and (2009) 10, 687–701 & 2009 Macmillan Publishers Limited All rights reserved 1466-4879/09 $32.00 www.nature.com/gene

ORIGINAL ARTICLE and expression analysis in target organs of acute graft-versus-host disease

A Bouazzaoui1, E Spacenko1, G Mueller1, S Miklos1, E Huber2, E Holler1, R Andreesen1 and GC Hildebrandt1 1Department of Hematology and Oncology, University of Regensburg Medical School, Regensburg, Germany and 2Department of Pathology, University of Regensburg Medical School, Regensburg, Germany

Acute graft-versus-host disease (aGVHD) is a major complication after allogeneic transplantation (allo-BMT), and infiltration of donor leukocytes into aGVHD target organs is partially orchestrated by . Using a murine BMT model, the expression of 30 chemokines or chemokine receptors in the , liver, gut and tongue was analyzed using real-time PCR at 1, 2, 3 and 6 weeks after BMT during the development of clinical aGVHD and target organ histopathology. CXCL9–11 expression was linked to elevated expression of CXCR3 in the gut, lung and tongue. In contrast, hepatic CXCR3 expression was not changed, whereas a clear association was seen for CXCL16 and CXCR6 expression. An elevated intestinal CCL3 expression 1 week after allo-BMT was associated with an increased expression of CCR5 but not CCR1 or CCR3, and in the lung and liver CCL3–CCL5 expression was associated with increases in CCR1 and CCR5. Overexpression of CCL2, CCL8, CCL12 and their receptor CCR2 was found in the liver and lung, but not in the gut and tongue. On the basis of the differences in kinetics and organ distribution, more studies are required to better characterize specific targets within this network, as this will allow the development of novel preventive and therapeutic approaches by using single or multiple targeting reagents. Genes and Immunity (2009) 10, 687–701; doi:10.1038/.2009.49; published online 2 July 2009

Keywords: stem cell transplantation; graft-versus-host disease; chemokine; chemokine receptor; leukocyte recruitment

Introduction seven-transmembrane domain-containing G - coupled receptors.5,6 Chemokine expression can be Allogeneic bone marrow transplantation (allo-BMT) enhanced by inflammatory ,7 and chemokines often presents the only curative treatment for a number have an important role in recruiting cells of the innate of malignant and non-malignant diseases.1 However, its and adaptive to the sites of inflamma- utility is limited by the development of severe treatment- tion.8–10 Furthermore, they can have additional functions related complications, most importantly, the occurrence in the context of leukocyte activation,11 ,12,13 of acute graft-versus-host disease (aGVHD). The patho- hematopoiesis14 and organization or function of second- physiology of aGVHD is complex and can be divided ary lymphoid tissues.15–17 The chemokine-chemokine into three phases as follows: (1) conditioning toxicity to receptor system shows high complexity and redundancy host tissue, leading to the secretion of inflammatory with more than one chemokine binding to one receptor cytokines and enhanced allo-stimulatory capacity of host and multiple receptors interacting with one chemokine. antigen-presenting cells (APCs); (2) induction of donor Increased expression of a number of chemokines and T-cell activation and T-cell expansion; and (3) cellular chemokine receptors in GVHD target organs has been effector phase, during which activated T cells are described earlier,18–21 and previous studies have further recruited into target organs and cause severe tissue shown that targeting one specific chemokine or chemo- injury.2 kine receptor can result in the reduction of aGVHD- Chemokines comprise a family of predomi- related injury to various target organs.11,22–32 Interest- nantly between 8 to 14 kDa, which at this point can be ingly, tropism and efficacy of such a strategy differed divided into four different subgroups: CC group (CCL1– depending on the chemokine and/or the chemokine 28), CXC group (CXCL1–16), C group (XCL1–2) and receptor involved,33 suggesting that the enhanced CX3C group (CX3CL1).3,4 They exert their effects through expression of chemokines after allo-BMT may differ the interaction with one or more members of a family of between organs with respect to quality, quantity and kinetics. In this study, we analyzed chemokine and chemokine Correspondence: Dr GC Hildebrandt, Department of Hematology receptor expression patterns in GVHD target organs and Oncology, University of Regensburg Medical Center, Franz- using a well-established murine BMT model, in which Josef-Strau Allee 11, Regensburg D-93053, Germany. E-mail: [email protected] alloreactivity occurs across MHC (major histocompat- Received 26 February 2009; revised and accepted 28 May 2009; ibility complex) class I, MHC class II and minor published online 2 July 2009 histocompatibility antigen mismatches. Degree and Chemokines and their receptors in acute GVHD A Bouazzaoui et al 688 kinetics of the expression of 16 chemokines and 14 which are similar to rete ridges in the human epidermis chemokine receptors were analyzed in the lung, liver, but are absent in the murine epidermis and may present gastrointestinal (GI) tract and tongue and were corre- possible primary sites of injury.44–46 lated with systemic and histopathological CXCL9-11 and CXCR3 expression was increased early organ changes. in all four target organs after allo-BMT (Figure 2) and correlated with systemic IFNg and TNFa levels (Supple- mentary Table S1). CXCL9 expression was highest in the Results tongue, followed by the lung, GI tract and liver, and as shown for the liver and gut, mRNA expression levels Development of aGVHD after allo-BMT reflected increased protein expression levels at indivi- To assess mRNA expression profiles of chemokines and dual time points (Figure 7). Except for the lung, in which chemokine receptors in aGVHD target organs, a well- maximum expression occurred at week 2, CXCL9 established murine BMT model was used for GVHD expression peaked in all other target organs at 1 week bxd induction. Lethally irradiated B6D2F1 (H-2 ) mice after allo-BMT and then declined (Figure 2b). CXCL10 6 received 5 Â 10 bone marrow cells supplemented with expression was also at its maximum early after allo-BMT 6 10 Â 10 splenocytes from either syngeneic (B6D2F1) or in the tongue, followed by the lung and liver, and then b allogeneic (C57BL/6; H-2 ) donors. Syngeneic recipients in each organ the expression decreased over time, showed minor changes in weight, fur texture and remaining significantly higher in allogeneic recipients mobility because of radiation toxicity at week 1, which than in syngeneic controls (Figure 2b). In the gut, no was followed by complete clinical recovery after BMT significant changes were seen. A similar time course of (Figures 1a and b). In contrast, allogeneic recipients expression was seen for CXCL11 in the lung, tongue and developed severe clinical aGVHD within the first week liver, and similar to CXCL9, CXCL11 expression was after BMT in parallel to systemic inflammation (Figures significantly upregulated in the gut at any time point 34,35 1c and d) and early T-cell expansion. During this after allo-BMT. period, prominent histopathological changes were al- CXCL9–11 mediate the recruitment of CXCR3 þ ready seen in the GI tract, whereas injury to other leukocytes to inflammatory sites.47–49 Interestingly, ex- aGVHD target organs such as the liver, lung and skin cept for the liver, CXCR3 expression was significantly was only of minor severity and extent (Figure 1e), increased in all other target organs at any time. It was reflecting the combined damaging effects of conditioning highest at week 1 and then decreased by week 2, toxicity and inflammatory cytokines, such as tumor followed by a second increase in the GI tract and tongue, necrosis factor-a (TNFa) and -g (IFNg), on the but not in the lung until week 6. Supported by the GI tract. Subsequently, subsiding levels of serum TNFa correlation analysis (Supplementary Table S1), these and IFNg indicated a decrease in systemic inflammation findings suggest that early IFNg expression and TNFa in allogeneic recipients, but clinical recovery was only are two major factors for inducing these chemokines at partial and temporary, and was observed before pro- this time point, leading to the initial recruitment of gressive clinical symptoms along with increasing target CXCR3 þ T cells as shown earlier for the lung.24 In 2 organ injury to the gut, liver, lung and skin developed. addition, the CXCL9–11:CXCR3 interactions may have As shown for the lung, liver and gut, the immunohisto- only a minor role in the recruitment of CXCR3 þ T cells chemical detection of CD3 þ T cells in aGVHD target to the liver, and at later time points, the recruitment of organs showed progressive T-cell infiltration from week CXCR3-expressing T cells to other target organs may be 1 to week 3 in aGVHD target organs of animals receiving mediated not only by CXCL9–11 but also independently allo-BMT, confirming the migratory effects of donor T from these ligands through other axes in the chemokine– cells to target organs during the cellular effector phase of chemokine network. aGVHD (Figure 1f). Expression of CCL3–5 chemokines and CCR1, CCR3 Expression of the IFNg-inducible chemokines, CXCL9–11, and CCR5 and CXCR3 CCL3–5 chemokines are expressed by various cell types Interferon-g levels have been correlated with aGVHD including T and natural killer (NK) cells, endothelial and severity, but whether this correlation implies a deleter- epithelial cells, fibroblasts, neutrophils and APCs. They ious role of IFNg in GVHD development, reflects only can be induced in response to inflammatory stimuli such the presence of large numbers of activated T cells or may as TNFa and IFNg;9 and in this study, a close correlation even prove beneficial for distinct organ injury is not between their expression and systemic levels completely understood.36–40 We analyzed the expression was seen (Supplementary Table S2). CCL3–5 chemokines of the IFNg-inducible chemokines, CXCL9–11,41–43 and bind to (among others) CCR1, CCR3 and CCR5, which their receptor CXCR3 in GVHD target organs. RNA are expressed on T cells, NK cells, and was isolated from organs at weeks 1, 2, 3 and 6 after dendritic cells.33 syngeneic or allogeneic BMT as described in Materials CCL3 expression was significantly increased in all four and methods, and was determined organs after allo-BMT and at any time point when using real-time reverse transcription (RT)-PCR. The compared with syngeneic controls (Figure 3). Interest- tongue was used as an equivalent for skin GVHD based ingly, in the GI tract and tongue, CCL3 expression was on earlier studies showing comparable kinetics of highest at week 1 after BMT and then decreased, whereas lymphocytic infiltration and epithelial cell apoptosis in the liver and lung, it constantly increased from week 1 between the tongue and the ear skin tissue as well as to week 3. CCL4 expression in the GI tract was not based on the presence of rete-like prominences of the significantly changed, and similar to CCL3, it was squamous epithelium of the murine dorsal tongue, significantly increased in the tongue, liver and lung,

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 689

Figure 1 Characterization of aGVHD after allo-BMT. Lethally irradiated B6D2F1 mice received 5 Â 106 bone marrow cells supplemented with 10 Â 106 splenocytes from either syngeneic (B6D2F1, white bar, ----E----)orallogeneic (C57BL/6, black bar, ) donors. (a) Weight loss, (b) clinical GVHD scores, and (c and d) IFNg and TNFa serum levels were determined at indicated time points after BMT. (e) Organ injury was assessed by histopathology at weeks 1, 3 and 6 after BMT using a semiquantitative scoring system as described in Materials and methods. (f) CD3 þ immunohistochemistry (brown stain) of the liver, lung, GI tract and skin from syngeneic and allogeneic animals was carried out at weeks 1 and 3 after allo-BMT. Data are presented as mean±s.e.m. and are from one representative experiment of two similar transplantations, syngeneic (n ¼ 4 at each time point) and allogeneic (n ¼ 4–7 at each time point); *Po0.05 and **Po0.01.

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 690

Figure 2 mRNA expression of the IFNg-inducible chemokines CXCL9–11 and their receptor CXCR3. Lethally irradiated B6D2F1 mice received 5 Â 106 bone marrow cells supplemented with 10 Â 106 splenocytes from either syngeneic (B6D2F1, white bar) or allogeneic (C57BL/6, black bar) donors. Expression levels of IFNg-inducible chemokines CXCL9–11 and their receptor CXCR3 were determined in GVHD target organs by real-time PCR at weeks 1, 2, 3 and 6 after BMT. Data are combined from two independent experiments and presented as mean±s.e.m.; syngeneic (n ¼ 4–8 at each time point) and allogeneic (n ¼ 7–14 at each time point) animals; *Po0.05 and **Po0.01.

peaking at around week 2 to week 3 after allo-BMT. not be directly associated with organ involvement during CCL5 expression levels were highest at week 1 only in aGVHD, and that for the development of mucocutaneous the tongue and then constantly decreased over time. aGVHD, chemokine receptors other than CCR1 and Expression kinetics in the liver and lung were similar to CCR5 have a more critical role. In the gut, CCR1 those of CCL3 and CCL4, reaching maximum levels in expression was not changed for the first 2 weeks and the liver by week 3 and in the lung by week 2. CCL3 then showed a slight but significant increase at weeks 3 protein levels did not only directly correlate with respect and 6, whereas in the liver and lung, it was already high to their changed kinetics when compared with mRNA at week 1 and stayed at the same level until week 3 and expression levels but were also significantly increased in week 2, respectively. CCR5 expression was increased the liver and the GI tract at 1 and 3 weeks after allo-BMT, early in the GI tract at week 1 after allo-BMT, decreased and CCL5 showed a similar expression kinetics on both at week 2 and then increased from week 3 until week 6 in the mRNA and the protein levels at these time points parallel to CCR1 expression. In the liver, CCR5 expres- (Figure 7). sion increased over the first 3 weeks after allo-BMT, and We then analyzed the expression of CCR1, CCR3 and in the lung, it was increased only at week 2. It is CCR5 to determine whether upregulation of their ligands reasonable to assume that, although early CCL3 expres- is associated with increased receptor expression in sion in the gut, which has been previously shown to be individual target organs of aGVHD. CCR3 expression host-cell-dependent,50 correlates with increased GI in organs of animals after allo-BMT was not statistically CCR5, but not CCR1 expression (Supplementary Table different from the CCR3 organ expression in naive S2), CCL4 and CCL5 may contribute to the gut GVHD to animals or in syngeneic recipients (Supplementary a much lesser extent during this period. Changes in Tables S6–S9), and despite a strong upregulation of CCL3, CCL4 and CCL5 expression observed later CCL3, CCL4 and CCL5 in the tongue, no significant coincide with increased CCR1 and CCR5 expression in changes were seen for CCR1 or CCR5 expression. These all organs, most likely reflecting the recruitment of findings suggest that CCR3 expression, in general, may CCR1- and CCR5-expressing leukocytes,11,31,50,51 which

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 691

Figure 3 mRNA expression of CCL3–5, CCR1 and CCR5. Lethally irradiated B6D2F1 mice received 5 Â 106 bone marrow cells supplemented with 10 Â 106 splenocytes from either syngeneic (B6D2F1, white bar) or allogeneic (C57BL/6, black bar) donors. Expression levels of CCL3–5, CCR1 and CCR5 were determined in GVHD target organs by real-time PCR at week 1, 2, 3 and 6 after BMT. Data are combined from two independent experiments and presented as mean±s.e.m.; syngeneic (n ¼ 4–8 at each time point); allogeneic (n ¼ 7–14 at each time point) animals; *Po0.05 and **Po0.01.

themselves are capable of producing CCL3 and CCL5 as Despite no increase in expression of its ligand, a mechanism of positive reinforcement for effector cell CXCL16, was seen in the lung, CXCR6 was highly recruitment to organ sites.50 Supportive of these findings, expressed in this organ as early as 1 week after allo-BMT, a correlation was found for CCR5 and, to a lesser extent, followed by a continuous decrease. In the liver, the for CCR1 with target organ pathology at both early (the expression of CXCR6 paralleled the expression of gut and skin) and later (the gut, skin, liver and lung) time CXCL16. In the GI tract, CXCR6 expression was strongly points after allogeneic BMT (Supplementary Table S2). elevated at week 1, temporarily declined and then, along with rising ligand expression levels, it increased again Expression of CXCL16 and CXCR6 until week 6. No statistical differences in CXCR6 CXCL16 is expressed in both lymphoid and non- expression were seen in the tongue after syngeneic lymphoid tissues33,52–54 and binds to the T-cell-expressed versus allogenic BMT. receptor, CXCR6, regulating both to the and to inflammatory sites.9 After BMT, differ- Expression of XCL1 and XCR1 ences in CXCL16 expression between syngeneic and XCL1 belongs to the C chemokine family and is allogeneic recipients were seen in the liver and GI tract, produced by different cell types, including mast cells,33 but not in the lung and tongue (Figure 4a). Hepatic activated CD4 þ and CD8 þ T cells,55–58 and NK cells.59 CXCL16 expression was significantly increased between Its receptor, XCR1, is expressed on T cells, NK cells, weeks 1 and 3, and then decreased again. In the GI tract, neutrophils and B cells.33 CXCL16 expression was not different early between After BMT, XCL1 expression levels significantly syngeneic and allogeneic recipients, but then constantly differed between syngeneic and allogeneic recipients increased in allogeneic recipients until week 6. in all four organs (Figure 4b), whereas the upregulation

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 692

Figure 4 mRNA expression of CXCL16, CXCR6, XCL1 and XCR1. Lethally irradiated B6D2F1 mice received 5 Â 106 bone marrow cells supplemented with 10 Â 106 splenocytes from either syngeneic (B6D2F1, white bar) or allogeneic (C57BL/6, black bar) donors. (a) Expression levels of CXCL16 and CXCR6 and of (b) XCL1 and XCR1 were determined in GVHD target organs by real-time PCR at week 1, 2, 3 and 6 after allo-BMT. Data are combined from two independent experiments and presented as mean±s.e.m.; syngeneic (n ¼ 4–8 at each time point) and allogeneic (n ¼ 7–14 at each time point) animals; *Po0.05 and **Po0.01.

of XCR1 was only seen after allo-BMT in the liver CCR2 have been shown in earlier studies to be critically and lung. involved in T-cell and / migration Intestinal XCL1 expression started to increase 1 week to GVHD target organs including the lung.25,29 CCL2, after allo-BMT and constantly increased until week 6. In CCL8 and CCL12 expression increased in the lung as contrast, XCL1 expression peaked in the tongue at early as 1 week after allo-BMT, peaked by week 2 and week 1 and then decreased from week 2 to week 6, and then rapidly declined (Figure 5). In the liver, peak similarly, in the lung it was increased between week 1 expression of CCL2 and CCL12 was seen at week 1 after and week 2, before decreasing until week 6. In the liver, it allo-BMT, followed by a constant decrease until week 6, was elevated between week 1 and week 3, and then when expression of neither of these two chemokines returned to syngeneic levels at week 6. Differences in differed from syngeneic controls. No significant changes XCR1 expression between syngeneic and allogeneic in expression were seen for CCL8 in the liver, for CCL2, recipients were less evident. In correlation with XCL1 CCL8 and CCL12 in the GI tract, and for CCL2 and expression, both in the liver and in the lung XCR1 CCL12 in the tongue at any time point, the latter organ expressions were increased by week 2 (Supplementary showing significant increases only for CCL8 at weeks 2 Table S3) and then decreased during week 3 to week 6. and 3 after allo-BMT. Early upregulation of CCL2 and CCL12 in the liver and of CCL2, CCL8 and CCL12 Expression of CCL2, CCL8, CCL12 and CCR2 in the lung after allo-BMT was associated with a In addition to CCL8 and CCL12, CCL2 is the major delayed increase in CCR2 expression levels in these ligand of CCR2, and interactions between CCL2 and organs, which reached maximum values around

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 693

Figure 5 mRNA expression of CCL2, CCL8, CCL12 and CCR2. Lethally irradiated B6D2F1 mice received 5 Â 106 bone marrow cells supplemented with 10 Â 106 splenocytes from either syngeneic (B6D2F1, white bar) or allogeneic (C57BL/6, black bar) donors. Expression levels of CCL2, CCL8, CCL12 and their receptor CCR2 were determined in GVHD target organs by real-time PCR at week 1, 2, 3 and 6 after allo-BMT. Data are combined from two independent experiments and presented as mean±s.e.m.; syngeneic (n ¼ 4–8 at each time point); allogeneic (n ¼ 7–14 at each time point) animals; *Po0.05 and **Po0.01.

week 3 followed by a subsequent decrease, and was organ, except for the lung at week 1. Interestingly, closely correlated with histopathology (Supplementary however, CXCL2 expression levels both in the GI tract Table S4). and in the tongue were strongly elevated at 1 week after both syngeneic and allo-BMT, suggesting that the Expression of CXCL1, CXCL2 and CXCR2 upregulation of CXCL2 expression at this time is rather CXCL1 and CXCL2 belong to the CXC chemokine related to radiation toxicity than to alloreactivity. In subfamily and are produced by endothelial cells, contrast to the early increase in ligand expression in the epithelial cells and APCs.33 They have a direct role as GI tract, tongue and lung, CXCR2 expression in these chemoattractants for CXCR2-expressing neutrophils,60,61 organs was not significantly increased (Figure 6a). but can also mediate T-cell recruitment through the However, in the liver, in which extreme upregulation of induction of CXCL9 and CXCL10.60 Except for the GI CXCL1 expression was observed, CXCR2 mRNA levels tract, CXCL1 expression was strongly increased between were significantly elevated at weeks 2 and 3 in associa- week 1 and week 2 after allo-BMT in all target organs tion with increased numbers of infiltrating neutrophils (Figure 6a) and correlated with (Figures 6b and c), and correlated with organ pathology expression (Supplementary Table S5). Upregulation was at week 6 (Supplementary Table S5). highest in the liver, followed by the tongue and lung, and whereas expression in the latter two organs as well as in Early inflammation after allo-BMT induces the expression the GI tract did not differ from that of syngeneic controls of inflammatory chemokines in the at later time points, CXCL1 expression remained Finally, we analyzed whether systemic inflammation significantly elevated in the liver throughout the dura- early after allo-BMT also induces the expression tion of the study and correlated with the severity of of inflammatory chemokines in secondary lymphoid hepatic pathology (Supplementary Table S5). In addition, tissues such as the spleen. As expected, increased an association between increased mRNA levels and systemic levels of IFNg or TNFa led to enhanced protein levels of CXCL1 expression was seen when tested expression of various inflammatory chemokines, includ- for the liver and gut at 1 or 3 weeks after BMT (Figure 7). ing CXCL9–11, CCL3–5, CCL2, CCL8, CCL12 and XCL1 No significant differences in CXCL2 expression between (Supplementary Table S10). However, no increases in the syngeneic and allogeneic recipients were seen in any expression of the respective chemokine receptors, that is,

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 694

Figure 6 mRNA expression of CXCL1, CXCL2 and CXCR2. Lethally irradiated B6D2F1 mice received 5  106 bone marrow cells supplemented with 10  106 splenocytes from either syngeneic (B6D2F1, white bar) or allogeneic (C57BL/6, black bar) donors. (a) Expression levels of CXCL1, CXCL2 and CXCR2 were determined in GVHD target organs using real-time PCR at week 1, 2, 3 and 6 after allo-BMT. Data presented are combined from two independent experiments and presented as mean±s.e.m.; syngeneic (n ¼ 4–8 at each time point); allogeneic (n ¼ 7–14 at each time point) animals. (b) Hepatic neutrophil infiltration in syngeneic and allogeneic animals at week 1 and 3 after allo-BMT (hematoxylin–eosin; original magnification,  200 (left and middle panels) or  400 (right panels). (c) Semiquantitative analysis (0 ¼ normal, 2 ¼ focal mild, 4 ¼ diffuse mild, 6 ¼ diffuse severe) of neutrophils in the liver at 1 and 3 weeks after BMT from syngeneic (B6D2F1, white bar) or allogeneic (C57BL/6, black bar) donors from one experiment representative of two comparable transplantations; syngeneic: n ¼ 3 (week 1) and n ¼ 4 (week 3); allogeneic: n ¼ 4 (week 1) and n ¼ 7 (week 3) animals. *Po0.05 and **Po0.01.

CXCR3, CCR1, CCR5, CCR2 and XCR1 were seen such as cytokines and chemokines, regulate and promote (Supplementary Table S10), suggesting that cells expres- the interactions of cellular components, such as T cells sing these receptors may rapidly leave secondary and APCs, and their recruitment to target sites. Two lymphoid tissues and migrate to classical GVHD target earlier studies18,19 on the expression of chemokines after organs. allo-BMT focused on the first 10 days after allo-BMT, comparing the effects of lethal with those of non- myeloablative conditioning, or described chemokine Discussion expression and T-cell infiltration in GVHD target organs of unconditioned SCID recipient mice,18,19 both studies Allogeneic bone marrow transplantation is an important considering only a small number of chemokines. Other therapy for many disorders affecting the hematopoietic studies, in which gene expression profiling after allo- and immune system.1 However, its broader application is BMT was carried out, included a high number of limited by the development of aGVHD. Soluble factors, chemokines, but restricted their study to only one

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 695

Figure 7 Chemokine expression is increased both on mRNA and on protein levels in aGVHD target organs after allogeneic BMT. Protein expression levels of CXCL1, CXCL9, CCL3 and CCL5 were determined in the liver and GI tract at 1 and 3 weeks after BMT as described in Materials and methods. Data are combined from two independent experiments and presented as mean±s.e.m.; syngeneic (white): n ¼ 8 (week 1) or 7 (week 3); allogeneic (black): n ¼ 11 (week 1) or n ¼ 13 (week 3); **Po0.01. organ.20,62,63 In our study, we analyzed the expression of functionally involved in the early recruitment of these chemokines and chemokine receptors in four major cells to the liver, but not to the gut.66,67 Consistent with aGVHD target organs at four different time points during these findings, in this study, hepatic expression of both a time course of 6 weeks after allo-BMT in order to better CXCR6 and its ligand, CXCL16, was strongly elevated understand the kinetics of chemokine expression in from week 1 to week 3, supporting the concept that the distinct target organs and their role in organ-specific migration of CXCR6 þ T cells to the liver in response to aGVHD pathophysiology on the transcriptional level. increased ligand expression is operative in aGVHD of the Inflammatory chemokines are expressed in inflamed liver. Intestinal CXCL16 expression was not altered at 1 tissues by both hematopoetic and non-hematopoetic cells week after allo-BMT and then continuously increased upon stimulation by pro-inflammatory cytokines, in- from week 2 to week 6. In the lung, it was not increased cluding TNFa and IFNg.64 We found that, except for GI at any time point. In contrast, intestinal and pulmonary CXCL10 and hepatic CXCR3 expression, increased CXCR6 expression was significantly elevated as early as systemic IFNg levels early after allo-BMT were generally 1 week after allo-BMT. As Ueha et al.67 reported no effect associated with higher organ expression levels of the of anti-CXCL16 treatment on intestinal GVHD within the IFNg-inducible chemokines CXCL9–11 and of their first 2 weeks after BMT, it can be speculated that the receptor, CXCR3. After early peak expression, CXCL9– CXCR6:CXCL16 interactions potentially contribute to T- 11 expression levels declined until week 6, whereas cell infiltration of the intestine at later time points. Early CXCR3 expression reached a second peak at 6 weeks CXCR6 expression in the absence of increased ligand again in the gut and tongue. In addition, the CXCR3 expression instead may be due to co-expression of expression was not significantly altered in the liver at CXCR6 on T cells being recruited through other any time point after allo-BMT. These findings support mechanisms. A similar phenomenon can be proposed previous reports on the expression of CXCR3 or its for increased CXCR6 expression in the lung without any ligands during aGVHD18,21,22,24 and on the role of IFNg- changes in CXCL16 expression. CXCL9–11-regulated recruitment of CXCR3 þ T cells to Inflammatory chemokines of the CC, C or CXC3C aGVHD target organs.22,24,28,31 However, to what extent families are also increasingly expressed after allo- the IFNg-CXCL 9–112CXCR3 cascade contributes to BMT.11,25–27,50,68 As suggested by Wysocki et al.,32 cellular the multi-facetted mechanism of leukocyte recruitment sources of chemokines may differ between specific target to specific GVHD target organs is not yet completely organs. Our results showing peak CCL3 expression in understood and may further depend on conditioning the gut and tongue early after allo-BMT, followed by a intensity18,31 or specific CXCR3 þ T-cell subsets being continuous decline, strengthen the concept that increased recruited.39,65 CCL3 production in these organs derives from host CXCR6 has been shown to be expressed on liver- cells,50 whereas in contrast, the observed increase in infiltrating CD8 þ T cells after allo-BMT and to be CCL3 expression in the liver and lung over the first 3

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 696 weeks after allo-BMT may be dependent on the expres- The role of neutrophils in the development of aGVHD sion by infiltrating T cells.50 Interestingly, expression is not well understood. CXCR2 expression was only levels of CCL4 and CCL5, which both share CCR5 as a increased in the liver, and expression of its ligand CXCL1 receptor with CCL3, were not significantly elevated in was several fold stronger in the liver than in the lung and the GI tract of allogeneic recipients. Also, in contrast to tongue, being associated with increased neutrophil CCR5, CCR1 expression was not increased in the gut infiltration. Therefore, the CXCL1:CXCR2-mediated re- early after allo-BMT, suggesting that the CCL3:CCR5 cruitment of granulocytes potentially contributes to interactions rather than CCL4, CCL5 or CCR1 are hepatic organ injury,83,84 and as CXCL1 has been further involved in the early recruitment of T cells to the GI attributed an inductive role on the production of CXCL9– tract.32,50,51 In the tongue, CCL4 levels persisted highly 10 by T cells,60 an indirect mechanism may exist to elevated for the first 3 weeks after allo-BMT, whereas promote and perpetuate the development of aGVHD. after an initial peak expression, CCL5 levels continu- Many chemokines, such as CXCL9–11, CCL2, CCL8, ously decreased. In addition, despite the increased CCL12, were highly expressed in parallel to systemic expression of their ligands, CCR1 and CCR5 levels were TNFa or IFNg protein levels and before the appearance not elevated, suggesting that in this organ system other of significant histopathological changes in individual chemokines and chemokine receptors may mediate aGVHD target organs. However, expression of others, for aGVHD-related injury.69 Similar to CCL3, pulmonary example, CCL3–CCL5 or XCL1, clearly coincided with and hepatic expression levels of CCL4 and CCL5 the development of lymphocytic infiltrates in specific increased until weeks 2–3 and were associated with organs, supporting the hypothesis that chemokine increased CCR1 and CCR5 expression. Previous reports induction in GVHD target organs after allo-BMT is prone indicated increased CCL3 and CCL5 expression by to two major mechanisms: (1) ‘ectopic’ early inflamma- alloreactive T cells in the course of aGVHD develop- tory cytokine production as soluble mediators, leading to ment;26,50 hence it is likely that in addition to CCL350,70 in situ chemokine expression in an endocrine manner20 and CCL5 expression,26 CCL4 production by infiltrating and (2) chemokine induction through infiltrating donor T cells potentially contributes to a chemokine-mediated leukocytes within target organs in situ, contributing to a feedback mechanism on the recruitment of CCR1 þ or chemokine-mediated feedback mechanism on the re- CCR5 þ donor T cells to these organs. cruitment of these cells. Furthermore, although the The role of XCL1 (lymphotactin a) in the development former mechanism most likely drives early aGVHD of aGVHD has not yet been defined. Predominantly development, the possibility that specifically the latter produced by activated Th1 CD4 þ , CD8 þ T cells and mechanism, in addition to perpetuating local alloreactive NK cells,71,72 XCL1 ligation to its receptor XCR1 not only injury, may be involved in suppressing prolonged and induces T and NK cell recruitment and co-activates dysregulated inflammation during the later course of macrophages but also can increase regulatory T-cell disease, and therefore may be partially responsible for function, suppress the IFNg production of CD4 þ T cells decreased expression levels of many chemokines as and seems to be involved in organ rejection.71,73–78 observed at later time points as well cannot be Increased levels of XCL1 have been reported in the excluded.32,73 bronchoalveolar lavage fluid of mice suffering from Alloantigen presentation by host- and potentially idiopathic pneumonia syndrome after allogeneic donor-derived DCs with subsequent T-cell activation is BMT,79 and antigen priming of CTLs was enhanced in a critical mechanism operative in the development of the presence of XCL1.80 Therefore, it can be speculated, GVHD. For the activation and maturation of these cells, that XCL1–XCR1 interactions can promote the inflam- TNFa and induction of the nuclear transcription factor matory process of aGVHD through an autocrine NF-kB are important,85,86 and in parallel to the decrease or a paracrine pathway or potentially mediate a down- in these factors over time, DC maturation may not be regulation of inflammation through regulatory and completed, potentially leading to the inhibition and tolerogenic effects. alteration of T-cell responses87,88 and co-contributing to In our study, increased expression levels of CCR2 were the downregulation of chemokine expression at later preceded by elevated CCL2 and CCL12 expression in the time points. liver and lung and were associated with the observed In summary, our study shows the complexity of a kinetics of histopathological changes in these organs, consistently growing network of chemokines and che- whereas neither receptor nor ligand was significantly mokine receptors in the development of aGVHD. Future altered in the GI tract or tongue. Using the same experimental studies, which will extend the current haploidentical BMT model, a critical role of donor knowledge on specific chemokine–chemokine interac- leukocyte expression of CCR2 has been shown for the tions and their role in the development of organ development of pulmonary GVHD,25 and in a fully complications after allo-BMT, are needed to identify mismatched BMT model, CCR2 expression of the host specific targets for novel preventive and therapeutic did not contribute to pulmonary dysfunction after allo- strategies against the tremendous clinical challenges of BMT.81 Terwey et al.29 reported the requirement of CCR2 aGVHD initiation and progression. expression on donor cells for intestinal and hepatic aGVHD, but not for cutaneous CD8 þ T-cell-mediated aGVHD. In CD4 þ T-cell-mediated hepatic aGVHD, CCR2 deficiency of donor T cells rather led to increased Materials and methods T-cell infiltrates.82 Therefore, the exact role of CCR2 may Mice and stem cell transplantation differ in relation to the degree of MHC mismatch and a Female C57BL/6 (H-2b) and B6D2F1 (H-2bxd) mice were leukocyte subpopulation, and seems to vary between purchased from Charles River Laboratories (Sulzbach, specific GVHD target organs. Germany) and acclimatized in our animal facility for at

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 697 least 1 week before the experiments. The animals were frozen in liquid nitrogen. At the time of analysis, 70 mg between 10 and 20 weeks at the time of BMT. All animal of organ tissue was homogenized in 1 ml buffer solution experiments were approved by the local institutional (1 Â PBS and 1 tablet of Complete Protease Inhibitor animal committee of the University of Regensburg and Cocktail (Roche Diagnostics, Basel, Switzerland)), cen- were in accordance with German Animal Protection Laws. trifuged at 1500 r.p.m. for 15 min at 4 1C, and the Naive, untreated mice were transplanted according to supernatant was harvested. Total protein concentration a standard protocol as described previously.27 On the day in the supernatant was determined using the Bio-Rad of BMT, B6D2F1 recipient mice received lethal total body Protein Assay (Bio-Rad, Hercules, CA, USA) before irradiation with a total dose of 13 Gy, delivered in two chemokines were measured by ELISA as described fractions 3 h apart to reduce GI toxicity, using a linear below. Chemokine concentration in the supernatant accelerator at 150 cGy minÀ1. Total body irradiation was was normalized to picograms per milligram of total followed by the infusion of cell mixtures of 5 Â 106 bone protein (pg per ml of total protein). marrow cells supplemented with 10 Â 106 splenocytes from either syngeneic (B6D2F1) or allogeneic (C57BL/6) Measurements of IFNg and TNFa serum levels and chemokine donors. Syngeneic and allogeneic groups were analyzed levels in the liver and GI tract by ELISA for chemokine and chemokine receptor expression levels Serum protein levels of IFNg and TNFa and of CCL3, and for GVHD target organ histopathology at weeks 1, 2, CCL5, CXCL1 and CXCL9 in the liver and gut were 3 and 6 after transplantation. Transplanted mice were measured by ELISA (IFNg and TNFa: OptEIA, BD housed in micro-isolator cages with autoclaved bedding Pharmingen, Heidelberg, Germany; CCL3, CCL5, CXCL1 and fed autoclaved chow and autoclaved or hyperchlori- and CXCL9: Quantikine M, R&D Systems, Wiesbaden, nated water. Germany). Assays were carried out according to the manufacturer’s protocols. Clinical GVHD and survival Immunohistochemistry Survival was monitored daily until day þ 42 after BMT, For immunostaining, 4 mm of organ slices were dewaxed and clinical GVHD scores were assessed weekly by a and re-hydrated, followed by incubation in 3% H2O2 for scoring system incorporating five clinical parameters, 10 min to quench endogenous peroxidase activity. CD3 þ namely weight loss, posture (hunching), mobility, fur cells were then detected by rabbit IgG anti-mouse CD3 texture and skin integrity.89 Each parameter was graded antibody (Sigma, St Louis, MO, USA) and by a goat-anti- between 0 and 2. Once an animal reached a cumulative rabbit IgG-HRP (horseradish peroxidase) secondary score of X6.5 or a weight loss of more than 35%, it was antibody (Dako, Hamburg, Germany). Immunoenzy- killed and counted as death due to transplantation- matic reactions were carried out using HRP and stains related mortality. were shown using 3,30-diaminobenzidine (Dako). The slides were counterstained with hematoxylin. Semiquantitative histopathology of the lung, liver, intestine and skin Determination of organ mRNA expression levels using At weeks 1, 2, 3 and 6 after transplantation, animals real-time PCR were killed for histopathological analysis. Organs were RNA RT. Total cellular RNA was extracted using removed, fixed in 4% paraformaldehyde dissolved in RNeasy Mini Kits (Qiagen, Hilden, Germany) according phosphate-buffered saline, transferred into 70% ethanol to the manufacturer’s protocol. First-strand cDNA was after 48 h, paraffin-embedded and then sectioned into synthesized using 1 mg of total RNA (DNase-treated): 4-mm-thick slices. Hematoxylin–eosin-stained lung sec- 1 mg of total RNA in 8 ml of RNase-free water was tions from individual mice were coded without reference denatured for 5 min at 75 1C, then 10 ml of RT reaction to mouse type and prior treatment and independently mixture (Invitrogen, Karlsruhe, Germany) and 2 mlofRT examined by a pathologist to establish an index of enzyme mixture (Invitrogen, Karlsruhe, Germany) were GVHD injury. Detailed analysis of changes in the GI added. The RNA and RT mixture was incubated for tract, liver and lung histology were carried out in a 10 min at 25 1C, followed by 60 min at 44 1C and a blinded manner at 1, 2, 3 and 6 weeks after transplanta- denaturing step for 5 min at 85 1C. cDNA was stored at 24–26 tion as described earlier. Histopathology of the skin À20 1C until use for real-time PCR. was assessed at the same time points by grading the extent of epithelial cell apoptosis and lymphocellular Real-time PCR. All real-time PCRs were carried out in a infiltrates from 0 to 4 (none, mild, mild-moderate, 10-ml mixture containing 4 ml of sample, 5 ml SYBR Green moderate and severe). The numbers of neutrophils in mix (Invitrogen, Karlsruhe, Germany), 0.25 ml of sense the liver at weeks 1 and 3 was assessed by an primer (10 mM), 0.25 ml of antisense primer (10 mM) and independent pathologist in a semiquantitative and 0.5 ml water. Real-time PCR quantifications were carried blinded manner and graded with 0 (normal), 2 (focal out using the ABI PRISM 7900HT Sequence Detection mild), 4 (diffuse mild) or 6 (diffuse severe). The analysis System (Applied Biosystems, Foster City, CA, USA). The was carried out using microscopy (Axioskop 2 plus, fluorescence threshold value was calculated using the Carl Zeiss GmbH, Jena, Germany). Photographs of ABI PRISM 7900HT Sequence Detection System software histopathology were obtained by AxioCam HRc (Carl (version SDS 2.2, Applied Biosystems). GAPDH was Zeiss GmbH) and processed using AxioVision Release used as a reference. The primers used for real-time PCR 4.6.3 (Carl Zeiss GmbH). are listed in Table 1. Lung, liver and gut chemokine analysis Statistical considerations After exsanguination, the organs were harvested at 1 or 3 For statistical analysis, an algorithm of statistical tests weeks after transplantation and immediately snap- was applied. First, we used analysis of variance

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 698 Table 1 Primers used for real-time PCR Table 1 Continued

GAPDH sense CATCACTGCCACCCAGAAGA GAPDH antisense CAGATCCACGACGGACACAT CCR10 sense CCAGCAAGCGCAAGGATCTA CCR10 antisense AGCAGGAAGAAAGGCGGAGT XCL1 sense AGCGGCTGCCAGTTCAAA XCL1 antisense GCTCCTGTGGGAACAGTTTCAG CXCR1 sense TGGGGGTGATCTTTGCTGTT CXCR1 antisense TTTGGCCAACGAAGGCATAG CCL2 (MCP-1) sense GCTGACCCCAAGAAGGAATG CCL2 (MCP-1) antisense GTGCTTGAGGTGGTTGTGGA CXCR2 sense CATCTTCGCTGTCGTCCTTGT CXCR2 antisense GCTGTGGAGGAAGCCAAGAA CCL3 (MIP-1a) sense TGCCTGCTGCTTCTCCTACA CCL3 (MIP-1a) antisense TGGACCCAGGTCTCTTTGGA CXCR3 sense TAGTGGTGGTGGCAGCCTTT CXCR3 antisense AGGCATAGAGCAGCGGATTG CCL4 (MIP-1b) sense CCAGGGTTCTCAGCACCAA CCL4 (MIP-1b) antisense GCTCACTGGGGTTAGCACAGA CXCR6 sense CTTTCGGGCTTGCCTTAACC CXCR6 antisense CATTGTGGGAGGCAGAACAA CCL5 (RANTES) sense CTCACCATATGGCTCGGACA CCL5 (RANTES) antisense CTTCTCTGGGTTGGCACACA CX3CR1 sense CTGAGGTTGGCCCTCAGTGT CX3CR1 antisense GAATGCTGTCCTGCCTGCTC CCL8 (MCP-2) sense GCAGTGCTTCTTTGCCTGCT CCL8 (MCP-2) antisense ACAGCTTCCATGGGGCACT Abbreviations: GADPH, glyceraldehyde 3-phosphate dehydrogenase; I-TAC, interferon-inducible T-cell a chemoattractant; IP-10, inducible CCL12 (MCP-5) sense CAGTCCTCAGGTATTGGCTGGA protein-10; MCP-1, monocyte chemotactic protein-1; MIP, macrophage CCL12 (MCP-5) antisense TCCTTGGGGTCAGCACAGAT inflammatory protein; RANTES, regulated upon activation, normal CCL17 (TARC) sense TGCTCTGCTTCTGGGGACTT expressed and secreted. CCL17 (TARC) antisense ACAGATGAGCTTGCCCTGGA

CCL22 (MDC) sense TGGCTCTCGTCCTTCTTGCT (ANOVA) to test whether organ-specific expression CCL22 (MDC) antisense AGGCTTGCGGCAGGATTT levels of individual genes at any time point in animals CXCL1 (KC) sense GCCTATCGCCAATGAGCTG after allo-BMT and in naive, untreated animals differed. CXCL1 (KC) antisense CTGAACCAAGGGAGCTTCAGG The minimal level of confidence at which results were judged significant was Po0.05. Next, expression levels of CXCL8 (MIP-2) sense GGCTGTTGTGGCCAGTGAA those genes tested significant by ANOVAwere compared CXCL8 (MIP-2) antisense GCTTCAGGGTCAAGGCAAAC between syngeneic and allogeneic recipients at indivi- CXCL9 (MIG) sense TGGGCATCATCTTCCTGGAG dual time points using the nonparametric unpaired CXCL9 (MIG) antisense CCGGATCTAGGCAGGTTTGA Mann–Whitney U-test. Increases in expression were considered significant if changes were at least two-fold CXCL10 (IP-10) sense CCTCATCCTGCTGGGTCTG when compared with naive animals, and when the CXCL10 (IP-10) antisense CTCAACACGTGGGCAGGA P-value for the difference between syngeneic and CXCL11 (I-TAC) sense CGGGATGAAAGCCGTCAA allogeneic animals was o0.05. Data are presented as CXCL11 (I-TAC) antisense TATGAGGCGAGCTTGCTTGG mean±s.e.m. (standard error of the mean). Spearman’s r-test was used for correlation analysis of chemokine CXCL16 sense CAACCCTGGGAGATGACCAC expression with cytokine expression, pathology and CXCL16 antisense CTGTGTCGCTCTCCTGTTGC chemokine receptor expression. CX3CL1 sense CCACTGCAGATCCCCAGAAA CX3CL1 antisense GCGGAGGCCTTCTACCATTT Conflict of interest XCR1 sense TGTTTCGCACAAGGTCCAGA XCR1 antisense AGGCCAAATGGCGACAGATA The authors declare no conflict of interest.

CCR1 sense TTCCTCCTCTGGACCCCCTA CCR1 antisense TTGAAACAGCTGCCGAAGGT Acknowledgements CCR2 sense CCACACCCTGTTTCGCTGTA CCR2 antisense TGCATGGCCTGGTCTAAGTG This work was support by the Deutsche Krebshilfe e.V., CCR3 sense TTTGGACCCCGTACAACCTG Project #106647. We thank the Biostatistician Dr Chris- CCR3 antisense TTTCCGGAACCTCTCACCAA toph Ehret (Department of Hematology and Oncology, University of Regensburg Medical Center) for his advice CCR4 sense GGCCTGATGCCTTACACCAA and assistance with the statistical analyses. CCR4 antisense GGCAGCCGCTGATGTTATTC

CCR5 sense CAGGGCTGTGAGGCTCATCT CCR5 antisense GGCAGCAGTGTGTCATTCCA References

CCR7 sense CCAATAGCAGCTGCGAAACC 1 Armitage JO. Bone marrow transplantation. N Engl J Med CCR7 antisense GCAGCCCAAGTCCTTGAAGA 1994; 330: 827–838. 2 Goker H, Haznedaroglu IC, Chao NJ. Acute graft-vs-host CCR9 sense GGTCACCTTGGGGTTTTTCC disease: pathobiology and management. Exp Hematol 2001; 29: CCR9 antisense TAAGCGTCAACAGCCTGCAC 259–277.

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 699 3 Murphy PM, Baggiolini M, Charo IF, Hebert CA, Horuk R, 25 Hildebrandt GC, Duffner UA, Olkiewicz KM, Corrion LA, Matsushima K et al. International union of pharmacology. Willmarth NE, Williams DL et al. A critical role for CCR2/ XXII. Nomenclature for chemokine receptors. Pharmacol Rev MCP-1 interactions in the development of idiopathic pneu- 2000; 52: 145–176. monia syndrome after allogeneic bone marrow transplanta- 4 Murphy PM. International Union of Pharmacology. XXX. tion. Blood 2004; 103: 2417–2426. Update on chemokine receptor nomenclature. Pharmacol Rev 26 Hildebrandt GC, Olkiewicz KM, Choi S, Corrion LA, 2002; 54: 227–229. Clouthier SG, Liu C et al. Donor T-cell production of RANTES 5 Zlotnik A, Yoshie O. Chemokines: a new classification system significantly contributes to the development of idiopathic and their role in immunity. Immunity 2000; 12: 121–127. pneumonia syndrome after allogeneic stem cell transplanta- 6 Rot A, von Andrian UH. Chemokines in innate and adaptive tion. Blood 2005; 105: 2249–2257. host defense: basic chemokinese grammar for immune cells. 27 Hildebrandt GC, Olkiewicz KM, Corrion LA, Chang Y, Annu Rev Immunol 2004; 22: 891–928. Clouthier SG, Liu C et al. Donor-derived TNF-alpha regulates 7 Mackay CR. Chemokines: immunology’s high impact factors. pulmonary chemokine expression and the development of Nat Immunol 2001; 2: 95–101. idiopathic pneumonia syndrome after allogeneic bone mar- 8 Pease JE, Williams TJ. The attraction of chemokines as a target row transplantation. Blood 2004; 104: 586–593. for specific anti-inflammatory therapy. Br J Pharmacol 2006; 28 Piper KP, Horlock C, Curnow SJ, Arrazi J, Nicholls S, 147(Suppl 1): S212–S221. Mahendra P et al. CXCL10–CXCR3 interactions play an 9 Moser B, Wolf M, Walz A, Loetscher P. Chemokines: multiple important role in the pathogenesis of acute graft-versus-host levels of leukocyte migration control. Trends Immunol 2004; 25: disease in the skin following allogeneic stem-cell transplanta- 75–84. tion. Blood 2007; 110: 3827–3832. 10 Rollins BJ. Chemokines. Blood 1997; 90: 909–928. 29 Terwey TH, Kim TD, Kochman AA, Hubbard VM, Lu S, 11 Choi SW, Hildebrandt GC, Olkiewicz KM, Hanauer DA, Zakrzewski JL et al. CCR2 is required for CD8-induced graft- Chaudhary MN, Silva IA et al. CCR1/CCL5 (RANTES) versus-host disease. Blood 2005; 106: 3322–3330. receptor–ligand interactions modulate allogeneic T-cell re- 30 Varona R, Cadenas V, Gomez L, Martinez AC, Marquez G. sponses and graft-versus-host disease following stem-cell CCR6 regulates CD4+ T-cell-mediated acute graft-versus-host transplantation. Blood 2007; 110: 3447–3455. disease responses. Blood 2005; 106: 18–26. 12 Addison CL, Daniel TO, Burdick MD, Liu H, Ehlert JE, Xue 31 Wysocki CA, Burkett SB, Panoskaltsis-Mortari A, Kirby SL, YY et al. The CXC chemokine receptor 2, CXCR2, is the Luster AD, McKinnon K et al. Differential roles for CCR5 putative receptor for ELR+ CXC chemokine-induced angio- expression on donor T cells during graft-versus-host genic activity. J Immunol 2000; 165: 5269–5277. disease based on pretransplant conditioning. J Immunol 2004; 13 Strieter RM, Burdick MD, Gomperts BN, Belperio JA, Keane 173: 845–854. MP. CXC chemokines in angiogenesis. Cytokine 32 Wysocki CA, Jiang Q, Panoskaltsis-Mortari A, Taylor PA, McKinnon KP, Su L et al. Critical role for CCR5 in the function Rev 2005; 16: 593–609. of donor CD4+CD25+ regulatory T cells during acute graft- 14 Broxmeyer HE. Chemokines in hematopoiesis. Curr Opin versus-host disease. Blood 2005; 106: 3300–3307. Hematol 2008; 15: 49–58. 33 Wysocki CA, Panoskaltsis-Mortari A, Blazar BR, Serody JS. 15 Ohl L, Henning G, Krautwald S, Lipp M, Hardtke S, Leukocyte migration and graft-versus-host disease. Blood Bernhardt G et al. Cooperating mechanisms of CXCR5 and 2005; 105: 4191–4199. CCR7 in development and organization of secondary lym- 34 Ewing P, Miklos S, Olkiewicz KM, Muller G, Andreesen R, phoid organs. J Exp Med 2003; 197: 1199–1204. Holler E et al. Donor CD4+ T-cell production of tumor necrosis 16 Cyster JG. Chemokines and cell migration in secondary factor alpha significantly contributes to the early proinflam- lymphoid organs. Science 1999; 286: 2098–2102. matory events of graft-versus-host disease. Exp Hematol 2007; 17 Muller G, Hopken UE, Lipp M. The impact of CCR7 and 35: 155–163. CXCR5 on lymphoid organ development and systemic 35 Zhang Y, Louboutin JP, Zhu J, Rivera AJ, Emerson SG. immunity. Immunol Rev 2003; 195: 117–135. Preterminal host dendritic cells in irradiated mice prime 18 Mapara MY, Leng C, Kim YM, Bronson R, Lokshin A, Luster A CD8+ T cell-mediated acute graft-versus-host disease. J Clin et al. Expression of chemokines in GVHD target organs is Invest 2002; 109: 1335–1344. influenced by conditioning and genetic factors and amplified 36 Burman AC, Banovic T, Kuns RD, Clouston AD, Stanley AC, by GVHR. Biol Blood Marrow Transplant 2006; 12: 623–634. Morris ES et al. IFNgamma differentially controls the devel- 19 New JY, Li B, Koh WP, Ng HK, Tan SY, Yap EH et al. T cell opment of idiopathic pneumonia syndrome and GVHD of the infiltration and chemokine expression: relevance to the gastrointestinal tract. Blood 2007; 110: 1064–1072. disease localization in murine graft-versus-host disease. Bone 37 Welniak LA, Blazar BR, Anver MR, Wiltrout RH, Murphy WJ. Marrow Transplant 2002; 29: 979–986. Opposing roles of interferon-gamma on CD4+ T cell-mediated 20 Sugerman PB, Faber SB, Willis LM, Petrovic A, Murphy GF, graft-versus-host disease: effects of conditioning. Biol Blood Pappo J et al. Kinetics of gene expression in murine cutaneous Marrow Transplant 2000; 6: 604–612. graft-versus-host disease. Am J Pathol 2004; 164: 2189–2202. 38 Kim YM, Sachs T, Asavaroengchai W, Bronson R, Sykes M. 21 Jaksch M, Remberger M, Mattsson J. Increased gene expres- Graft-versus-host disease can be separated from graft-versus- sion of chemokine receptors is correlated with acute graft- lymphoma effects by control of trafficking with versus-host disease after allogeneic stem cell transplantation. FTY720. J Clin Invest 2003; 111: 659–669. Biol Blood Marrow Transplant 2005; 11: 280–287. 39 Mauermann N, Burian J, von Garnier C, Dirnhofer S, Germano 22 Duffner U, Lu B, Hildebrandt GC, Teshima T, Williams DL, D, Schuett C et al. Interferon-gamma regulates idiopathic Reddy P et al. Role of CXCR3-induced donor T-cell migration pneumonia syndrome, a Th17+CD4+ T-cell-mediated in acute GVHD. Exp Hematol 2003; 31: 897–902. graft-versus-host disease. Am J Respir Crit Care Med 2008; 23 Hancock WW, Lu B, Gao W, Csizmadia V, Faia K, King JA et al. 178: 379–388. Requirement of the chemokine receptor CXCR3 for acute 40 Yang YG, Wang H, Asavaroengchai W, Dey BR. Role of allograft rejection. J Exp Med 2000; 192: 1515–1520. interferon-gamma in GVHD and GVL. Cell Mol Immunol 2005; 24 Hildebrandt GC, Corrion LA, Olkiewicz KM, Lu B, Lowler K, 2: 323–329. Duffner UA et al. Blockade of CXCR3 receptor:ligand 41 Farber JM. A macrophage mRNA selectively induced by interactions reduces leukocyte recruitment to the lung and gamma-interferon encodes a member of the the severity of experimental idiopathic pneumonia syndrome. family of cytokines. Proc Natl Acad Sci USA 1990; 87: J Immunol 2004; 173: 2050–2059. 5238–5242.

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 700 42 Vanguri P, Farber JM. Identification of CRG-2. An interferon- vascularized corneal allografts on T cell recruitment and graft inducible mRNA predicted to encode a murine . rejection. Transplantation 2008; 85: 615–625. J Biol Chem 1990; 265: 15049–15057. 61 Reutershan J, Morris MA, Burcin TL, Smith DF, Chang D, 43 Widney DP, Xia YR, Lusis AJ, Smith JB. The murine Saprito MS et al. Critical role of endothelial CXCR2 in LPS- chemokine CXCL11 (IFN-inducible T cell alpha chemoattrac- induced neutrophil migration into the lung. J Clin Invest 2006; tant) is an IFN-gamma- and lipopolysaccharide-inducible 116: 695–702. glucocorticoid-attenuated response gene expressed in lung 62 Ichiba T, Teshima T, Kuick R, Misek DE, Liu C, Takada Y et al. and other tissues during endotoxemia. J Immunol 2000; 164: Early changes in gene expression profiles of hepatic GVHD 6322–6331. uncovered by oligonucleotide microarrays. Blood 2003; 102: 44 Gilliam AC, Whitaker-Menezes D, Korngold R, Murphy GF. 763–771. Apoptosis is the predominant form of epithelial target cell 63 Zhou L, Askew D, Wu C, Gilliam AC. Cutaneous gene injury in acute experimental graft-versus-host disease. J Invest expression by DNA microarray in murine sclerodermatous Dermatol 1996; 107: 377–383. graft-versus-host disease, a model for human scleroderma. 45 Sale GE, Raff RF, Storb R. Stem cell regions in filiform papillae J Invest Dermatol 2007; 127: 281–292. of tongue as targets of graft-versus-host disease. Transplanta- 64 Luster AD. Chemokines—chemotactic cytokines that mediate tion 1994; 58: 1273–1275. inflammation. N Engl J Med 1998; 338: 436–445. 46 Whitaker-Menezes D, Jones SC, Friedman TM, Korngold R, 65 Hasegawa H, Inoue A, Kohno M, Lei J, Miyazaki T, Yoshie O Murphy GF. An epithelial target site in experimental graft- et al. Therapeutic effect of CXCR3-expressing regulatory T cells versus-host disease and cytokine-mediated cytotoxicity is on liver, lung and intestinal damages in a murine acute GVHD defined by cytokeratin 15 expression. Biol Blood Marrow model. Gene Therapy 2008; 15: 171–182. Transplant 2003; 9: 559–570. 66 Sato K, Nakaoka T, Yamashita N, Yagita H, Kawasaki H, 47 Baggiolini M. Chemokines and leukocyte traffic. Nature 1998; Morimoto C et al. TRAIL-transduced dendritic cells protect 392: 565–568. mice from acute graft-versus-host disease and leukemia 48 Sallusto F, Baggiolini M. Chemokines and leukocyte traffic. relapse. J Immunol 2005; 174: 4025–4033. Nat Immunol 2008; 9: 949–952. 67 Ueha S, Murai M, Yoneyama H, Kitabatake M, Imai T, 49 Flier J, Boorsma DM, van Beek PJ, Nieboer C, Stoof TJ, Shimaoka T et al. Intervention of MAdCAM-1 or fractalkine Willemze R et al. Differential expression of CXCR3 targeting alleviates graft-versus-host reaction associated intestinal in- chemokines CXCL10, CXCL9, and CXCL11 in different types jury while preserving graft-versus-tumor effects. J Leukoc Biol of skin inflammation. J Pathol 2001; 194: 398–405. 2007; 81: 176–185. 50 Serody JS, Burkett SE, Panoskaltsis-Mortari A, Ng-Cashin J, 68 Panoskaltsis-Mortari A, Strieter RM, Hermanson JR, Fegeding McMahon E, Matsushima GK et al. T-lymphocyte production KV, Murphy WJ, Farrell CL et al. Induction of monocyte- and of macrophage inflammatory protein-1alpha is critical to the T-cell-attracting chemokines in the lung during the generation recruitment of CD8(+) T cells to the liver, lung, and spleen of idiopathic pneumonia syndrome following allogeneic during graft-versus-host disease. Blood 2000; 96: 2973–2980. murine bone marrow transplantation. Blood 2000; 96: 834–839. 51 Murai M, Yoneyama H, Ezaki T, Suematsu M, Terashima Y, 69 Morita NI, Matsumura Y, Morita K, Miyachi Y. Expression of Harada A et al. Peyer’s patch is the essential site in initiating CCR5 in graft-versus-host disease (GVHD) of the skin: murine acute and lethal graft-versus-host reaction. Nat immunohistochemical staining of 38 cases. J Dermatol 2007; Immunol 2003; 4: 154–160. 34: 254–257. 52 Wilbanks A, Zondlo SC, Murphy K, Mak S, Soler D, Langdon 70 Murai M, Yoneyama H, Harada A, Yi Z, Vestergaard C, Guo B P et al. Expression cloning of the STRL33/BONZO/TYMSTR- et al. Active participation of CCR5(+)CD8(+) T in ligand reveals elements of CC, CXC, and CX3C chemokines. J Immunol 2001; 166: 5145–5154. the pathogenesis of liver injury in graft-versus-host disease. 53 Matloubian M, David A, Engel S, Ryan JE, Cyster JG. A J Clin Invest 1999; 104: 49–57. transmembrane CXC chemokine is a ligand for HIV-corecep- 71 Dorner BG, Scheffold A, Rolph MS, Huser MB, Kaufmann SH, tor Bonzo. Nat Immunol 2000; 1: 298–304. Radbruch A et al. MIP-1alpha, MIP-1beta, RANTES, and 54 Dong H, Toyoda N, Yoneyama H, Kurachi M, Kasahara T, ATAC/lymphotactin function together with IFN-gamma as Kobayashi Y et al. Gene expression profile analysis of the type 1 cytokines. Proc Natl Acad Sci USA 2002; 99: 6181–6186. mouse liver during bacteria-induced fulminant hepatitis by a 72 Muller K, Bischof S, Sommer F, Lohoff M, Solbach W, Laskay cDNA microarray system. Biochem Biophys Res Commun 2002; T. Differential production of macrophage inflammatory 298: 675–686. protein 1gamma (MIP-1gamma), lymphotactin, and MIP-2 55 Kelner GS, Kennedy J, Bacon KB, Kleyensteuber S, Largae- by CD4(+) Th subsets polarized in vitro and in vivo. Infect spada DA, Jenkins NA et al. Lymphotactin: a cytokine Immun 2003; 71: 6178–6183. that represents a new class of chemokine. Science 1994; 266: 73 Nguyen KD, Fohner A, Booker JD, Dong C, Krensky AM, 1395–1399. Nadeau KC. XCL1 enhances regulatory activities of CD4+ 56 Muller S, Dorner B, Korthauer U, Mages HW, D0Apuzzo M, CD25(high) CD127(low/À) T cells in human allergic asthma. Senger G et al. Cloning of ATAC, an activation-induced, J Immunol 2008; 181: 5386–5395. chemokine-related molecule exclusively expressed in CD8+ T 74MiddelP,ThelenP,BlaschkeS,PolzienF,ReichK,BlaschkeV lymphocytes. Eur J Immunol 1995; 25: 1744–1748. et al. Expression of the T-cell chemoattractant chemokine 57 Tikhonov I, Kitabwalla M, Wallace M, Malkovsky M, Volkman lymphotactin in Crohn’s disease. Am J Pathol 2001; 159: 1751–1761. B, Pauza CD. Staphylococcal induce lympho- 75 Kurt RA, Bauck M, Harma S, McCulloch K, Baher A, Urba WJ. tactin production by human CD4+ and CD8+ T cells. Cytokine Role of C chemokine lymphotactin in mediating recruitment 2001; 16: 73–78. of antigen-specific CD62L(lo) cells in vitro and in vivo. Cell 58 Palena C, Arlen P, Zeytin H, Greiner JW, Schlom J, Tsang KY. Immunol 2001; 209: 83–88. Enhanced expression of lymphotactin by CD8+ T cells is 76 Wang CR, Liu MF, Huang YH, Chen HC. Up-regulation of selectively induced by enhancer agonist of tumor- XCR1 expression in rheumatoid joints. Rheumatology (Oxford) associated antigens. Cytokine 2003; 24: 128–142. 2004; 43: 569–573. 59 Hedrick JA, Saylor V, Figueroa D, Mizoue L, Xu Y, Menon S 77 Ordway D, Higgins DM, Sanchez-Campillo J, Spencer JS, et al. Lymphotactin is produced by NK cells and attracts both Henao-Tamayo M, Harton M et al. XCL1 (lymphotactin) NK cells and T cells in vivo. J Immunol 1997; 158: 1533–1540. chemokine produced by activated CD8T cells during the 60 Amescua G, Collings F, Sidani A, Bonfield TL, Rodriguez JP, chronic stage of infection with Mycobacterium tuberculosis Galor A et al. Effect of CXCL-1/KC production in high risk negatively affects production of IFN-gamma by CD4T cells

Genes and Immunity Chemokines and their receptors in acute GVHD A Bouazzaoui et al 701 and participates in granuloma stability. J Leukoc Biol 2007; 82: 83 Nikolic B, Lee S, Bronson RT, Grusby MJ, Sykes M. Th1 1221–1229. and Th2 mediate acute graft-versus-host disease, each 78 Streblow DN, Kreklywich C, Yin Q, De La Melena VT, Corless with distinct end-organ targets. J Clin Invest 2000; 105: CL, Smith PA et al. Cytomegalovirus-mediated upregulation 1289–1298. of chemokine expression correlates with the acceleration 84 Nonomura A, Kono N, Mizukzmi Y, Nakanuma Y, Matsubara of chronic rejection in rat heart transplants. J Virol 2003; 77: F. Histological changes in the liver in experimental graft- 2182–2194. versus-host disease across minor histocompatibility barriers. 79 Panoskaltsis-Mortari A, Hermanson JR, Haddad IY, Wangens- VI. A light and electron microscopic study of the periportal teen OD, Blazar BR. Intercellular adhesion molecule-I (ICAM- changes. Liver 1991; 11: 278–286. I, CD54) deficiency segregates the unique pathophysiological 85 Koski GK, Lyakh LA, Cohen PA, Rice NR. CD14+ requirements for generating idiopathic pneumonia syndrome as precursors: diverse maturation-inducing (IPS) versus graft-versus-host disease following allogeneic pathways lead to common activation of NF-kappab/RelB. Crit Rev Immunol 2001; 21: 179–189. murine bone marrow transplantation. Biol Blood Marrow 86 Ouaaz F, Arron J, Zheng Y, Choi Y, Beg AA. Dendritic cell Transplant 2001; 7: 368–377. development and survival require distinct NF-kappaB sub- 80 Matsuyoshi H, Senju S, Hirata S, Yoshitake Y, Uemura Y, units. Immunity 2002; 16: 257–270. Nishimura Y. Enhanced priming of antigen-specific CTLs in 87 Buonocore S, Flamand V, Goldman M, Braun MY. Bone vivo by embryonic stem cell-derived dendritic cells expressing marrow-derived immature dendritic cells prime in vivo chemokine along with antigenic protein: application to alloreactive T cells for -4-dependent rejection of antitumor vaccination. J Immunol 2004; 172: 776–786. major histocompatibility complex class II antigen-disparate 81 Panoskaltsis-Mortari A, Hermanson JR, Taras E, Wangensteen cardiac allograft. Transplantation 2003; 75: 407–413. OD, Charo IF, Rollins BJ et al. Post-BMT lung injury occurs 88 Nouri-Shirazi M, Guinet E. Direct and indirect cross-tolerance independently of the expression of CCL2 or its receptor, of alloreactive T cells by dendritic cells retained in the CCR2, on host cells. Am J Physiol Lung Cell Mol Physiol 2004; immature stage. Transplantation 2002; 74: 1035–1044. 286: L284–L292. 89 Cooke KR, Kobzik L, Martin TR, Brewer J, Delmonte Jr J, 82 Rao AR, Quinones MP, Garavito E, Kalkonde Y, Jimenez F, Crawford JM et al. An experimental model of idiopathic Gibbons C et al. CC chemokine receptor 2 expression in donor pneumonia syndrome after bone marrow transplantation: I. cells serves an essential role in graft-versus-host-disease. The roles of minor H antigens and endotoxin. Blood 1996; 88: J Immunol 2003; 171: 4875–4885. 3230–3239.

Supplementary Information accompanies the paper on Genes and Immunity website (http://www.nature.com/gene)

Genes and Immunity