A novel chemokine receptor for SDF-1 and I-TAC involved in cell survival, cell adhesion, and tumor development

Abstract

The chemokine stromal cell-derived factor (SDF-1; also known as chemokine ligand 12 [CXCL12]) regulates many essential biological processes, including cardiac and neuronal development, stem cell motility, neovascularization, angiogenesis, apoptosis, and tumorigenesis. It is generally believed that SDF-1 mediates these many disparate processes via a single cell surface receptor known as chemokine receptor 4 (CXCR4). This paper characterizes an alternate receptor, CXCR7, which binds with high affinity to SDF-1 and to a second chemokine, interferon-inducible T cell alpha chemoattractant (I-TAC; also known as CXCL11). Membrane-associated CXCR7 is expressed on many tumor cell lines, on activated endothelial cells, and on fetal liver cells, but on few other cell types. Unlike many other chemokine receptors, ligand activation of CXCR7 does not cause Ca2+ mobilization or cell migration. However, expression of CXCR7 provides cells with a growth and survival advantage and increased adhesion properties. Consistent with a role for CXCR7 in cell survival and adhesion, a specific, high affinity small molecule antagonist to CXCR7 impedes in vivo tumor growth in animal models, validating this new receptor as a target for development of novel cancer therapeutics.

Figure 1.

Evidence for a novel SDF-1 binding protein. (A) ¹²⁵I SDF-1 binds to E13 fetal liver from wild-type (CXCR4 ^+/+), heterozygous (CXCR4 ^+/−), and homozygous (CXCR4 ^−/−) mouse embryos. In each case, binding specificity is demonstrated by the ability of 100 nM of unlabeled SDF-1α to effectively compete radiolabeled ligand binding to fetal liver calls. Data represent the means of three determinations ± SD. (B) FACS analysis of CXCR4 expression with the anti-CXCR4 mAb 12G5 (left), assessment of radiolabeled ¹²⁵I SDF-1 binding (middle), and 10 nM SDF-1–induced calcium mobilization (right) using MCF-7 cells (top) or CEM-NKr cells (bottom). FACS histogram indicates isotype control (open) and specific staining (shaded). Arrows indicate ligand (SDF-1α) addition to cells. This set of analyses has been reproduced in at least 10 independent studies, and error bars represent SEM. (C) SDF-1 binding site on MCF-7 cells is pharmacologically distinct from CXCR4. ¹²⁵I SDF-1 binding to CEM-NKr and MCF-7 cells is performed in the presence of SDF-1α, CXCR7 antagonist CCX451, or CXCR4 antagonist AMD3100. Error bars represent SEM.

Figure 2.

The novel SDF-1/I-TAC binding protein is RDC1. (A)Binding of ¹²⁵I SDF-1 or ¹²⁵I I-TAC to MCF-7 cells was effectively competed using either unlabeled SDF-1 or unlabeled I-TAC. Error bars represent SEM. (B) Binding of ¹²⁵I SDF-1 to either human breast tumor MDA MB 435s wild-type cells (top) or MDA MB 435s transfected with CXCR7 (bottom) in the presence of increasing concentrations of SDF-1α, I-TAC, CXCR7 antagonist CCX451, and CXCR4 antagonist AMD3100. Total counts bound for each condition are shown and represent the mean of four determinations ± SD. RDC1 Homo sapiens is available from GenBank/EMBL/DDBJ under accession no. P25106. (C) Mouse mAb to human CXCR7 (11G8) binds MDA MB 435s cells transfected with CXCR7 (right) but not MDA MB 435s wild-type cells (left). FACS histogram indicates isotype control (open) and specific staining (shaded). (D) Mouse anti–human CXCR7 mAbs 11G8 and 6E10 but not irrelevant isotype control, 11H3, largely inhibit binding of radiolabeled SDF-1 (top) or I-TAC (bottom) to CXCR7 transfectants. Error bars represent SEM. Data shown in (C) and (D) have been obtained in at least five experiments, and the percent inhibition relative to untreated controls (total bound) is shown.

Figure 3.

Expression of CXCR7 by transformed human and mouse cell lines. (A) Membrane CXCR7 and CXCR4 expression on mouse B cell lymphoma, BCL1, and human cervical carcinoma HeLa cells is demonstrated by anti-CXCR7 antibody 11G8 staining and anti-CXCR4 antibody 12G5. Staining is represented as shaded versus open for isotype control. (B) ¹²⁵I SDF-1 binding profile reveals membrane CXCR7 expression on BCL1 and HeLa cell surfaces. CXCR7 binding is defined by inhibition of the ¹²⁵I SDF-1 binding with 100 nM of nonradiolabeled ligand or 10 μM CXCR7 antagonist CCX451, but not by 10 μM CXCR4 antagonist AMD3100. Error bars represent SEM. (C) Northern blot analysis of mRNA expression of CXCR7 and CXCR4 in a panel of human transformed cell lines.

Figure 4.

Nontransformed human and mouse tissues express little membrane CXCR7 but frequently express CXCR7 by Northern blot analysis. (A) Binding of ¹²⁵I SDF-1 to various tissue cells obtained from the mouse. Error bars represent SEM (B) Anti-CXCR7 antibody 11G8 fails to bind mouse blood, liver, lung, or heart cells. FACS histogram indicates isotype control (open) and specific staining (shaded). (C) Northern blot analysis shows CXCR7-specific mRNA in normal mouse tissues, but this does not correlate with cell surface protein expression.

Figure 5.

CXCR7 is expressed in early mouse embryonic development. (A) Binding of ¹²⁵I SDF-1 to embryonic fetal liver (E11–17) in the presence of increasing concentrations of SDF-1α, I-TAC, the CXCR7 antagonist CCX451, or the CXCR4 antagonist AMD3100. Error bars represent SEM. (B) A subset of E13 mouse fetal liver cells stain positive using anti-CXCR7 antibody 11G8, as indicated by the boxed areas. (C) Northern blot analysis shows expression of CXCR7 mRNA in E11 and E13 but not E15 and E17 fetal liver. Total RNA from CXCR7-positive MCF7 cells and CXCR7-negative CEM-NKr cells are included as controls. Data shown were obtained in at least five experiments.

Figure 6.

Introduction of CXCR7 into human breast tumor cell line MDA MB435s confers a growth advantage to these cells. In all experiments, wild-type or CXCR7-transfected MB435s cells were cultured in vitro in media containing a suboptimal concentration of serum (1% instead of the standard 10%). (A) Cell counts of live, dead, and total wild-type or CXCR7-transfected cells cultured over time. (B) Summary of the day 4 time point from growth experiments in histogram form to emphasize the fact that the wild-type and CXCR7-transfected cells have similar total cell numbers but are distinguished by how many of these total cells are dead. (C) Apoptotic cells in these cultures were identified (see oval) as cells excluding propridium iodide and that stained with the Annexin V marker for apoptotic cells. (D) CCX754 inhibits CXCR7-mediated growth advantage in a dose-dependant manner. Wild-type or CXCR7-transfected cells were incubated with various concentrations of CXCR7 compound throughout assay.

Figure 7.

CXCR7 mediates cell adhesion in vitro. Adherence is measured (A) by capturing a brightfield image to visualize both the HUVEC monolayer and the labeled adherent cells (bright circles) and (B) by fluorescence to quantitate adhesion. Error bars represent SEM. (C) Activated endothelium expresses CXCR7 by radiolabeled binding assay. HUVECs were stimulated with TNF-α and IL-1β (HUVEC act.) or sham-treated in the presence or absence of cycloheximide (CHX) and incubated with radiolabeled SDF-1. Unlabeled SDF-1α and I-TAC (100 nM each), as well as CCX451 and AMD3100 (10 μM each), were examined for the ability to compete with ¹²⁵I SDF-1 binding. (D) Northern blot analysis of CXCR7-specific mRNA expressed by unstimulated (HUVEC) or TNF-α/IL-1β (HUVEC act.) cells cultured in the presence or absence of Cycloheximide (CHX). MCF7 (native expression) and CEM-NKr (transfectant) CXCR7 transcripts are shown as controls. Data for A–C were reproduced in at least five experiments. Data for D were obtained twice.

Figure 8.

Pharmacological properties of CXCR7 antagonist (CCX754) used in in vivo experiments. The experiments are shown in Fig. 9. (A) Pharmacokinetic properties of CXCR7 antagonist in mice show a serum half-life (T1/2) of 6 h, bioavailability of 20% (F), and an acceptable liver clearance rate of 30 ml/min/kg (CL). These pharmacokinetic properties are compatible with once a day dosing in mouse animal models. (B) CCX754 inhibits binding of ¹²⁵I SDF-1 to MCF-7 human breast tumor cells with an IC₅₀ of 5 nM. Error bars represent SEM in both panels.

Figure 9.

In vivo efficacy of CXCR7 antagonist. Efficacy of CCX754 was evaluated in syngeneic or xenograft mouse models engrafted with (A) human B lymphoma IM9 cells, (C) human A549 lung carcinoma cells, or (D) mouse LL/2 Lewis lung carcinoma cells. Control groups receiving vehicle alone were included in all experiments. Horizontal bars in C indicate means, and error bars in D represent SEM. Images in A show peritoneal cavity of mice bearing IM9 tumor cells and receiving either CXCR7 antagonist (right) or vehicle alone (left). (B) Immunohistochemical analysis of CXCR7 expression on a human biopsy sample of a malignant lung carcinoma. n = 8 in all groups. Statistical differences between treatment groups were calculated using survival curve statistics and the Student's t test using GraphPad Prism software.