Farber, and I. Gery. 2004. A unique pattern by guest on September 25, 2021 progress from insulitis to diabetes in non-obese diabetic mice. Eur. J. Immunol. of up- and down-regulation of chemokine receptor CXCR3 on inﬂammation- 34: 548–557. inducing Th1 cells. Eur. J. Immunol. 34: 2885–2894.