GPR15^gfp/gfp mice from Dan Littman (4) were bred on a C57BL/6 background. Wild-type C57BL/6 mice were from JAX. Mice were maintained and bred in specific pathogen-free conditions in the animal facility at the Veterans Affairs Palo Alto Health Care Systems (VAPAHCS). Mice were used at 12–14 weeks of age, except for the neonatal thymocyte migration experiments in which day 0 pups were used. Animals were maintained in accordance to US National Institutes of Health guidelines, and experiments were approved by Stanford University Institutional Animal Care and Use Committee. Human peripheral blood mononuclear cells (PBMCs) were obtained from healthy donors. This study was carried out in accordance with the recommendations of the US National Institutes of Health guidelines, with written informed consent from all subjects. The protocol was approved by the Stanford University Institutional Review Board.

Thymic lymphocytes were isolated by removing the thymus and straining the whole organ first through a 100 µm mesh. Then, the first cell suspension was filtered through a 70 µm nylon mesh to obtain a single cell suspension. Caudal mesenteric lymph node (cMLN) cells were isolated from two of the colon-draining lymph nodes that are clearly separated from the chain of mesenteric lymph nodes that drain the small intestine (12). The lymph nodes were strained through a 100 µm mesh, and the cell suspension was filtered through a 70 µm nylon mesh to obtain a single cell suspension.

Human PBMCs were isolated from heparinized peripheral blood (10–40 ml) that was obtained via venipuncture from healthy donors. Blood samples were processed using Ficoll density gradient centrifugation (Histopaque-1077, Sigma-Aldrich). The resulting interface containing the PBMC layer was extracted, washed twice with PBS, and resuspended in the chemotaxis medium.

In all cases, after isolation cells were incubated for 2 h at 37°C/5% CO₂ in chemotaxis medium [RPMI-1640 supplemented with heat inactivated fetal bovine serum (FBS) 10% (v/v), 100 U/ml penicillin, 0.1 mg/ml streptomycin, 2 mM glutamine, 1× MEM non-essential amino acids, and 1 mM sodium pyruvate] for receptor expression recovery, before the migration assay. Where indicated, Pertussis toxin (Tocris, Bristol, UK) treatment (200 ng/ml) was performed during these 2 h of incubation.

CHO-K1 cells transfected for human GPR15 expression were from DiscoverX, Inc. (Fremont, CA, USA). We confirmed the expression and surface location of the human GPR15 in the cells by FACS using a PE-anti-human GPR15 antibody (Biolegend, San Diego, CA, USA).

In the standard-binding assays, GPR15 transfectants or normal CHO-K1 cells were incubated in binding buffer [HBSS Ca²⁺/Mg²⁺ with FBS 2% (v/v)] for 30 min with 0.1 nM of either human IgG Fc or a chimeric protein comprising the full-length human GPR15L linked to human Fc through the GPR15L N terminus (LakePharma Inc., Belmont, CA, USA).

In the “cold competition assay,” GPR15 transfectants were incubated for 30 min with several concentrations (100 pM to 10 µM) of the native full-length human GPR15L (Peprotech, Rocky Hill, NJ, USA) in binding buffer. Then the “tracer,” here human GPR15L linked to human Fc, was added to the cells at a concentration of 1 nM and incubated for 30 min.

The cells were washed in staining buffer (PBS BSA 0.5%), then stained with PE-Goat F(ab’)2 anti human IgG (Invitrogen), and acquired on a LSRII from BD, using FACS Diva Software.

CHO-K1 β-arrestin reporter cells for human GPR15 and human CMKLR1 from Discover X, Inc. (Fremont, CA, USA) were cultured normally in phenol-free RPMI-1640 supplemented with FBS 10% (v/v), 100 U/ml penicillin, 0.1 mg/ml streptomycin, 2 mM glutamine, MEM non-essential amino acids 1×, hygromycin 0.3 mg/ml, and G418 0.8 mg/ml. Briefly, cells were seeded the day before the assay in 96-well plates (Greiner bio-one, Ref:655090) at a density of 30,000 cells per well. On the day of the assay, cells were exposed for 90 min at 37°C/5% CO₂ to different peptides in solution in the regular culture medium, then incubated for 2 h at room temperature with Tropix^® Gal-screen™ reagent per manufacturer’s instructions, and finally, the luminescence signal measured with an integration time of 500 ms on a SpectraMax M5 (Molecular Devices, Sunnyvale, CA, USA) plate reader. The peptides used in some of the β-arrestin assays were produced by the Stanford Protein and Nucleic Acid Facility.

In vitro migration assays were conducted and data presented as described (13). 1–2 × 10⁶ lymphocytes were added to the top wells of 5 µm pore, polycarbonate 24w tissue culture inserts (Costar, Cambridge, MA, USA) in 100 µl, with 600 µl of chemokine solution (or medium) in the bottom chamber. All migrations were conducted in chemotaxis medium at 37°C/5% CO₂ for 3 h. The chemokines used were recombinant mouse and human CXCL12 (100 nM) (Gryphon Sciences, South San Francisco, CA, USA), and C10orf99 recombinant protein, Ref: NBP2-14716PEP (Novusbio, Littleton, CO, USA), here referred as human GPR15L-His. In preliminary data, we have also observed a comparable chemotactic activity for the human C10orf99 from Peprotech. The total volume of the bottom chamber of each well was harvested separately, and the migrated cells recovered were washed in staining buffer (PBS 0.5% BSA) prior to analysis by flow cytometry. Cell migration was assessed by analyzing the percentage of cells of each indicated phenotype that migrated from the top chamber to the bottom compartment for each individual insert (Figure 3). Specific migration was calculated as percentage of input cells of the indicated cell subset that migrated, minus the mean background migration of that subset in the no chemokine control group (Figures 4 and 5).

Original input and migrated cells were first blocked with staining buffer containing 10% human serum in the case of human PBMC or 1 µg/ml of anti CD16/CD32 (clone 2.4 G2, BD) for mouse cells. For intracellular Foxp3 staining, we used the transcription factor staining buffer set (00-5523-00, Thermo Scientific) following the manufacturer’s protocol. The following antibodies were used for staining of human cells: CD127-FITC (clone HIL-7R-M21, BD), Foxp3-PerCP-Cy5.5 (clone 236A/E7, BD), GPR15-PE (clone SA302A10, Biolegend), CD3-PE-Cy7 (clone SK7, BD), CLA-Biotin (559950, BD) with streptavidin-APC (BD) as secondary, CD25-APC-Cy7 (clone M-A251, BD), CD8a-Pacific Blue (clone RPA-T8, Biolegend), CD45R0-BV605 (clone UCHL1, Biolegend), and CD4-BV711 (clone SK3, BD). The following antibodies were used for staining mouse cells: CD24-PE (clone 30-F1, Biolegend), CD3e-PE-Cy7 (clone 145-2C11, BD), TCRVγ3-APC (clone 536, Biolegend), TCRγ/δ-BV421 (clone GL3, Biolegend), CD25-PerCp-Cy5.5 (clone PC61, BD), Foxp3-PE (clone MF23, BD), CD103-APC (clone M290, BD), CD44-AF700 (clone IM7, BD), CD8a-Pacific Blue (clone 56-3.7, eBioscience), CD45RB-BV650 (clone 16A, BD), CD4-BV711 (clone L3T4, Biolegend), and α4β7-APC (clone Act-1). Dead cells were excluded by DAPI, aqua dead cell staining, or fixable blue staining kit (Thermo Scientific) depending on the experiment. FACS data were acquired on a Fortessa from BD using FACS Diva software. Further analysis was performed using FlowJo from Treestar.

Total RNA from dissected segments of the intestines was isolated using the RNeasy Mini Kit (Qiagen, Redwood City, CA, USA), according to the manufacturer’s instructions. Quantification was determined by 260/280 nm absorbance ratio; 0.5 µg RNA was subjected to reverse transcription with iScript™ Reverse Transcription kit (Biorad, Hercules, CA, USA); Power SYBR^® Green PCR Master Mix (Applied Biosystems, Warrington, UK) was used for amplification, and specific DNA sequences were amplified with an Applied Biosystems 7900HT sequence detection system. No PCR products were detected when the mock-transcribed mixture was used as template. Results are expressed as 2^−ddCt using as housekeeping genes mouse HPRT1, RPL13a, and Actin. The primers used for these housekeeping genes were commercially available primer sets from RealTimePrimers.com. The sequences of the primers for mouse 2610528A11Rik amplification were: 5′-CAC CAC CCA TGA CTT GAC TG-3′ and 5′-CTT CTA GCC CTT TCC GGT CT-3′, purchased from Integrated DNA Technologies (Redwood City, CA, USA).

Results are expressed as mean ± SEM. Statistics were calculated using GraphPad Prism, where differences among means were tested for statistical significance by one-way or two-way ANOVA and a posteriori Fisher’s least significant difference tests on preselected pairs. Differences were considered significant at P < 0.05.

Our initial discovery efforts testing known gut- and skin-expressed chemoattractants and cytokines including CCL27, CCL28, CCL25, chemerin, and dermokine, failed to identify a binding partner for GPR15. We noted that the colon-expressed polypeptide C10orf99 was reported to trigger calcium signaling in GPR15 transfectants in a published patent application, although functional responses of lymphocytes were not evaluated (14). The C10orf99 gene was predicted to encode a polypeptide with basic pI, low molecular weight, and structural features similar to those of known chemokines. The exon structure, position of cysteine residues (including a “CC-motif”), and the C-terminal amino acid sequence are highly conserved across vertebrate classes from human to rodents (Figure 1), consistent with a conserved function.

We, therefore, tested the ability of the candidate GPR15L protein to bind GPR15. We linked GPR15L via its N-terminus to the Fc domain of human IgG. The human GPR15L-Fc chimeric protein binds huGPR15/CHO cells but not CHO controls (Figure 2A). GPR15L/GPR15 binding was inhibited by preincubation of GPR15-expressing cells with native human GPR15L, with an IC50 of 100 nM (Figure 2B). Using a β-arrestin assay to detect receptor activation, we showed that the native full length human GPR15L and the huGPR15L-Fc chimera trigger signaling with EC50’s of 20 and 40 nM, respectively (Figure 2C). Moreover, a 16 amino acid peptide corresponding to the conserved C-terminal sequence of mouse and human GPR15L was sufficient to trigger activation, albeit at higher concentrations (Figure 2D).

A subset of perinatal thymic CD3⁺ CD24⁻ TCRγ/δ⁺ TCRVγ3⁺ dendritic epidermal cell precursors (pDETC) expresses high levels of gpr15, and GPR15 mediates pDETC homing to the fetal epidermis (3). Using Transwell chemotaxis assays, we found that GPR15L selectively attracts pDETC, compared to their CD3⁺ CD24⁻ TCRγ/δ⁺ TCRVγ3⁻ GPR15⁻ thymic counterparts (Figures 3A,B). To determine if GPR15L increases random cell motility vs directed chemotaxis, we performed a “checkerboard” assay and found that the migration to the chemokine was predominantly chemotactic (Figure 3C). Migration was inhibited by pertussis toxin pretreatment of lymphocytes, indicating that the chemotactic effect of GPR15L is Gαi dependent (Figure 3D). Moreover, pDETC from Gpr15-deficient mice failed to migrate in response to GPR15L, confirming the GPR15-dependence of the chemotactic response (Figure 3E).

GPR15 mediates the colon localization of CD4⁺ CD25⁺ effector and regulatory cells to the colon and has been implicated in colitis in models of infectious (4) and autoimmune (5) intestinal inflammation. Colon homing T cells are generated in the colon-draining lymph nodes, including the cMLN (12). We isolated cMLN cells from dextran sodium sulfate (DSS) colitic mice (day 7) and assessed their migration to GPR15L. CD3⁺ CD4⁺ CD44⁺ CD45RB^low CD25⁺ CD103⁺ memory CD4 cells and CD3⁺ CD8⁺ CD44⁺ effector memory phenotype CD8 T cells in the cMLN migrate efficiently to GPR15L, in contrast to naïve CD4 or CD8 cells (Figures 4A,B, respectively). ~80% of the migrated memory CD4⁺ T cells, and 20–30% of the migrated memory phenotype CD8⁺ T cells were α4β7^hi (not shown). CD3⁺ CD4⁺ Treg, which are enriched in GPR15 expression in mice (4), also migrated to GPR15L (Figure 4C). Chemotaxis of cMLN T cells to GPR15L was GPR15/Gpr15 dependent, as T cells from Gpr15-deficient mice migrated normally to CXCL12, but did not respond to the new chemokine (Figures 4D,E).

GPR15 is expressed by subsets of memory, regulatory, and effector lymphocytes in human blood but not on naïve T cells (7, 15). We analyzed the chemotactic effect of GPR15L on GPR15⁺ subsets of CD8 and CD4 effector/memory, Treg, CLA⁺ skin-homing, and α4β7⁺ gut-homing T cells using an anti-GPR15 antibody. GPR15⁺ subsets migrated to GPR15L, while CD4 and CD8 naïve T cells, which lack the receptor, did not respond (Figures 5A–F). Migration decreased at high concentrations, as is typical of chemotactic responses (Figures 5A–F) (16). GPR15-dependent human T cell migration to GPR15L was pertussis toxin sensitive, indicating that GPR15 couples to Gαi in human as well as mouse lymphocytes (Figure 5G).

Previous studies have reported tissue-selective expression of C10orf99 and 2610528A11Rik, with highest gene and protein expression in the colon and stomach mucosa, and substantial expression in the oral cavity and genitourinary tract as well (17). We confirmed selective expression in human large intestine by analysis of independent data sets in the NCBI gene expression repository (17, 18) (Figure 6A). In addition, we have confirmed a comparable pattern of expression in the mouse, with selective and high expression of 2610528A11Rik/GPR15L in the distal segments of the gastrointestinal tract (cecum and colon) (Figure 6B). In these sites, GPR15L appears restricted to epithelial cells (Human Protein Atlas available from www.proteinatlas.org). The colonic expression of GPR15L is developmentally programmed, since the gene is already expressed in the caudal endoderm by day 9.5–13 of embryonic development (19).

As mentioned above, pDETC migrate from thymus to epidermis in a GPR15-dependent manner between day 16 and 18 of embryonic development (3). Consistent with this, by analyzing public data sets of the NCBI gene expression repository, we found that the 2610528A11Rik is highly expressed in keratinocytes of embryonic (day 16) and neonatal skin. Expression is highest in the spinous layer, which harbors recruited DETC (20) (Figure 6C). In contrast, GPR15L message is low or absent in adult skin in both mouse and human (Figures 6A,C); but it is significantly increased during wound healing, peaking 3 days after a burn injury (Figure 6D). Moreover, in humans, C10orf99 is highly upregulated in psoriasis (10, 11).

Analyzing public data sets of the NCBI gene expression repository, we have observed that colonic expression of 2610528A11Rik/GPR15L is not altered by intestinal microbiota (Figures 6E,F) (21, 22). It is only minimally affected by intestinal inflammation, since it is not significantly modulated in two classic experimental models of colitis like the DSS and the T naïve cell transfer model (Figures 6G,H) (23, 24). In line with this, in humans, C10orf99/GPR15L colonic expression is not significantly altered by inflammation in the context of ulcerative colitis (UC) (Figure 6I) (25).