https://orcid.org/0000-0002-8556-8433
Abstract
The G protein coupled receptor GPR15 is expressed on and functionally important for T cells homing to the large intestine. However, the precise mechanisms by which GPR15 controls gut homing have been unclear. Thus, we aimed to elucidate these mechanisms as well as to explore the potential of targeting GPR15 for interfering with T cell recruitment to the colon in inflammatory bowel disease [IBD].
We used dynamic adhesion and transmigration assays, as well as a humanised in vivo model of intestinal cell trafficking, to study GPR15-dependent effects on gut homing. Moreover, we analysed GPR15 and integrin expression in patients with and without IBD, cross-sectionally and longitudinally.
GPR15 controlled T cell adhesion to MAdCAM-1 and VCAM-1 upstream of α4β7 and α4β1 integrin, respectively. Consistently, high co-expression of these integrins with GPR15 was found on T cells from patients with IBD, and GPR15 also promoted T cell recruitment to the colon in humanised mice. Anti-GPR15 antibodies effectively blocked T cell gut homing in vitro and in vivo. In vitro data, as well as observations in a cohort of patients treated with vedolizumab, suggest that this might be more effective than inhibiting α4β7.
GPR15 seems to have a broad, but organ-selective, impact on T cell trafficking and is therefore a promising target for future therapy of IBD. Further studies are needed.
1. Introduction
Targeting the trafficking of immune cells to the inflamed intestine has emerged as an important pillar in the therapy of inflammatory bowel diseases [IBD] such as ulcerative colitis [UC] and Crohn’s disease [CD].1 In particular, the anti-α4β7 integrin antibody vedolizumab2,3 and the sphingosine-1 phosphate receptor agonist ozanimod4 have been approved, and several further compounds5,6 are still in the pipeline.
However, these approaches are not sufficient to induce remission in a substantial portion of patients.7 Vedolizumab has been shown to inhibit the interaction of α4β7 integrin on the surface of immune cells with mucosal addressin cell adhesion molecule [MAdCAM-]1 on the intestinal endothelium during the multi-step adhesion process of gut homing, and is therefore considered a very safe and gut-specific therapeutic.8 Yet, evidence suggests that other integrins may functionally compensate for such blockade and thereby limit the effect of the antibody.9 Ozanimod, on the other hand, leads to the sequestration of naïve and central memory T cells in secondary lymphoid organ, and thereby unspecifically depletes these cells from the circulation, which might also lead to reduced surveillance of other organs.10 Thus, further insights into the regulation of immune cell recruitment to the gut are required to derive novel approaches for therapeutic interventions that combine improved efficacy with high safety and gut selectivity.
The G protein coupled receptor 15 [GPR15] is expressed on gut-homing T cells and has been reported to play a functional role for T cell homing to the large intestinal lamina propria in mice in 2013.11 Further studies indicated that it has a similar function in humans, where it is predominantly expressed on effector T cells.12 Following the de-orphanisation of GPR15 with the description of its ligand GPR15L in 2017,13,14 chemotaxis depending on GPR15L-GPR15 interactions has been described,13 but with rather small effect size. The exact mechanism of GPR15 in the regulation of gut homing, however, is unclear to date.
Here, we hypothesised that since the polypeptide sequence of GPR15 is related to chemokine receptors,15 it might play a role in the ‘activation’ step of lymphocyte homing.16 According to the classical paradigm of lymphocyte homing, the interactions of selectins with their ligands, or of integrins in an inactive conformation with cell adhesion molecules [CAMs], lead to loose ‘tethering’ and ‘rolling’ of the cells along the vessel wall. This allows tissue-derived chemokines to bind to their cognate receptors on the T cell surface, which results in ‘activation’, including conformational changes in integrin heterodimers leading to high affinity for the respective CAMs. This serves as the prerequisite for subsequent integrin-dependent ‘firm adhesion’ and, finally, ‘transmigration’ to the tissue.17–19
Indeed, we show that GPR15 mediates T cell adhesion and in vivo homing upstream of integrins such as α4β7 and α4β1. Consistently, anti-GPR15 antibodies inhibit integrin-dependent homing, and data from a cohort of IBD patients treated with vedolizumab suggest that GPR15-expressing T cells may evade α4β7 blockade. Our findings therefore strongly support the concept that targeting GPR15 might be a promising option for broad, yet gut-specific, inhibition of intestinal T cell homing.
2. Materials and Methods
2.1. Patients
Peripheral blood samples used in this study were obtained following informed written consent of the donors and according to approval of the Ethics Committee of the Friedrich-Alexander-Universität Erlangen-Nürnberg. Patients with Crohn’s disease and ulcerative colitis were included. Blood samples from healthy donors served as control. The clinical data of the patients donating material for this study are summarised in Supplementary Table 1 and Supplementary Table 2.
2.2. T cell isolation
Peripheral blood mononuclear cells [PBMCs] were isolated with density gradient centrifugation using Ficoll [anprotec]. PBMC pellets were suspended and CD4+ T cells and memory CD4+ T cells were purified with the CD4+ T cell isolation kit or memory CD4+ T cell isolation kit, respectively, using LS columns and a MACS Separator [all Miltenyi Biotec].
2.3. Dynamic adhesion assays to MAdCAM-1 and/or VCAM-1
Rectangular glass capillaries [0,20 mm × 2,0 mm inner diameter, CM Scientific] were coated with 5 µg/ml of Fc chimera of the cell adhesion molecules MAdCAM-1 or VCAM-1 [R&D Systems] in coating buffer [150 mM NaCl + 1 mM HEPES] for 1 h at 37°C. Capillaries were blocked for 1 h with 20 µL of blocking solution [1 × PBS with 5% BSA] at 37°C and subsequently fixed on rubber tubings [Ismatec]. Purified T cells were stained with CellTrace CFSE [Invitrogen] for 15 min at 37°C. Cells were incubated with or without 10 µg/ml vedolizumab [Takeda], 10 µg/ml natalizumab [Biogen], 0.5 µg/ml anti-GPR15 antibody [R&D Systems, clone 367902], and/or 0.2 µg/ml pertussis toxin [List Biological Laboratories] as indicated in RPMI 1640 cell medium [supplemented with 1% penicillin/streptomycin] for 1 h at 37°C/5% CO2 and at a cell density of 1.5 × 106 cells/ml. For the adhesion assays, the cells were suspended in adhesion buffer [150 mM NaCl + 1 mM HEPES + 1 mM MgCl2 + 1 mM CaCl2] at a concentration of 1.5 × 106 cells/ml. Where indicated, cells were then treated with 100 nM recombinant GPR15L [Bachem or Peprotech] for 1 min at room temperature, or with 350 nM GPR15L for 15 min at 37°C, or left untreated. Cells were subsequently perfused through the capillaries at a flow rate of 10 µl/min for 3 min with a peristaltic pump [Baoding Shenchen]. Next, capillaries were rinsed for 5 min at a speed of 50 µL/min to remove non-adherent cells. The number of adhering cells was evaluated with an inverted microscope [Leica DMi8, Leica] by selecting eight representative pictures at 20x magnification, and the adhering cells were counted. To better visualise the results, pictures were merged and digitally coloured using ImageJ [NIH].
2.4. Transmigration assays
Transmigration assays with human CD4+ memory T cells were performed as previously described.20 5 µg/ml MAdCAM-1 or VCAM-1 were coated on the upper wells of a 3-µm transmembrane plate [Corning] for 1 h at 37°C. Isolated T cells were resuspended in X-Vivo medium [Lonza] to a concentration of 2 × 106 cells/ml. Then, 350 nM GPR15L with or without 10 µg/ml vedolizumab or 0.5 µg/ml CCL25 [Immunotools] were added to the cells directly prior to loading them into MAdCAM-1 coated wells [1.6 × 105 cells per well], whereas 350 nM GPR15L with or without 10 µg/ml natalizumab or 0.5 µg/ml CXCL10 [Immunotools] were added to the cells directly prior to loading them into VCAM-1 coated wells [1.6 × 105 cells per well]. The lower wells contained X-Vivo medium with 10% fetal calf serum [FCS]. Plates were incubated for 4 h at 37°C, and the number of transmigrated cells in the lower wells was analysed by flow cytometry.
2.5. Flow cytometry
Flow cytometry was performed according to standard protocols using the following fluorochrome-conjugated extracellular antibodies: CD3 [APC/Cy7, 145-2C11 or HIT3a, FITC, OKT3, both BioLegend, Vio Blue, Rea613, Miltenyi Biotec], integrin α4/CD49d [PE/Cy7, 9F10, BioLegend or FITC, MZ18-24A9, Miltenyi Biotec], integrin β7 [PE or PerCP/Cy5.5, FIB27, BioLegend], integrin β1/CD29 [PerCP/Cy5.5 or PE or AF647, TS2/16, BioLegend], GPR15 [APC or PE, SA302A10, BioLegend, AF700, #367902, R&D Systems], CD4 [VioBlue or VioGreen, VIT4, Miltenyi Biotec]. Intracellular staining was performed following fixation and permeabilisation with the Fix/Perm kit [Thermo Fisher] with antibodies against Foxp3 [PE, 236A/E7, Thermo Fischer Scientific], IL-13 [APC, JES10-5A2, BioLegend], IL-17 [BV605, BL168, BioLegend] and IFN-γ [PE/Cyanine7, 4S.B3, BioLegend]. For co-expression analysis, staining with anti-GPR15 antibody [mouse anti-human-GPR15, unconjugated, #367902, R&D Systems] and secondary detection with anti-mouse-IgG2B antibody [APC, #332723, R&D Systems] was performed before using further staining antibodies. Where indicated, MAdCAM-1 [rh Fc Chimera Protein, R&D Systems] was labelled using Alexa Fluor Antibody Labelling Kits [AF674/AF488, Life Technologies] according to the manufacturer’s instructions. VCAM-1 [rh Fc Chimera Protein, BioLegend] was detected by secondary staining with Alexa Fluor 647 anti-human IgG Fc Recombinant Antibody [BioLegend].
For the quantification of GPR15-dependent binding of MAdCAM-1 or VCAM-1 to cell surface integrins, RPMI 8866 [ECACC] or HuT 78 [Merck] human lymphocyte cell lines, respectively, were incubated with 350 nM GPR15L for 15 min at 37°C and subsequently stained with 5 µg/ml labelled MAdCAM-1 or VCAM-1 for 15 min at 37°C.
For the characterisation of GPR15 expression on CD4+ T cell subsets in UC and CD, peripheral blood mononuclear cells were isolated and stimulated with anti-CD3/CD28 antibodies for 48 h. Subsequently, they were treated with 50 ng/mL PMA [Sigma], 10 ng/µL Brefeldin A [Applichem], and 1 µM ionomycin [Cayman] for 4 h and the expression of cytokines was assessed by flow cytometry.To address a potential influence of GPR15L on integrin expression, CD4+ T cells isolated from the peripheral blood of healthy donors were stimulated with anti-CD3/28 antibodies [T Cell Activation & Expansion Kit, Miltenyi Biotec] in 48-well plates at a density of 1.0 million cells/mL for 72 h. Where indicated, cells were treated with 150 nM recombinant GPR15L. Subsequently, cells were collected and analysed by flow cytometry.
Data were acquired on LSR Fortessa [BD Bioscience], MACSQuant 10, and MACSQuant 16 [Miltenyi Biotec] instruments. Data were analysed using FlowJo™ v10.8 Software [BD Life Sciences].
2.6. Mice
B6.129S7-Rag1 < tm1Mom>/J [Rag1-/-] knockout mice were housed in individually ventilated cages with a regular day-night cycle, and used for experiments according to approval of the Government of Lower Franconia in compliance with all relevant ethical regulations.
2.7. In vivo gut homing experiments and intravital microscopy
In vivo gut homing assays were performed as previously described.21 Briefly, Rag1-/- mice were given 1.5% dextran sodium sulphate [DSS] in drinking water for 7 days to induce colitis. Purified peripheral blood CD4+ T cells from patients with IBD were stained with CellTrace FarRed for flow cytometry or CellTrace Yellow [both Invitrogen] for in vivo microscopy. Then, 2–6 × 106 fluorescently labelled cells were injected into the ileocolic artery of the Rag1-/- recipient mice together with vedolizumab and/or anti-GPR15 antibodies, and left to circulate for 1 h.
For in vivo microscopy, 10 µg of anti-CD31 antibodies [AF647, MEC13.3, BioLegend] and 50 µl Hoechst 33342 [ThermoFisher Scientific] were additionally injected together with cells in a total volume of 100 µl into the ileocolic artery. After 15 min, the proximal part of the colon was opened longitudinally and placed onto a coverslip for intravital microscopy with a confocal SP8 microscope [Leica].
Lamina propria mononuclear cells [LPMCs] were isolated using the Lamina Propria Dissociation Kit [Miltenyi Biotec] according to the manufacturer’s instructions, followed by density gradient centrifugation with Percoll [GE Healthcare], and used for flow cytometry to quantify CellTrace+ cells in the lamina propria.
2.8. Re-analysis of microarray data
In order to compare GPR15 expression before and under vedolizumab treatment, the microarray dataset with the accession number GSE73661 was downloaded from the Gene Expression Omnibus using the Python v 3.11.5 package GEOparse v 2.0.3.22 The data were quantile normalised and the normalised intensity levels of GPR15 at Week 0 versus Week 6 were compared.
2.9. Statistics
Statistical analyses were performed with GraphPad Prism [GraphPad Software]. Graphs show mean and standard error of the mean [SEM]. Unless otherwise stated, the following statistical tests were applied: Two-tailed Student’s t test was used if two groups were compared; If more than two groups were compared, one-way analysis of variance [ANOVA] with Tukey, Newman–Keuls or Bonferroni post hoc comparison was performed. Grouped analyses were carried out with two-way ANOVA with Bonferroni’s post hoc comparison; p-values below 0.05 were considered significant. Levels of significance are indicated by asterisks [*p <0.05, **p <0.01, ***p <0.001, ****p <0.0001].
3. Results
3.1. GPR15 modulates adhesion and transmigration upstream of α4β7 and α4β1 integrin
To investigate whether GPR15 promotes gut homing via integrin-dependent adhesion of T cells to the respective CAMs, we used a dynamic adhesion assay in which T cells are perfused through CAM-coated capillaries and their adhesion is studied by fluorescence microscopy.23
As expected, we observed adhesion of CD4+ T cells from the peripheral blood to MAdCAM-1 [Figure 1A]. When these T cells were pre-treated with the GPR15 ligand GPR15L in vitro, the adhesion to MAdCAM-1 was substantially increased. Additional treatment with pertussis toxin, which inhibits the Gi subunit of G protein-coupled receptors, abrogated this effect indicating that this is indeed due to signalling via GPR15. Strikingly, treatment of the T cells with vedolizumab in addition to GPR15L even further decreased dynamic adhesion to MAdCAM-1, confirming that α4β7 integrin is required downstream of GPR15 to mediate firm adhesion.
![GPR15L modulates integrin-dependent adhesion and transmigration of CD4+ T cells in vitro. [A] Left panels: Representative pictures from dynamic adhesion assays with labelled CD4+ T cells treated as indicated and perfused through MAdCAM-1- or VCAM-1-coated capillaries. Scale bars—50 µm. Right panels: Statistical analysis [n = 10 for MAdCAM-1, n = 6 for VCAM-1]. [B] Migration of CD4+ memory T cells from the peripheral blood and treated as indicated over porous membranes coated with or without MAdCAM-1 or VCAM-1. Representative [upper panels] and quantitative [lower panels] flow cytometry of transmigrated cells. CCL25, CC chemokine ligand 25; CXCL10, C-X-C motif chemokine ligand 10; MAdCAM-1, mucosal addressin cell adhesion molecule-1; NTZ, natalizumab; PTX, pertussis toxin, VCAM-1, vascular cellular adhesion molecule-1; VDZ, vedolizumab. *p <0.05; **p <0.01; ***p <0.001 as determined by one-way-ANOVA with Bonferroni post hoc test.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ecco-jcc/18/7/10.1093_ecco-jcc_jjae012/1/m_jjae012_fig1.jpeg?Expires=1750213648&Signature=TZlKrXhqbM83jD-nwRk2JH37Irs3TjPpbl0I4jUYhGjB8tO0-DGnu6e8mKyl8RpJlXAq39bfaSXXki3L94PU3E4THoThLO2UAoIj1m3EsXyG1XGfwp3tLYblTH57f~SY~vqYQze8jADILvwmfSwjrZl5QlMINJl3QDs9q6nLomP9qSfWXF2Bv0cI2TH2pkyf06iSU6DZueLACii4GjAR--p9Hw3Muhzh6uc7OeCrrgZ6YbYGOk7sixBsLYizUgVGXImX4BV3vgf-RWweASgD0AMhONd3UhBo6Sv3Kzb9-WLw6tmqBAmMwsmvz72SRP4kF13ubnWBqXf6jTlZymAdgA__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
GPR15L modulates integrin-dependent adhesion and transmigration of CD4+ T cells in vitro. [A] Left panels: Representative pictures from dynamic adhesion assays with labelled CD4+ T cells treated as indicated and perfused through MAdCAM-1- or VCAM-1-coated capillaries. Scale bars—50 µm. Right panels: Statistical analysis [n = 10 for MAdCAM-1, n = 6 for VCAM-1]. [B] Migration of CD4+ memory T cells from the peripheral blood and treated as indicated over porous membranes coated with or without MAdCAM-1 or VCAM-1. Representative [upper panels] and quantitative [lower panels] flow cytometry of transmigrated cells. CCL25, CC chemokine ligand 25; CXCL10, C-X-C motif chemokine ligand 10; MAdCAM-1, mucosal addressin cell adhesion molecule-1; NTZ, natalizumab; PTX, pertussis toxin, VCAM-1, vascular cellular adhesion molecule-1; VDZ, vedolizumab. *p <0.05; **p <0.01; ***p <0.001 as determined by one-way-ANOVA with Bonferroni post hoc test.
We subsequently performed similar assays with capillaries coated with VCAM-1 and made very similar observations [Figure 1B]: GPR15L treatment clearly increased the adhesion of peripheral blood CD4+ T cells, which was blocked by additional pertussis toxin. Blockade of α4 integrins with natalizumab in addition to GPR15L reduced adhesion to uncoated control levels.
Although we applied GPR15L for 15 min only in these assays and effects on protein expression were therefore very unlikely, we additionally excluded that GPR15L induces the expression of α4β7 and/or α4β1 integrin on peripheral blood CD4 + T cells treated with or without recombinant GPR15L [Supplementary Figure 1].
To interrogate our findings in a second model, we coated porous membranes with or without MAdCAM-1 or VCAM-1 and explored the transmigration of peripheral blood memory CD4+ T cells. Although we did not observe any specific effects when using uncoated membranes, GPR15L significantly increased T cell transmigration over MAdCAM-1- or VCAM-1-coated membranes [Figure 1C]. Similar to our previous results, vedolizumab or natalizumab, respectively, blocked these effects. Very interestingly, CCL25 and CXCL10, which have previously been shown to induce an active conformation of the integrins α4β7 and α4β1,24 respectively, led to very similar effects on transmigration as GPR15L did.
Together, these data promoted the concept that signalling downstream of GPR15 supports integrin-dependent adhesion to cell adhesion molecules expressed on the gut endothelium.
3.2. GPR15 is co-expressed with α4 integrin on T cells in patients with IBD
To investigate how relevant such a mechanism might be in patients with IBD, we profiled the co-expression of GPR15 with α4β7 and α4β1 integrins on circulating CD4+ T cells from healthy controls and patients with CD or UC by flow cytometry. GPR15+ T cells expressed substantially higher levels of α4β7 as well as α4β1, regardless of the entity studied [Figure 2A, Supplementary Figure S2A]. Vice versa, the expression of GPR15 on α4 integrin-expressing T cells was also clearly higher than on α4 integrin-negative T cells [Figure 2B]. Thus, these data make it plausible that GPR15-derived signalling contributes to integrin-dependent gut homing of individual cells.
![Substantial co-expression of GPR15 with integrins α4β7 and α4β1 on circulating CD4+ T cells. [A] Representative [left panels] and quantitative [right panels] flow cytometry of integrin α4β1 [upper panels] and α4β7 [lower panels] expression on GPR15+ and GPR15- CD4+ T cells from healthy donors [CON], or patients with ulcerative colitis [UC], or Crohn’s disease [CD], as indicated. [B] Quantitative flow cytometry of GPR15 expression on α4+β1+ vs. α4-β1- and on α4+β7+ vs. α4-β7- CD4+ T cells from the peripheral blood, n = 15–18 per group. [C, D] Quantitative flow cytometry of GPR15 expression on Foxp3+ and Foxp3- CD4+ T cells [C] and on IFN-γ+, IL-13+ and IL-17+ CD4+ T cells [D] from healthy donors [CON] or patients with ulcerative colitis [UC] or Crohn’s disease [CD]; n = 10 per group; *p <0.05; **p <0.01; ***p <0.001 as determined by two-tailed Student’s t test [A, C] or one-way ANOVA with Tukey post hoc test [B, D].](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ecco-jcc/18/7/10.1093_ecco-jcc_jjae012/1/m_jjae012_fig2.jpeg?Expires=1750213648&Signature=l9Ar7Jd3A~mgEEaHaVWiGlElUeK1uV7PJH21y5-HS9KgJyrmWZeLZeGwJhAvQ194LHHnA1sx7jI8Un094GpE8s9dhugd1g4dh-jbpEHSU3ewR-0v2sUngRSBkwc6rsgg5A9coXHTKFKBZucLQEjo3zMFgN6EKaeKCLP1UWjcg34tiyFfk5U01e5ThtaiF52eOx1uM40NXn61Y7LYYHihPDb1M0cwZcoGnvoGro2C9cJvnlbin31IGccP9-DkWlPjCrs7QZVrvPQ6LYojS2HgAVHnLniuLo14R2Fuf0gTLc8hzMt1jdLf-ptrz0OKN2TcgVXJgzCTsvoDG501tl6AYQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Substantial co-expression of GPR15 with integrins α4β7 and α4β1 on circulating CD4+ T cells. [A] Representative [left panels] and quantitative [right panels] flow cytometry of integrin α4β1 [upper panels] and α4β7 [lower panels] expression on GPR15+ and GPR15- CD4+ T cells from healthy donors [CON], or patients with ulcerative colitis [UC], or Crohn’s disease [CD], as indicated. [B] Quantitative flow cytometry of GPR15 expression on α4+β1+ vs. α4-β1- and on α4+β7+ vs. α4-β7- CD4+ T cells from the peripheral blood, n = 15–18 per group. [C, D] Quantitative flow cytometry of GPR15 expression on Foxp3+ and Foxp3- CD4+ T cells [C] and on IFN-γ+, IL-13+ and IL-17+ CD4+ T cells [D] from healthy donors [CON] or patients with ulcerative colitis [UC] or Crohn’s disease [CD]; n = 10 per group; *p <0.05; **p <0.01; ***p <0.001 as determined by two-tailed Student’s t test [A, C] or one-way ANOVA with Tukey post hoc test [B, D].
It has previously been suggested that GPR15 is predominantly expressed on effector T cells in humans in contrast to mice.12 To validate this notion, we determined the expression of GPR15 on stimulated CD4+ T cells from the peripheral blood [Supplementary Figure S2B]. The expression of GPR15 was clearly higher on Foxp3- compared with Foxp3+ cells, suggesting that regulatory T cells express lower levels of GPR15 [Figure 2C]. Within the Foxp3- compartment, higher expression was found on IL13+ and also on IFN-γ+ compared with IL17+ T cells [Figure 2D], suggesting that TH2 cells are particularly enriched for GPR15 expression. Overall, these data were well in line with previous literature and supported the view that GPR15 is mainly restricted to the effector cell compartment in humans.
3.3. GPR15 promotes integrin-dependent homing to the inflamed gut in vivo
Signaling downstream of other chemokine receptors, such as CXCR3 or CCR9, promotes the adhesion of T cells to endothelial ligands by driving integrins to switch from an inactive conformation that only loosely interacts with CAMs to an active conformation that mediates firm adhesion.24
Although our functional observations on GPR15-mediated adhesion presented above strongly suggested that a similar mechanism applies for GPR15L-induced activation of GPR15, we aimed to generate further evidence supporting this assumption. To this end, we labelled recombinant MAdCAM-1 and VCAM-1 with fluorescent dyes and used them to incubate integrin+/GPR15+ lymphoid cell lines either pre-exposed to GPR15L or not. We subsequently performed extensive washing steps to avoid any signals due to interaction with the inactive integrin confirmation. On analysis by flow cytometry, we found a clear increase of the T cell fraction that stained positive for MAdCAM-1 and VCAM-1 after GPR15L treatment [Figure 3A], indicating that GPR15L-GPR15 interactions lead to increased presence of integrins in an active conformation on the cell surface.
![GPR15 licenses integrin-dependent gut homing in vivo. [A] Representative and quantitative flow cytometry of labelled MAdCAM-1 binding to RPMI 8866 cells [upper panels] or labelled VCAM-1 to HuT 78 cells [lower panels], respectively, after treatment with or without GPR15L. Treatment with MnCl2 or CXCL10 is shown as positive control; n = 5-7 per group. [B] In vivo gut homing of fluorescently labelled human CD4+ T cells treated with or without GPR15L and/or vedolizumab after injection into the ileocolic artery of Rag1-/- with DSS colitis. Upper panels: representative intravital confocal microscopy. Lower panels: representative and quantitative flow cytometry of stained human cells among the mouse lamina propria cells; n = 6–9 per group. DSS, dextran sodium sulphate; VDZ, vedolizumab; *p <0.05; **p <0.01 as determined by two-tailed Student’s t test [A] or one-way ANOVA with Tukey post hoc test [B].](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ecco-jcc/18/7/10.1093_ecco-jcc_jjae012/1/m_jjae012_fig3.jpeg?Expires=1750213648&Signature=rYS1YwFbWTZu~bdnb1qHTBt8CrHNW4op9GiT0teZRaww54gqbxqCZNVUyvRTwO253tXV1TE8ORNgItxIzLAqjXBZd3iGnjuQ421u3tOnVNsqxEIoCRLUb8JNNba2pvlWN-tvqyNUGmHbdocGJO~bSNHQ19hP0OPV4fw1GYhuT9adL0LjHz6ngVfWYwDB7JvA192cY-huvEe7XY9M2akiO9mRE2xnBmhCp6ImF62OaVBwaQvHeruVWOarvw4eikqdlq-BdrbWZykiWHLTtNuiDM0GbSklopvW8fy80vUDrPWSuwEdYVlwQOp9YLrJe4DOwYb7bd6KntHnAZGjH663vw__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
GPR15 licenses integrin-dependent gut homing in vivo. [A] Representative and quantitative flow cytometry of labelled MAdCAM-1 binding to RPMI 8866 cells [upper panels] or labelled VCAM-1 to HuT 78 cells [lower panels], respectively, after treatment with or without GPR15L. Treatment with MnCl2 or CXCL10 is shown as positive control; n = 5-7 per group. [B] In vivo gut homing of fluorescently labelled human CD4+ T cells treated with or without GPR15L and/or vedolizumab after injection into the ileocolic artery of Rag1-/- with DSS colitis. Upper panels: representative intravital confocal microscopy. Lower panels: representative and quantitative flow cytometry of stained human cells among the mouse lamina propria cells; n = 6–9 per group. DSS, dextran sodium sulphate; VDZ, vedolizumab; *p <0.05; **p <0.01 as determined by two-tailed Student’s t test [A] or one-way ANOVA with Tukey post hoc test [B].
We now wondered whether the observed GPR15-driven effects on integrin confirmation, adhesion, and transmigration also mediate T cell homing to the inflamed gut in vivo. Hence, we used a previously established humanised in vivo model21 and adoptively transferred fluorescently labelled CD4+ T cells from the peripheral blood to Rag1-/- mice with mild dextran sodium sulphate [DSS] colitis. These T cells were either treated with GPR15L, or vedolizumab alone, or in combination, or left untreated. We subsequently assessed gut homing by intravital confocal microscopy and flow cytometry, which showed that—similar to the in vitro findings—GPR15L substantially increased T cell recruitment to the gut, whereas vedolizumab abrogated this effect [Figure 3B].
3.4. Anti-GPR15 antibodies efficiently block the recruitment of T cells to the gut
Having thus shown that GPR15 promotes integrin-dependent gut homing in vivo, we aimed to explore the feasibility of targeting GPR15 to impede T cell recruitment to the intestine.
Accordingly, we used anti-GPR15 antibodies in dynamic adhesion assays with peripheral blood CD4+ T cells on MAdCAM-1 or VCAM-1. Consistent with our previous observations, treatment of the CD4+ T cells with GPR15L clearly increased dynamic adhesion. Strikingly, additional treatment with anti-GPR15 completely suppressed this effect, and anti-GPR15 alone even decreased adhesion below baseline levels [Figure 4A].
![GPR15 blockade impedes gut homing to the inflamed colon in vivo. [A] Representative [left panels] and quantitative [right panels] dynamic adhesion of peripheral blood CD4+ T cells from patients with IBD, treated with or without GPR15L and/or anti-GPR15 in capillaries coated with MAdCAM-1 or VCAM-1 as indicated. Scale bars—50 µm; n = 5–6 per group. [B] In vivo gut homing of fluorescently labelled human CD4+ T cells after treatment with or without GPR15L or anti-GPR15 and injection into the ileocolic artery of Rag1-/- mice with DSS colitis. Upper panels: representative intravital confocal microscopy. Lower panels: representative and quantitative flow cytometry of stained human cells among the mouse lamina propria cells; n = 4 per group; *p <0.05; ***p <0.001 as determined by one-way ANOVA with Tukey post hoc test. IBD, inflammatory bowel disease; DSS, dextran sodium sulphate.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ecco-jcc/18/7/10.1093_ecco-jcc_jjae012/1/m_jjae012_fig4.jpeg?Expires=1750213648&Signature=Lw0g60S~HM-YPcz93K-RdVTaPgQx08olEWA896YKRRz2vAfqB5~D6tBIdfXQS2-uT27lLjY0cDFGjKXjfQpDHECKkXRulN83aQQrZax0NoRufxb7tyltQ50WJyOb62CAIldjW6wzpoxPC3lU0EdPErjjc2B~VQlmfVecHw18qPbrBS1kT~h-k27lM7RpxHRw2b9hNs5mBbsZzuXXXLw3G-GxtMxCHt1Ohkwc9TpLwl85SR9omeurx34RIeV84J9PB16n3wY5q88hA4QFy9qOx1dVToJFZVI-7mnNXzWWyYZyMCpI0LvYtdtgCJ2qaE28XYqpmAC0Q5lQwA7DiX8ZSA__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
GPR15 blockade impedes gut homing to the inflamed colon in vivo. [A] Representative [left panels] and quantitative [right panels] dynamic adhesion of peripheral blood CD4+ T cells from patients with IBD, treated with or without GPR15L and/or anti-GPR15 in capillaries coated with MAdCAM-1 or VCAM-1 as indicated. Scale bars—50 µm; n = 5–6 per group. [B] In vivo gut homing of fluorescently labelled human CD4+ T cells after treatment with or without GPR15L or anti-GPR15 and injection into the ileocolic artery of Rag1-/- mice with DSS colitis. Upper panels: representative intravital confocal microscopy. Lower panels: representative and quantitative flow cytometry of stained human cells among the mouse lamina propria cells; n = 4 per group; *p <0.05; ***p <0.001 as determined by one-way ANOVA with Tukey post hoc test. IBD, inflammatory bowel disease; DSS, dextran sodium sulphate.
Therefore, in a next step, we wondered whether anti-GPR15 also counteracts gut homing in vivo. Taking advantage of the humanised mouse model, we transferred peripheral blood CD4+ T cells treated with GPR15L, and/or anti-GPR15L, or left untreated. In consistency with the in vitro data, the increase of T cell accumulation in the inflamed colon induced by GPR15L was reversed by additional anti-GPR15 [Figure 4B], suggesting that GPR15 is a suitable target to interfere with large intestinal T cell recruitment.
3.5. Reduction of circulating GPR15+ but not GPR15- T cells expressing α4β7 under therapy with vedolizumab
Finally, we aimed to obtain insights into the role of GPR15-driven integrin-dependent gut homing in patients with IBD in vivo. To this end, we profiled α4β7 integrin and GPR15 expression on T cells in a cohort of patients with IBD25 initiating therapy with the anti-α4β7 antibody vedolizumab at baseline and after 6 weeks of treatment [Figure 5A].
![Persistence of GPR15L-dependent gut homing via α4β1/VCAM-1 under α4β7 blockade with vedolizumab. [A] Schematic depiction of longitudinal assessment of GPR15-expressing gut-homing T cells in the peripheral blood of patients with IBD undergoing therapy with vedolizumab. Generated with licensed Biorender application. [B] Representative gating strategy for circulating CD4+α4+β7+GPR15+/- T cells [upper and left panels]. Quantification of CD4+α4+β7+GPR15+ and CD4+α4+β7+GPR15- T cells at T0 compared with T2; n = 25. [C] GPR15 RNA expression in mucosal biopsies obtained at Week 0 and Week 6 of vedolizumab treatment from patients with UC as determined by microarray (GSE73661). [D, E] Representative [left panels] and quantitative [right panels] dynamic adhesion assays with peripheral blood CD4+ T cells from patients with IBD in capillaries co-coated with MAdCAM-1 and VCAM-1. Cells in [E] were treated with GPR15L, cells in [D] not. Additionally, cells were treated with natalizumab, vedolizumab, and/or anti-GPR15 as indicated. Scale bars—50 µm; n = 10 per group. NTZ, natalizumab; VDZ, vedolizumab; IBD, inflammatory bowel disease; UC, ulceratice colitis; ns—not significant; *p <0.05; **p <0.01; ***p <0.001; ****p <0.0001 as determined by two-tailed Student’s t test [B, C] or one-way ANOVA with Tukey post hoc test.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ecco-jcc/18/7/10.1093_ecco-jcc_jjae012/1/m_jjae012_fig5.jpeg?Expires=1750213648&Signature=cBGEhb4~yDTzl1y0-Ojfo5~qTFpvU1bIrxj5gyMrQLkaIJ-JSYTtKDxHfhR4rHqTc1MeKWR7IdlpaamcrRKCTmoUi~~PHT6b6saejzWFpgN~Ls7~~1cVgAT9V~8oB-z5BMWF22nPWHYTcVlhQLLGuvrgEKQZl3SwXByOMZSdI9Appr1DqJcBC1F0RexGnvs2qna0sBh~oZvaoSE-Rz4YoamVmfCOCplDBylNtQZbg09vejaZSGWp62vLaL2lKr9l-Ckz1MV5j9D3eh6DguyuEeVM6ilPIqgjUbcGB8ZLnaTgDR9h~SQ7r1~39dtjAsSfvaUOtboYxEHL7V5fVIVY5A__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Persistence of GPR15L-dependent gut homing via α4β1/VCAM-1 under α4β7 blockade with vedolizumab. [A] Schematic depiction of longitudinal assessment of GPR15-expressing gut-homing T cells in the peripheral blood of patients with IBD undergoing therapy with vedolizumab. Generated with licensed Biorender application. [B] Representative gating strategy for circulating CD4+α4+β7+GPR15+/- T cells [upper and left panels]. Quantification of CD4+α4+β7+GPR15+ and CD4+α4+β7+GPR15- T cells at T0 compared with T2; n = 25. [C] GPR15 RNA expression in mucosal biopsies obtained at Week 0 and Week 6 of vedolizumab treatment from patients with UC as determined by microarray (GSE73661). [D, E] Representative [left panels] and quantitative [right panels] dynamic adhesion assays with peripheral blood CD4+ T cells from patients with IBD in capillaries co-coated with MAdCAM-1 and VCAM-1. Cells in [E] were treated with GPR15L, cells in [D] not. Additionally, cells were treated with natalizumab, vedolizumab, and/or anti-GPR15 as indicated. Scale bars—50 µm; n = 10 per group. NTZ, natalizumab; VDZ, vedolizumab; IBD, inflammatory bowel disease; UC, ulceratice colitis; ns—not significant; *p <0.05; **p <0.01; ***p <0.001; ****p <0.0001 as determined by two-tailed Student’s t test [B, C] or one-way ANOVA with Tukey post hoc test.
Interestingly, the percentage of α4β7-expressing CD4+ T cells negative for GPR15 remained virtually stable under therapy, indicating effective blockade of their gut homing [Figure 5B]. To the contrary, there was a clear reduction of CD4+ T cells co-expressing GPR15 and α4β7 integrin during the first 6 weeks of treatment, which could be interpreted by preserved recruitment to the intestine. Consistently, the re-analysis of publicly available microarray data of gut tissue from a prospective patient cohort (GSE73661), where samples were obtained at baseline and after 6 weeks of vedolizumab treatment, showed no differences in the expression of GPR15.
As we had shown above that GPR15 also promotes T cell adhesion to VCAM-1 via α4β1 integrin, VCAM-1 is expressed on the gut endothelium during intestinal inflammation,26 and flow cytometry of peripheral blood CD4+ T cells showed high co-expression of α4β7 and α4β1 on GPR15+ cells [Supplementary Figure S3], we hypothesised that this finding might display the GPR15-driven gut homing via alternative homing pathways such as α4β1. Indeed, further dynamic adhesion assays with T cells in capillaries co-coated with MAdCAM-1 and VCAM-1 showed that vedolizumab only partially blocked GPR15L-induced adhesion, whereas the pan-α4 integrin antibody, natalizumab, much more effectively interfered with T cell adhesion [Figure 5D, E]. This indicated that GPR15L-dependent residual T cell homing via α4β1 and VCAM-1 occurs under α4β7 blockade by vedolizumab. Interestingly, anti-GPR15 had around the same effect size as natalizumab, suggesting that—similar to non-organ-specific targeting of α4 integrin—the adhesion of α4β1 and α4β7 integrin can be targeted via GPR15 in a probably more gut-specific way.
4. Discussion
Already since its description as a molecule expressed on gut-homing T cells, GPR15 has been discussed as a potential therapeutic target.11,12,19 However, detailed mechanistic data substantiating the assumption that targeting GPR15 indeed interferes with gut homing pathways were still scarce. Here, for the first time, we show that anti-GPR15 antibodies impede the gut homing procedure in vitro and in vivo, which strongly supports the status of GPR15 as a potential therapeutic candidate for colonic IBD. Moreover, we reveal the mechanism responsible for GPR15-dependent gut homing by demonstrating that GPR15 acts upstream of integrins and thereby controls integrin-dependent adhesion and transmigration.
A similar function has previously been described for other GPCRs, namely CCR9 and CXCR3, which are the receptors for CCL25 and CXCL10. Indeed, signalling pathways downstream of these receptors selectively controlled the integrin-dependent adhesion of α4β7 integrin to MAdCAM-1 and VCAM-1, respectively.24 Later on, detailed structural investigations demonstrated that, to induce these effects, both chemokines differentially control the switch of α4β7 to an active conformation, which is required for firm ligation of MAdCAM-1.27 For GPR15, so far, only GPR15L-directed chemotaxis of small subsets of GPR15+ T cells had been described,13 whereas other studies had not been able to demonstrate clear chemotactic effects.14 However, since GPR15 shows structural similarities to chemokine receptors as does GPR15L to chemokines, we found it worth investigating whether GPR15L-GPR15 interactions might also play a role for integrin-dependent adhesion. Our data suggest that GPR15 promotes both α4β1- and α4β7-dependent adhesion to VCAM-1 and MAdCAM-1, respectively. Further studies are warranted to characterise the responsible intracellular signaling pathways and to elucidate whether this role also extends to other integrins such as αLβ2, as well as to investigate the obvious hypothesis that increased GPR15-driven adhesion might also be due to conformational changes.
The above-mentioned role of CCL25-CCR9 and CXCL10-CXCR3 interactions for gut homing28 also materialised in the observation that interfering with these pathways mitigates experimental colitis,29,30 and therefore triggered subsequent trials of the CCR9 small molecule antagonist vercirnon in CD and the anti-CXCL10 antibody eldelumab in UC and CD. Disappointingly, both strategies failed.31–33 These findings have led to the impression that chemokine signalling pathways are difficult to target, and might limit the enthusiasm for a further development of strategies interfering with GPR15. However, in our eyes, several aspects support the notion that despite these previous setbacks in targeting chemokines, GPR15 is worth following up:
A clear preferential expression of the GPR15 ligand GPR15L in the colon has previously been described.13,14 While some expression has also been noted in the small intestine, stomach, tonsils, cervix, skin, and bladder, no relevant expression has been detected in numerous other organs such as brain, liver, lungs, kidneys, or heart. This means that blocking GPR15 will probably only have impact on T cell homing to tissues where GPR15L is expressed, but not to tissues where GPR15L expression absent and a physiological role of GPR15 for T cell recruitment is therefore missing. This clearly differentiates GPR15 from other targets such as α4β1 integrin, which has a key function for T cell recruitment to the central nervous system, and might be important to forecast the safety profile of a potential anti-GPR15 treatment strategy. In this regard, it is also important to note that GPR15 seems to be predominantly expressed on effector T cell subsets in humans,12,34,35 although this was not confirmed in all studies.36 Thus, targeting GPR15 might also preferentially impact on pro-inflammatory T cells, while preserving regulatory T cell homing. A similar mechanism has previously been demonstrated to contribute to the effect of vedolizumab.20
Finally, when trying to forecast the efficacy of an anti-GPR15 strategy, it seems important to consider our finding that GPR15 is involved in both α4β7- and α4β1-dependent pathways. Since it has been established that alternative homing pathways may compensate for the blockade of one integrin,9,37 interfering with GPR15 as a shared upstream signal might lead to tighter blockade of gut homing. Indeed, our observations in patients treated with vedolizumab supported the notion that GPR15 may actually drive alternative gut homing in the presence of α4β7 blockade. Moreover, anti-GPR15 blocked the adhesion to a combination of MAdCAM-1 and VCAM-1 in vitro to a similar degree as the pan-α4-integrin antibody natalizumab did [which has, however, been associated with fatal cases of progressive multifocal leukoencephalopathy38,39].
At the same time, it is also obvious that GPR15 is not as gut-specific as α4β7 integrin is and that this must be considered as a caveat in terms of therapeutic applications. A key role of GPR15 in many homeostatic processes, as well as in pathological conditions, has been described. For instance, GPR15 seems to be involved in regulating angiogenesis via thrombomodulin, which is also a ligand of GPR15.40 Moreover, the GPR15-dependent recruitment of regulatory T cells to the colorectal cancer environment seems to promote tumorigenesis.41 GPR15 has furthermore been associated with myocardial infarction,42 rheumatoid arthritis,43 Grave’s disease,44 and multiple sclerosis.35 Thus, a potential negative impact of blocking GPR15 on other organ systems and/or conditions needs to be ruled out in future studies.
Collectively our data show, for the first time, how GPR15 drives T cell recruitment to the gut, and thereby mandate further preclinical and translational studies to determine its suitability as a target for the treatment of IBD.
Supplementary Data
Supplementary data are available at ECCO-JCC online.
Funding
This work was supported through grants by the Deutsche Forschungsgemeinschaft [DFG, German Research Foundation; ZU 377/4-1, TRR 241 – 375876048], the Else Kröner-Fresenius-Stiftung [2021_EKCS.23], and the Litwin IBD Pioneer’s initiative of the Crohn’s and Colitis Foundation of America.
Conflict of Interest
MFN has served as an advisor for Pentax, Giuliani, MSD, Abbvie, Janssen, Takeda, and Boehringer. SZ received speaker’s fees from Takeda, Roche, Galapagos, Ferring, Lilly, Falk, and Janssen. MFN and SZ received research support from Takeda, Shire [a part of Takeda,and Roche. The other authors declare no conflicts of interest.
Author Contributions
SS, LL, and MS performed the experiments. SS, LL, and SZ designed the study and analysed and interpreted the data. SS, LD, TMM, IA, CJV, RA, MFN, and SZ provided clinical samples, protocols, or reagents; SS, LL, and SZ drafted the manuscript; all authors critically revised the manuscript for important intellectual content.
Acknowledgements
The research of TMM, IA, RA, CJV, MFN, and SZ was supported by the Interdisciplinary Center for Clinical Research [IZKF] and the ELAN programme of the Universität Erlangen-Nürnberg, the Fritz-Bender-Stiftung, the Ernst Jung-Stiftung, the Else Kröner-Fresenius-Stiftung, the Thyssen-Stiftung, the German Crohn’s and Colitis Foundation [DCCV], the DFG Collaborative Research Centers 1181, KFO5024 and TRR241, and the Rainin Foundation. The authors thank J. Derdau, D. Dziony, J. Marcks, and J. Schuster (all Department of Medicine 1, University Hospital Erlangen) for their invaluable technical assistance.
Data Availability
Data are available from the authors upon reasonable request.
References
Author notes
Equal contributions.
