A Fluorescent Activatable AND‐Gate Chemokine CCL2 Enables In Vivo Detection of Metastasis‐Associated Macrophages

Abstract We report the novel chemical design of fluorescent activatable chemokines as highly specific functional probes for imaging subpopulations of immune cells in live tumours. Activatable chemokines behave as AND‐gates since they emit only after receptor binding and intracellular activation, showing enhanced selectivity over existing agents. We have applied this strategy to produce mCCL2‐MAF as the first probe for in vivo detection of metastasis‐associated macrophages in a preclinical model of lung metastasis. This strategy will accelerate the preparation of new chemokine‐based probes for imaging immune cell function in tumours.

Tumour metastasis is the leading cause of cancer-related death. Themetastatic potential of tumours is largely defined by the surrounding immune environment, where macrophages are among the most abundant cells. [1] Different subpopulations of macrophages are found in tumours,s uch as resident macrophages and active metastasis-associated macrophages,w ith the latter being recruited by chemokines (e.g.,C CL2 or C-C motif chemokine ligand 2) to accelerate the progression of metastasis. [2] Whereas novel therapeutic strategies to stop the recruitment of tumour-associated macrophages-including chemokine inhibitors-are under evaluation as anticancer treatments, [3] there are no chemical probes for detecting these subpopulations of macrophages in live tumours.
Themost widespread technology to track the mobilisation of macrophages are genetically-encoded fluorescent proteins (e.g.,G FP) expressed under macrophage-selective promoters. [4] Fori nstance,E ntenberg et al. recently used transgenic mice to study the interactions between macrophages and cancer cells in vivo. [5] However,t his technique cannot distinguish macrophage subtypes or provide cell activity readouts.A lternatively,s mall-molecule fluorophores have been developed to monitor the function of macrophages. [6] Our group and others have shown that fluorophores responding to different biomarkers (e.g. acidic pH, [7] reactive oxygen species, [8] MMPs, [9] cathepsins, [10] and SLC transporters [11] )can be used to detect active macrophages under physiological conditions.T hese probes provide generic readouts of macrophage activity but are not specific for the metastasisassociated macrophages linked to cancer progression.
CCL2 is apotent chemokine that regulates the migration of immune cells to tumours through the recognition of CCR2 receptors.R ecent studies have identified that metastasisassociated macrophages express high levels of CCR2 on the cell surface. [2,12] Binding of CCL2 to CCR2 promotes the recruitment of macrophages into metastatic sites,w hich accelerates the seeding and expansion of cancer cells.G iven that chemokines can enter cells via their functionally-active receptors,w eh ypothesized that fluorescently-labelled CCL2 analogues would allow us to detect CCR2 + metastasisassociated macrophages in tumours.P reviously,N ibbs et al. described CCL2 chemokines to broadly identify subpopulations of mouse leukocytes. [13] Theg roups of Beck-Sickinger and Liu and Qi also demonstrated that chemically photocaged chemokines can be used to temporally control leukocyte migration. [14] Herein, we have designed as trategy to prepare AND-gate [15] activatable chemokines that selectively target CCR2 + populations of metastatic macrophages in vivo.
We designed fluorescent activatable chemokines to work in at wo-step sequence ( Figure 1a). First, they recognise functional receptors that internalise upon ligand binding (step 1) AND second, they emit fluorescent signals in response to macrophage-related activity (step 2). Compared to fluorescently-labelled antibodies and always-on fluorescent chemokines ( Figure 2), this AND-gate strategy makes possible the recognition of both functional receptors as well as intracellular activity,e nabling the detection of highly specific cell populations.I nt his work, we have applied this concept to generate the first fluorescent activatable CCL2 chemokine for live imaging of active metastasis-associated macrophages.
In order to conjugate the chemokine CCL2 to fluorescent markers of active macrophages,w ef irst examined macrophage-activatable fluorophores (MAFs) responding to intracellular metabolites associated with macrophage activity,such as phagosomal pH, superoxide and reactive oxygen species. Forthis,weincubated inactive and LPS-activated RAW264.7 mouse macrophages with different MAFs and measured their emission by fluorescence microscopy ( Figure S1 in the Supporting Information). We observed that pH-sensitive BODIPY dyes triggered by phagosomal acidification showed bright intracellular fluorescence in active macrophages with the largest turn-on emission among the different MAFs.T herefore,w ed ecided to utilise pH-sensitive BOD-IPYs for the generation of activatable CCL2 chemokines.
We synthesized the pH-dependent BODIPY (5)containing am aleimide group for site-specific conjugation to the mouse chemokine CCL2 (mCCL2, Figure 1). Thek ey intermediate 4 was obtained in 3s teps from pyrrole 1 in good yields.Asreported by our group and others, [16] the assembling of the BODIPY fluorophore proceeded via condensation of two functionalized pyrroles (1 and 2,F igure 1) followed by addition of BF 3 OEt 2 to generate the nitro-derivatised BODIPY 3.Herein, we optimised the catalytic hydrogenation and subsequent derivatisation with chloroacetyl chloride and diethylamine to yield BODIPY 4 with an overall yield of 87 % including all 3steps.Finally,compound 4 was hydrolysed and coupled to 1-(2-aminoethyl)maleimide to generate BODIPY 5 for site-specific protein bioconjugation (see Supporting Information for synthetic and characterisation details).
We envisaged that the incorporation of fluorophores at the C-terminus of mCCL2 would not disrupt the recognition of chemokine receptors or impair its chemotactic activity. Therefore,weemployed an analogue of mCCL2 with an extra cysteine residue at the C-terminal end that would react sitespecifically with the BODIPY maleimide 5.The fluorophorechemokine conjugation was performed with an excess of compound 5 (4 equiv) in aqueous buffer at pH 6.5 for 1h,and the resulting mCCL2-MAF (6,F igure 1b)w as purified by HPLC.Weused mass spectrometry to confirm single coupling of the fluorophore to the C-terminal cysteineo fm CCL2 in a1:1 ratio ( Figure S2).
Next, we evaluated the spectral properties of mCCL2-MAF (6)t oc orroborate that the turn-on properties of the fluorophore remained unaffected after conjugation to mCCL2. Compound 6 showed similar photophysical properties to the parent BODIPY fluorophore 4 (l exc. :496 nm, l em. : 510 nm;F igure S3) with pH-dependent activation and ap K a around 4.9, suitable to detect active macrophages specifically undergoing phagosomal acidification ( Figures S4 and S5). This result suggests that the incorporation of alkyl spacers at the position 3ofthe BODIPY core is an effective strategy to generate bioconjugatable fluorophores with full retention of their optical properties.N ext, we performed cell migration transwell assays to examine whether mCCL2-MAF (6) retained the biological function of the native chemokine mCCL2. We cultured RAW264.7 mouse macrophages on top of cell-permeable membranes,incubated them with different concentrations of mCCL2 or mCCL2-MAF (6), and counted the migrated cells across the membrane.A ss hown in Figures 3a nd S6, mCCL2-MAF (6)p romoted macrophage migration to comparable levels of unlabelled mCCL2. These results confirm that the C-terminal modification of mCCL2 with BODIPY dyes endows the chemokine with fluorescent reporters of macrophage function without affecting the chemotactic properties of mCCL2.
Given the functional and fluorescent activatable properties of mCCL2-MAF (6), we then investigated its applicability as am arker for active CCR2 + metastasis-associated macrophages.First, we compared the emission of mCCL2-MAF (6) in inactive and LPS-activated CCR2 + mouse macrophages by confocal microscopy and flow cytometry.W ef ound that inactive macrophages showed weak fluorescence emission, which was dramatically increased in activated cells due to LPS-induced phagosomal acidification (Figures 4a,b). Importantly,t he fluorescence enhancement was reduced to basal levels when macrophages were pre-treated with bafilomycin A, which inhibits LPS-mediated phagosomal acidification, or anti-CCR2 antibodies,w hich block ligand binding to the  receptor and subsequent endocytosis (Figure 4c). Altogether these results confirm that the fluorescence emission of mCCL2-MAF (6)f ollows an activation mechanism with two sequential events:1 )C CR2-mediated internalisation, and 2) fluorescence amplification in active macrophages.
In contrast to mCCL2-MAF (6), AlexaFluor647-mCCL2 (always-on chemokine) emitted strong fluorescence in CCR2 + mouse macrophages regardless of their activation state (Figure 4d), which highlights the advantages of activatable chemokines as fluorescent reporters of intracellular activity.
Encouraged by these results,w en ext examined whether mCCL2-MAF (6)c ould selectively stain CCR2 + macrophages in the presence of other immune cells.F or these experiments,w eu sed wild-type C57/BL6 as well as Ccr2 knock-out (KO) mice so that we could determine the selectivity profile of compound 6 in cell populations expressing variable levels of CCR2 receptors.Weprepared single-cell suspensions from spleens of wild-type and Ccr2 KO mice, incubated them with mCCL2-MAF (6)(100 nm)and analysed them by flow cytometry.W eemployed mouse antibodies (F4/ 80, CD45, CD11b,Ly6G,CD3 and CD19) to identify different immune cells, [17] which included macrophages,neutrophils,B cells and Tc ells ( Figure S7 for gating strategy). From this analysis,w ef ound that as ubpopulation of wild-type macrophages showed bright fluorescence whereas this cell population was not detected in macrophages from Ccr2 KO mice (Figure 5a), indicating that the signals from mCCL2-MAF (6) are CCR2-dependent. We also found that mCCL2-MAF (6) did not stain neutrophils,Tcells or Bcells in either wild-type of KO mice (Figure 5b). These results highlight the value of mCCL2-MAF (6)f or the selective detection of CCR2 + macrophages with minimal off-target fluorescence in other cells.
Finally,w ee xamined whether mCCL2-MAF (6)c ould detect metastasis-associatedmacrophages in vivo.T othis end, we used am ouse model of lung metastasis where E0771-LG mouse mammary tumour cells are injected into the tail vein of syngeneic C57BL/6 mice to mimic the systemic distribution of cancer cells.I nt his model, metastasis-associated macrophages (MAMs) and their progenitor cells (MAMPCs) accumulate in the metastatic lungs via CCL2-CCR2 signaling. [17,18] Since the accumulation of MAMs and MAMPCs promotes metastasis,t hese cells are regarded as attractive targets in anticancer therapy. [17][18][19] Animals injected with E0771-LG cancer cells formed metastatic tumours in their lungs after 2weeks,asconfirmed by whole-body bioluminescence imaging ( Figure S8). At this point, we injected mCCL2-MAF (6)via the tail vein (10 ng in 50 mLsaline) and harvested the metastatic lungs after 2hours.Histological examination of the lungs corroborated the presence of metastatic nodules (Figure 6a). Notably,inthe metastatic nodules,wefound F4/ 80 + macrophages showing green fluorescence (Figure 6b), which indicates that the in vivo injection of mCCL2-MAF (6)  labels ad efined subpopulation of macrophages in metastatic tumours.W ealso analysed the tissues by flow cytometry and measured the fluorescence signals in major immune cells found in metastatic lungs ( Figure S9 for gating strategy and immune cell markers). As shown in Figure 6c,wefound high levels of green fluorescence in active MAMPCs and MAMs that require CCR2 signaling for their accumulation in metastatic sites.R emarkably,a ll other immune cells in the metastatic lungs [for example,r esident macrophages (RMAC), neutrophils,Tcells,Bcells,n atural killer (NK cells)] remained unstained by compound 6 (Figure 6c).
These results confirmed that the AND-gate activation mechanism of mCCL2-MAF (6)w as recapitulated in vivo. Specifically,wedid not observe staining in RMAC, which do not express CCR2, or in NK cells,which express high levels of CCR2 but cannot trigger the emission of macrophageactivated fluorophores (Figure 6c). In contrast, the alwayson chemokine AF647-mCCL2 showed bright fluorescence in NK cells from metastatic tumours under the same experimental conditions ( Figure S10). This observation confirms that the combination of functional chemokines with activatable fluorophores can generate highly specific probes for immune cells that cannot be targeted with conventional reagents,and validates mCCL2-MAF (6)asthe first chemical probe for selective detection and imaging of active metastasisassociated macrophages in tumours in vivo.
In summary,w eh ave designed AND-gate activatable chemokines as fluorescent agents for imaging active subpopulations of immune cells in live tumours.W eh ave synthesized mCCL2-MAF as the first chemical probe for metastasisassociated macrophages by coupling aC ys-modified chemokine CCL2 with at hiol-reactive macrophage-activatable fluorophore.m CCL2-MAF fully retains chemotactic activity and emits bright intracellular fluorescence in aCCR2-AND macrophage activity-dependent manner. We also demonstrated the utility of our strategy by showing selective staining of metastasis-associated macrophages in vivo,outperforming always-on fluorescent chemokines.This chemical strategy will accelerate the design of highly specific chemokine-based imaging agents to study immune cell subpopulations with enhanced spatial and temporal resolution.