A novel methodology for defining stromal expression of atypical chemokine receptors in vivo

Analysis of chemokine receptor, and atypical chemokine receptor, expression is frequently hampered by the lack of availability of high-quality antibodies and the species-specificity of those that are available. We have previously described methodology utilising Alexa-Fluor labelled chemokine ligands as versatile reagents to detect receptor expression. Previously this has been limited to haematopoietic cells and methodology for assessing expression of receptors on stromal cells has been lacking. Amongst chemokine receptors the ones most frequently expressed on stromal cells belong to the atypical chemokine receptor subfamily. These receptors do not signal in the classic sense in response to ligand but scavenge their ligands and degrade them and thus sculpt in vivo chemokine gradients. Here we demonstrate the ability to use either intratracheal, or intravenous, Alexa-Fluor labelled chemokine administration to detect stromal cell populations expressing the atypical chemokine receptor ACKR2. Using this methodology we demonstrate, for the first time, expression of ACKR2 on blood endothelial cells. This observation sets the lung aside from other tissues in which ACKR2 is exclusively expressed on lymphatic endothelial cells. In summary therefore we described a novel method for the in situ labelling of atypical chemokine receptor expressing cells appropriate for subsequent flow cytometric analysis. We propose that this methodology will work in a range of species and for a range of receptors and therefore will have significant versatility

Amongst chemokine receptors the ones most frequently expressed on stromal cells belong to the 23 atypical chemokine receptor subfamily. These receptors do not signal in the classic sense in response 24 to ligand but scavenge their ligands and degrade them and thus sculpt in vivo chemokine gradients. 25 Here we demonstrate the ability to use either intratracheal, or intravenous, Alexa-Fluor labelled 26 chemokine administration to detect stromal cell populations expressing the atypical chemokine 27 receptor ACKR2. Using this methodology we demonstrate, for the first time, expression of ACKR2 on 28 blood endothelial cells. This observation sets the lung aside from other tissues in which ACKR2 is 29 exclusively expressed on lymphatic endothelial cells. In summary therefore we described a novel 30 method for the in situ labelling of atypical chemokine receptor expressing cells appropriate for 31 subsequent flow cytometric analysis. We propose that this methodology will work in a range of species 32 and for a range of receptors and therefore will have significant versatility 33 34 35

Introduction. 36
In vivo leukocyte migration is regulated, in the main, by proteins belonging to the chemokine 37 family of chemotactic cytokines (1,2). This family is defined on the basis of a conserved cysteine motif 38 in the mature sequence of its members and is divided into CC, CXC, XC and CX3C subfamilies according 39 to the specific configuration of this motif. The chemokine family arose early in vertebrate evolution (3)  40 (pre-vertebrate species do not have chemokines) and the primordial chemokine was almost certainly 41 CXCL12 which plays essential roles in stem cell migration during embryogenesis(4-9). From this one 42 chemokine and its receptor CXCR4, through gene duplication, the family has expanded to the point at 43 which mammals have approximately 45 chemokines, and 18 signalling chemokine receptors, which 44 together orchestrate in vivo homeostatic and inflammatory leukocyte migration. Chemokine 45 regulation of cellular migration is extremely complex and, particularly in the case of inflammation(10), 46 poorly understood which has contributed to ongoing problems in therapeutically targeting 47 inflammatory chemokine receptors in immune and inflammatory diseases (11). 48 In addition to signalling chemokine receptors which belong to the G-protein coupled receptor 49 family(12), chemokines also bind to a subfamily of atypical chemokine receptors (ACKRs) which are 50 generally stromally-expressed and which fine-tune in vivo chemokine activity by scavenging 51 chemokines and therefore regulating chemokine availability(13-15). There are currently 4 members 52 of the atypical chemokine receptor family. ACKR1 (formerly known as DARC); ACKR2 (formerly known 53 as D6); ACKR3 (formerly known as CXCR7) and ACKR4 (formerly known as CCX-CKR). With the 54 exception of ACKR1, these receptors exhibit spontaneous internalisation and recycling activity and 55 scavenge chemokines from the environment and target them for lysosomal degradation. ACKR3 56 carries out this role in some essential developmental contexts and is strongly evolutionarily 57 conserved (5,6,16,17). ACKR4 scavenges chemokines within the lymph node to generate intra-lymph 58 node gradients and facilitate dendritic cell migration from the sub-capsular sinus into the T cell zone 59 of the lymph node(18). We have had a particular interest in ACKR2 which is the prototypic member of 60 the atypical chemokine receptor family(19). This receptor binds, internalises and degrades all 61

Materials and Methods 85
Mice 86 Animal experiments were performed using co-housed mice in ventilated cages in a barrier facility that 87 conformed to the animal care and welfare protocols approved by the University of Glasgow under the 88 revised Animal (Scientific Procedures) Act 1986 and the European Union Directive 2010/63/EU. Ackr2-89 deficient mice (Jamieson et al., 2005) were bred in-house (C57BL/6 background); wild type (WT) 90 C57BL6/J mice were from Charles River Research Models and Services. Prox-1 reporter mice were 91 obtained from Jackson laboratories. All experimental mice were sex and age matched. 92 93 qPCR 94 RNA was extracted using RNeasy columns with DNase treatment (Qiagen), and the amount of RNA 95 was quantified on a Nanodrop 1000 Spectrophotometer (Thermo Fisher Scientific). cDNA was 96 synthesized using High-capacity RNA-to-cDNA kit by Applied Biosystems (Thermofisher). For all qPCRs, 97 a final concentration of 0.2-mM primers was used for each PCR set up using PerfeCTa SYBR Green 98 FastMix and ROX qPCR Master Mix (Quanta BioSciences). qPCRs were performed on a Prism 7900HT 99 Fast Real-Time PCR System (Applied Biosystems). The thermal cycles for qPCR of TBP and ACKR2 were 100 95˚C (3 min) for one cycle and 95˚C (3 s) and 60˚C (30 s) for 40 cycles. Relative expression was 101 calculated using serial dilutions of cDNA standards. Primer sequences designed for qPCR and for 102 producing cDNA standards were designed using Primer3 software (http://frodo.wi.mit.edu/cgi-103 bin/primer3/primer3_www.cgi). The following primers were used: mouse ACKR2, 5'-104 TTCTCCCACTGCTGCTTCAC-3', 5'-TGCCATCTCAACATCACAGA-3'; mouse TBP primer: 5'-105 AAGGGAGAATCATGGACCAG-3', 5'-CCGTAAGGCATCATTGGACT-3'. 106 107

In-situ hybridisation 108
Mice were culled using increasing concentration of CO2. Lungs were placed in 10% neutral buffered 109 formalin at room temperature for 24-36 hours before they were processed by dehydration using rising 110 concentrations of ethanol, xylene stabilisation and paraffin embedding (Shandon citadel 1000, 111 Thermo Shandon). Tissue was then sectioned onto Superfrost plus slides (VWR) at 6μm using a 112 Microtome (Shandon Finesse 325 Microtome, Thermo). All slides for analysis were processed together. 113 Slides were baked at 60 o C for 1 hr before pre-treatment. Slides were deparaffinised with xylene (5 114 mins x 2) and dehydrated with ethanol (1 min x 2). In situ hybridisation was performed using the 115 RNAscope® 2.5 HD Reagent Kit-RED from Advanced Cell Diagnostics (cat. no. 322350) and according 116 to the manufacturer's instructions. Briefly tissues were incubated with Hydrogen peroxide for 10 mins 117 at RT. The slides were boiled in antigen retrieval buffer for 15 mins. Slides were treated with 'protease 118 plus' for 30 mins at 40 o C. Slides were then hybridised using the RNAScope 2.5 Red Manual Assay 119 (Advanced cell diagnostics) according to manufacturer's instructions using the Mm-ACKR2 probe 120 (NM_021609.4). Slides were mounted in DPX (Sigma Aldrich) and imaged on an EVOS M7000 121 microscope (Thermofisher). 122 123

Intratracheal and Intravenous Chemokine administration. 124
To administer fluorescent chemokine intratracheally mice were euthanised using an appropriate 125 schedule 1 method or CO2 exposure. The mice were then carefully dissected to remove the ribcage 126 and expose the intact lungs and trachea in situ. Using a pair of surgical scissors, a small incision was 127 made at the top of the exposed trachea toward the base of the jaw. A 2µg/ml solution of Alexa 647 TM 128 labelled CCL22 (Almac; Alexa-CCL22) dissolved in RPMI/25mM HEPES was prepared in a polypropolene 129 tube and preserved from light at room temperature until needed. Once the dissection was complete 130 a syringe with a 19G needle and loaded with 400µl of the Alexa-CCL22 solution was then inserted into 131 the exposed trachea via the incision. The needle should be tight within the trachea and care should 132 be taken not to pierce the trachea further down. The lungs were then inflated with the chemokine 133 solution and the trachea carefully tied off with surgical thread to prevent the leakage of chemokine 134 solution as the syringe is removed. The intact inflated lungs were then removed and placed into a 135 falcon tube containing enough RPMI to cover the intact inflated lungs. The lungs were then incubated 136 in a water bath for 1hour at 37 0 C. Following this time the lungs were removed from the waterbath 137 and the surgical thread cut to allow draining of the remaining chemokine solution. The lungs were 138 then digested for a single cell suspension as per the protocol. Data on ACKR2 expression profiles available through the Immgen database (www.immgen.org) 181 reveal that the lung is the tissue with the highest expression (Supplemental Figure 1). We have 182 previously described the versatile use of fluorescently labelled chemokines, instead of antibodies, to 183 detect their cognate receptors using flow cytometry as a read out(35-38). We used this approach with 184 Alexa-647 labelled CCL22 (a high affinity ACKR2 ligand: Alexa-CCL22) uptake to detect ACKR2 in lung 185 digests. Flow cytometric analysis of Alexa-CCL22 stained lung digests failed to detect any significant 186 expression on CD45+ leukocytes in either WT or ACKR2-/-lungs (Figure 1a). This therefore indicated 187 that ACKR2 expression in the lung was predominantly stromal in origin. We used RNA sequencing to 188 generate data on the transcriptional profile of FACS-sorted pulmonary stromal cell types at rest and 189 over the course of an influenza infection to examine possible pathogen-driven alterations in 190 expression. As shown in Figure 1b, ACKR2 expression was essentially undetectable in epithelial cells 191 but was present at low levels in fibroblasts and at very high levels in blood endothelial cells. Expression 192 did not vary significantly over the course of influenza infection. Examination of pulmonary ACKR2 193 expression by qPCR from embryonic day 13.5 to 9 weeks of age indicated that expression is low within 194 the embryo but that it is markedly upregulated immediately after birth and presumably coincident 195 with the onset of breathing. This increased level is maintained and increased as mice age (Figure 1c). 196 Next, we tried to use the in vitro Alexa-CCL22 detection method to examine ACKR2 expression 197 on non-leukocytic stromal cells in the digested lung. However, and as shown in Figure 1d Overall, therefore, these data indicate that ACKR2 is predominantly expressed on stromal cells 202 within the lung but that flow cytometry utilising fluorescent chemokine uptake with digested lung 203 tissue has limited sensitivity to detect the key stromal expressing cell types. 204 205 Intra-tracheal fluorescent chemokine administration identifies key ACKR2-expressing stromal 206

components. 207
We reasoned that the function of stromal ACKR2 expression may be dependent on 208 interactions with other stromal components and that the inability to detect it using flow cytometry 209 reflects the inability to take up Alexa-CCL22 due to absence of these interactions. We therefore 210 harvested intact lungs, inflated them with Alexa-CCL22 (in RPMI) intra-tracheally followed by 211 incubation at 37°C for 1 hour (Figure 2a Overall therefore these data demonstrate that it is possible to detect stromal cell populations 220 that bind and internalise Alexa-CCL22 via ACKR2 by introducing the chemokine intra-tracheally into 221 the intact lung. They also demonstrate novel stromal expression patterns for ACKR2 within the lung. 222

A subpopulation of Fibroblasts in the lung express ACKR2. 224
Cells identified in the R1 gate in Figure 2d, which were negative for markers of lymphatic and 225 vascular endothelial cells, were further phenotyped. Initially these cells were isolated by cell-sorting 226 and then grown in tissue culture. As shown in Figure 3a, the cells display a morphology suggestive of 227 a fibroblastic phenotype. Prox-1 is a definitive marker and an essential master regulator, of lymphatic endothelial cells(41). As 259 shown in Figure 4e, Prox-1+ve cells from reporter mouse lungs were exclusively co-positive for CD31 260 and Gp38 confirming the faithfulness of the lymphatic phenotype in the lung. These data further 261 confirm the blood endothelial nature of the ACKR2+ve stromal cells. 262 In the lung there are two major blood vascular beds: one associated with bronchial tissues 263 and one with alveolar tissues. These can be discriminated on the basis of CD54 and CD31 264 expression(42). As shown in Figure 4f the Alexa-CCL22 internalising blood endothelial cell population 265 is strongly co-positive for CD31 and CD54 demonstrating that this population is associated with 266 alveolar blood vessels and not peribronchial blood vessels. 267 268

In situ hybridisation and antibody expression confirm blood endothelial cell expression of ACKR2 269
To further validate ACKR2 expression by murine pulmonary vascular endothelial cells, we 270 carried out in situ hybridisation. As shown in Figure 5a reporter mouse-based approaches are available for analysis of atypical chemokine receptor 281 expression patterns, in many cases these are limited and applicable only to mouse and humans. Given 282 that many of the atypical chemokine receptors display strong evolutionary conservation, other more 283 versatile approaches would therefore represent a significant improvement in the methodological 284 repertoire for atypical chemokine receptor expression analysis. Here we demonstrate, using the lung, 285 that fluorescently-labelled chemokines can be used, with intact tissues, to precisely isolate and 286 phenotypically define stromal cell types expressing individual atypical chemokine receptors. This is 287 particularly important when, as in the current analyses, removal of the cells from their stromal 288 environment impairs ACKR2 function and thus frustrates this detection methodology. The ability to 289 chemically synthesise chemokines with relative ease and to introduce discrete fluorescent markers at 290 the carboxy terminus(39) means that this approach has full versatility and is appropriate for all 291 members of the chemokine receptor family and all species expressing either typical or atypical 292 chemokine receptors. Importantly, whilst we demonstrate the utility of this approach using WT and 293 KO mice, the use of appropriate unlabelled competing chemokines specific for the receptor being 294 studied will allow this technology to be used under circumstances, or in species, where KO models do 295 not exist. 296 Here we show that intra-tracheal, and intra-venous, administration of fluorescently labelled 297 CCL22 is capable of identifying stromal cell populations expressing ACKR2. Importantly, in the context 298 of vascular endothelial cell expression, intravenous administration has the advantage of detecting 299 chemokine receptor expression with a polarity favouring expression on the luminal side of the 300 endothelium. Together these approaches allowed us to define fibroblasts, lymphatic endothelial cells, 301 and surprisingly alveolar blood endothelial cells as key sites of stromal ACKR2 expression in the lung. 302 Surprisingly, and in contrast to previous reports, we did not detect ACKR2 activity on alveolar 303 macrophages (43,44). This may be due to species differences in expression or alternatively may be a 304 consequence of non-specific antibody internalisation by the alveolar macrophages. Further analysis is 305 required to address this discrepancy. 306 Importantly this technology is applicable to other atypical chemokine receptors as well as to 307 the classical chemokine receptors and to any tissue which can be incubated in vitro with fluorescent 308 chemokines. This approach would, for example, be readily amenable for use with intestinal tissue, 309 liver and skin. 310 In summary therefore we report a methodology appropriate for detecting atypical chemokine 311 receptor expression in the lung which we believe to be sufficiently versatile to be useful to detect