Mice were bred and maintained in a conventional animal facility at the Max Planck Institute of Immunobiology and Epigenetics (Freiburg, Germany) according to local regulations. Animal breeding was conducted with approval by the Regierungspräsidium Freiburg. All mouse strains used in this study were without health burden. Adult mice (>8 weeks) were used. WT controls were Cre-negative and age- and sex-matched for all experiments. Littermate controls were used for ex vivo migration experiments and analysis of endogenous immune cell populations. Removal of organs from CO₂ euthanized mice was performed by FELASA B-trained personnel in accordance with the relevant guidelines and regulations approved by the animal welfare review committee of the MPI-IE. No animal experimentation was performed in this study. The following mouse strains were on a C57BL6/J background and have previously been generated and published: Grk2 ^fl/fl, (JAX stock #012458) (20), Grk3 ^−/−, (kindly provided by Teresa Tarrant) (21), Grk5 ^fl/fl (JAX stock #010960) (22), Grk6 ^fl/fl (JAX stock #010962) (23), Mrp8-Cre (JAX stock #021614) (24), Vav1-iCre (JAX stock #008610) (25), Cd11c-Cre (JAX stock #008068) (26), Cd19^Cre (JAX stock #006785) (27), Rorc-Cre (JAX stock #022791) (28). Mouse crosses are indicated in the respective figures.

Neutrophils were isolated from bone marrow (tibia, femur, and os coxae) using the MACS neutrophil isolation kit for negative selection (Miltenyi Biotec) and an autoMACS^® Pro Selector cell separator (Miltenyi Biotec) according to the manufacturer’s protocol and as previously described (7). T- and B-cells were isolated from spleens by negative selection using the EasySep™ mouse T cell or EasySep™ mouse B cell isolation kit (Stemcell Technologies), respectively.

BMDCs were generated from bone marrow (isolated from femur, tibia and os coxae) in R10 culture medium (RPMI 1640 Medium, GlutaMAX™ Supplement (Thermo Fisher Scientific), 10 U/ml penicillin (Sigma), 10 µg/ml streptomycin (Sigma), 10% heat-inactivated fetal bovine serum (Sigma)) supplemented with 20 ng/ml recombinant murine GM-CSF (PeproTech). Cultures were maintained at 37° C and 5% CO₂ in a humidified incubator for 8-9 days. For BMDC stimulation, cells were cultured for 24 h in R10 supplemented with 20 ng/ml GM-CSF and 20 ng/ml LPS (from E.coliO127:B8, L45116, Sigma).

To generate Flt3-Ligand (FLT3L) dendritic cells, bone marrow was cultured for 8 days undisturbed in R10 culture medium supplemented with 55 μM 2-mercaptoethanol and 200 ng/ml recombinant murine FLT3L (PeproTech).

Neutrophils used for under-agarose migration assays were labeled for 25 min with CellTracker™ Green (CMFDA, 0.5 µM) or 5-carboxy-tetramethylrhodamine, succinimidyl ester (5-TAMRA SE, 10 µM) in PBS with 0.0002% Pluronic™ F-127 (all Thermo Fisher). BMDCs were labeled for 25 min with CellTracker™ Green (CMFDA, 1 µM) or 5-carboxytetramethylrhodamine, succinimidyl ester (5-TAMRA SE, 10 µM) in PBS with 0.0002% Pluronic™ F-127. The under-agarose chemotaxis assay was performed as described elsewhere (7, 29). In brief, 1.2% (w/v) (for neutrophils) or 0.6% (w/v) (for BMDCs) ultrapure agarose (Invitrogen) in a solution of phenolred-free RPMI, 2x HBSS (both Thermo Fisher Scientific), and 10% FCS was cast into tissue culture dishes and was allowed to solidify. A central attractor hole surrounded by four responder holes in a distance of 3 mm were punched into the polymerized gel and the gel was allowed to equilibrate for one hour at 37° C and 5% CO₂. The responder holes were loaded with differentially dye-labeled cells at a concentration of 5 × 10⁶ cells/ml (for neutrophils) or 1.25 × 10⁶ cells/ml (for BMDCs) in R10. The central well was loaded with chemokines at the indicated concentrations: LTB4 (Cayman), murine CXCL2 (PeproTech), WKYMVM (Tocris), WKYMVm (Tocris), murine C5a (PeproTech) for neutrophil chemotaxis experiments and murine CCL19 (PeproTech) for BMDC chemotaxis. In experiments with WKYMVm, neutrophils were 30 min preincubated with the FPR2 inhibitor WRW4 (Tocris, 100 µM) and also migrated in the presence of WRW4. Neutrophils were allowed to migrate for 4 h, BMDC for 16 h at 37° C and 5% CO₂ in a humidified incubator. Live-cell video microscopy was done with a spinning-disk confocal microscope (Cell Observer SD system, Carl Zeiss, with a CSU-X1 confocal scanner unit, Yokogawa and an AxioObserver Z1 inverted microscope stand) equipped with a stage-top incubator (Tokai-Hit). Images were taken using an Evolve^® back-illuminated EM-CCD camera (Teledyne Photometrics) and a plan-apochromat 10x 0.45 objective. End-point images were taken at the beforementioned spinning-disk confocal microscope or a LSM780 fluorescence confocal microscope (Zeiss). Tiles were stitched using ZEN blue software (Carl Zeiss Microscopy). For analysis of migration videos, cells were manually tracked using Imaris software (versions 7.5 – 9.5, Bitplane), coordinates were exported and further analyzed using a custom R script [R version 4.0.2, RStudio version 1.3.959, ggplot2 version 3.3.2 (7)]. For the analysis of endpoint pictures, cell center points were identified using the Imaris spot function. Displacement was measured from the border of the responder hole and the median per well was calculated as one technical replicate. In rare cases, technical replicates were excluded from analysis when cells did not respond or the agarose gel was damaged.

Transwell assays were used to analyze T-cell and B-cell migration towards various chemoattractants and chemokines. Lymphocytes isolated from spleen were incubated in serum-free medium (2% fatty-acid free BSA, 10 mM HEPES in RPMI) for 1 h at 37°C and 5% CO₂ in a humidified incubator prior to migration. HTS Transwell^® 96 well plates with polycarbonate membrane and 5-µm pores were used (Corning). The lower chamber was filled with chemoattractants in serum-free medium (CXCL12, CXCL13, CCL19, CCL21, S1P; concentrations as indicated in the graphs), and cells (400.000 cells/well in serum-free medium) were placed in the upper chamber and allowed to migrate for 3 h at 37° C and 5% CO₂ in a humidified incubator. Cells in the bottom chamber were harvested, stained with DAPI and counted using a LSRFortessa™ (BD) flow cytometer. Migration efficiency of DAPI-negative cells to all chemokines was calculated as % of input, after the rate of spontaneous migration was subtracted. As migration of WT lymphocytes to S1P was in the range of spontaneous migration, it was not subtracted in these experiments. All assays were performed in technical duplicates. In rare cases, experiments were excluded from analysis when WT cells did not show any migration.

To test B cell adhesion, splenocytes were isolated and erythrocyte lysis performed. Follicular B cells (Lin^-, CD45⁺, B220⁺, CD23⁺) and marginal zone B cells (Lin^-, CD45⁺, B220⁺, CD21⁺, CD23^-) were sorted on a FACSymphony™ S6 or FACSAria™ Fusion flow cytometer (BD). Angiogenesis µ-slides (Ibidi) were coated overnight with 2 µg/ml recombinant mouse ICAM-1 Fc chimera protein or 2 µg/ml recombinant mouse VCAM-1 Fc chimera protein (R&D Systems). 15.000 sorted cells were used per well, stimulated with 1 µg/ml CXCL12 and 1 µg/ml CXCL13, and incubated for 30 min at 37° C and 5% CO₂ in a humidified incubator. Wells were carefully washed three times, Hoechst was added and complete wells were imaged using a LSM780 fluorescence confocal microscope (Zeiss). Tiles were stitched using ZEN blue software (Zeiss) and cells were counted using the Imaris spot function (Bitplane).

In order to obtain GM-DC and GM-Mac precursors from bone marrow (isolated from tibia, femur, os coxae), lineage-negative cells were enriched using a FITC positive selection kit (Stemcell Technologies) and FITC-conjugated antibodies against Ly6G, B220, CD3 and MHCII (all Biolegend) according to the manufacturer’s instruction. Cells were sorted on a FACSymphony™ S6 (BD) and progenitors were defined as follows: MDPs as Lineage⁻ (Ly6G, CD19, CD3, NK1.1, B220, Ter119, CD11c, MHCII, Zombie NIR™), CD115⁺, CD117^hi, CD11b⁻, DNGR-1⁻, Ly6C⁻, CD135⁺. cMoPs as Lineage⁻, CD115⁺, CD117^hi, CD11b⁻, DNGR-1⁻, CD135⁻, Ly6C⁺. CDPs as Lineage⁻, CD115⁺, CD117^low-neg, CD135⁺, DNGR-1⁺ (antibodies listed in Supplementary Table 1 ) (30). For GM-CSF cultures, 10.000 sorted progenitors were mixed with 750.000 CD45.1 bone marrow cells, which are a source of growth factors for the sorted progenitors and were used as internal control of culture conditions, and cultured as described above (30).

For flow cytometric analysis of lymphocytes from lymph nodes and spleens, cell suspensions were obtained using a 70-µm cell strainer. Erythrocyte lysis was performed with ACK lysing buffer (ThermoFisher). For the analysis of endogenous dendritic cells, spleens and lymph nodes were digested for 20 min, ear skin for 1 h at 37°C shaking at 1000 rpm in digestion buffer (0.5 mg/ml collagenase IV (abnova), 0.25 U/ml dispase (StemCell), 40 µg/ml DNase I (Roche), 10 µM MgCl₂, 5% FCS and 1% penicillin/streptomycin in HBSS). Cultured BMDCs were harvested through gentle washing with PBS. Unspecific binding was blocked with anti-CD16/32 (BD) and cells stained for 1.5 h on ice (antibodies listed in Supplementary Table 1 ). For intracellular staining, the Foxp3/transcription factor staining buffer set (eBioscience) was used according to the manufacturer’s instructions. DAPI, Zombie NIR™ (BioLegend) or fixable viability dye eFluor 506 (eBioscience) were used as live/dead cell dyes. Data was acquired on a LSRFortessa™ (BD) and analyzed using FlowJo™ software (BD, version 10).

Gating strategies for bone marrow precursors in Supplementary Figures 3A, B were adapted from (31): CMPs/MDPs were defined as Lin⁻ (CD3, B220, CD19, NK1.1, Ly6G, Ter119, CD11c, CD11b), Sca-1⁻, CD117⁺, CD16/32^lo, CD34⁺, CD135⁺; GMPs as Lin⁻, Sca-1⁻, CD117⁺, CD16/32^hi, CD34⁺, CD135⁻, Ly6C^-; cMoPs and GPs as Lin⁻, Sca-1⁻, CD117⁺, CD16/32^hi, CD34⁺, CD135⁻, Ly6C⁺; BM monocytes as Lin⁻, Sca-1⁻, CD117⁻, CD16/32^hi, Ly6C⁺; and CDPs as Lin⁻, Sca-1⁻, CD117^lo, CD16/32^lo, CD34⁺, CD135⁺, Ly6C⁻ (antibodies listed in Supplementary Table 1 ).

Efficiencies of conditional GRK depletion in BMDCs were determined by immunoblot analysis. BMDCs were lysed (50 mM Tris, 150 mM NaCl, 5 mM EGTA, 5 mM EDTA, 0.5% IGEPAL^® CA-630, 1% Triton™ X-100, 1x cOmplete™ protease inhibitor cocktail (Roche)) and lysates loaded on 12% polyacrylamide gels for electrophoresis under reducing conditions (SDS-PAGE). Separated proteins were immunoblotted on PVDF membrane and unspecific binding blocked with 5% milk powder in Tris-buffered saline with 0.1% Tween^® 20 (TBS-T). Membranes were stained overnight at 4°C with primary antibodies (listed in Supplementary Table 1 ) and actin as loading control and, after three wash steps with TBS-T, stained with horseradish peroxidase-coupled secondary antibodies (Dako). Chemiluminescence was detected using Clarity™ Western ECL substrate and a ChemiDoc™ Imaging System (both BioRad).

Conditional knockout efficiency of GRKs in T- and B-cells was determined by RT-qPCR. RNA was extracted from isolated splenic T- or B-cells using TRI Reagent^® (Sigma-Aldrich) and chloroform extraction. Reverse transcription was performed with SuperScript IV™ reverse transcriptase and random hexamere primers (ThermoFisher). qPCR was performed using ABsolute QPCR Mix, SYBR Green, ROX (ThermoFisher) according to the manufacturer’s instructions and a StepOnePlus™ Real-Time PCR system (Applied Biosystems). Primers are listed in Supplementary Table 2 , Grk expression levels were normalized against Actin and Gapdh.

For immunofluorescence analysis of lymphoid organs, lymph nodes were fixed with 1% PFA overnight, incubated in 30% sucrose for 8 h, and embedded in Tissue-Tek^® O.C.T.™ Compound (Sakura). Spleens were snap-frozen in liquid nitrogen before embedding. A CM3050 S Cryostat (Leica) was used to cut 20 µm thick sections. Before staining, spleen sections were fixed in ice-cold methanol for 10 min. Unspecific binding was blocked in blocking/staining buffer (0.1% Triton™ X-100, 1% BSA in PBS). Staining was done overnight in a humidified chamber at 4°C, antibodies are listed in Supplementary Table 1 . Sections were mounted with Fluoromount-G™ (SouthernBiotech) and images acquired using a LSM780 fluorescent confocal microscope (Zeiss).

Ears were mechanically split into two halves in a way that the cartilage stayed on the dorsal side. The ventral half was placed in pre-warmed R10 medium, floating with the open, dermal side facing the liquid and incubated for 16 h at 37°C and 5% CO₂ in a humidified incubator. After incubation, the tissue was fixed overnight in 1% PFA and subsequently subjected to immunofluorescent staining as described above using anti-MHCII antibodies for DCs and anti-Lyve1 for lymphatic vessels (see Supplementary Table 1 ). Images were acquired with a LSM780 fluorescence confocal microscope (Zeiss) using the tile function and z-stacks. Tiles were stitched with ZEN blue software (Carl Zeiss Microscopy). For image analysis, the DC signal was quantified using the surface function of Imaris (Bitplane, versions 7.5 to 9.5) to mask lymphatic vessels. In rare cases, images were excluded from analysis when the density of lymphatics in the imaging field of view clearly deviated from all other images.

RNA was isolated from BMDCs (unstimulated or LPS-stimulated for 24h) by standard phenol-chloroform extraction in technical triplicates. Libraries were prepared using the TruSeq stranded mRNA protocol (Illumina) and sequenced on a NovaSeq6000 (Illumina) with a reading depth of 15 × 10⁶ reads per sample. RNA-seq analysis was performed using the mRNA-seq function in snakePipes (version 2.1.0) (32). In brief, raw fastq files were aligned to the mm10 reference genome using STAR (version 2.7.3a) (33) and expression count quantification was performed with featureCounts (version 2.0.0) (34). Quality control was performed using deeptools (version 3.3.2) (35). Differential gene expression analysis was performed using DESeq2 (version 1.26.0) (36) implemented in R (version 3.6.2). Expression data was further analyzed in R (R version 4.0.2, RStudio version 1.3.959). Results with a false discovery rate < 0.05 were considered significant.

Analyses were performed using Prism software (GraphPad Software, Inc. Version 9.1.0). First, normality of sample distribution was assessed. For comparison of two groups with normally distributed data, student’s t test was used. For comparison of two groups with not-normally distributed data, the nonparametric Mann-Whitney test was used. Ordinary one-way ANOVA and Tukey’s post-hoc tests for multiple comparisons were used for comparison of more than two groups of normally distributed data. Kruskal-Wallis tests and Dunn’s post-hoc tests were used for comparison of more than two groups of not-normally distributed data. For ratio datasets, log2 transformed values were computed and one sample t tests with a theoretical mean = 0 were used. “ns” indicates non-significant difference (P>0.05), stars indicate significance (*P<0.05, **P<0.01, ***P<0.001). For further statistical details, see Supplementary Table 3.

We began our study by examining neutrophils, which are key cells of the innate immune response with crucial roles for clearing invading bacteria and fungi in tissues. Neutrophils express a diverse repertoire of many GPCRs on their surface, allowing them to respond and migrate to a wide range of danger-associated signals, inflammatory chemokines and chemoattractants (37). In a previous study we carefully characterized the functional role of GRK2 for the desensitization of LTB4R1 and CXCR2, the two important GPCRs that mediate neutrophil swarming (7). We analyzed neutrophils responding to gradients of the swarm attractants LTB4 and CXCL2 in an under-agarose chemotaxis assay, and demonstrated that Grk2-deficient neutrophils show twice the displacement of control cells and continue to migrate in areas of high attractant concentrations. However, Grk2-deficient neutrophils did not show this enhanced migration in gradients of attractants binding to formyl peptide receptors (FPR1 and FPR2) or the complement component 5a anaphylatoxin chemotactic receptor 1 (C5aR1) (7). In that study, we also described Mrp8-Cre Grk2^fl/fl Grk3^−/− Grk5^fl/fl Grk6^fl/fl mice, which we used to generate and characterize primary mouse neutrophils with a complete depletion of all four expressed GRKs (referred to as 4xGrk ^−/−) ( Figure 1A ) (7). These cells showed the same migratory behavior as Grk2 ^−/− neutrophils in combined gradients of LTB4 and CXCL2, highlighting the particular role of GRK2 in controlling the swarm-mediating receptors LTB4R1 and CXCR2 (7). Here, we expanded our analysis of 4xGrk ^−/− neutrophils and investigate their chemotactic responses toward a more diverse set of attractants. Wildtype (WT) and 4xGrk ^−/− neutrophils were differentially dye-labeled before they migrated side-by-side in the under-agarose gel system from a responder hole toward a source of chemoattractants ( Figure 1B ). Confirming our previous results in combined gradients of LTB4 and CXCL2, 4xGrk ^−/− neutrophils migrated within 4 hours two to three times further than WT cells in response to single gradients of the intermediary attractants LTB4 or CXCL2 ( Figures 1C, D ). Neutrophils can respond to immune cell-derived “intermediate-target” chemotactic factors, but they can also move in response to so called “end-target” chemoattractants, which are factors released from damaged cells or pathogens (38). We then tested migration toward end-target chemoattractants (1), which are not involved in swarm-amplification but rather in the initial recruitment to sites of damage or infection ( Figure 1E ). In contrast to Grk2 ^−/− neutrophils (7), we observed significantly increased migration of 4xGrk ^−/− neutrophils towards W-peptides (ligands for the FPR1 and FPR2 receptors) and C5a ( Figures 1F, G ).

Thus, GRKs other than GRK2 contribute to the regulation of the chemotactic response toward end-target chemoattractants. Together, our data show that the complete depletion of all GRKs in neutrophils leads to an increase in GPCR-controlled migration across several GPCR types, supporting the canonical role of GRKs in regulating GPCR desensitization.

As we aimed for a comparative analysis of GRK functions in different primary immune cells, we expanded our experiments to subsets of lymphocytes. We started with T cells, key immune cells of the adaptive immune response, which as naïve T cells constitutively recirculate between blood and lymphoid organs under homeostatic conditions (39). T cells mainly express two GRK family members, GRK2 and GRK6 (4, 5), which we simultaneously targeted by generating Rorc-Cre Grk2^fl/fl Grk6^fl/fl (2xGrk ^ΔRorc) mice ( Figure 2A ). This approach allowed us to isolate primary splenic T cells lacking both GRKs (2xGrk ^−/−) and to analyze their chemotactic migration in vitro and homeostatic localization in lymphoid organs. Moreover, we could directly compare migratory responses of 2xGrk ^−/− T cells with single gene deficient Grk2 ^−/− and Grk6 ^−/− T cells ( Figure 2B ). By using transwell chemotaxis assays ( Figure 2C ), we systematically assessed the migration toward the most prominent chemoattractants and chemokines that promote the migration of naïve T-cells: sphingosin-1-phosphate (S1P), CCL19, CCL21 and CXCL12 (1).

Supporting the role of GRK2 as an important regulator of S1P-receptor (S1PR) desensitization (6), we confirmed previous findings that Grk2 ^−/− T cells show a substantially increased migration response to varying concentrations of S1P ( Figure 2D ; Supplementary Figures 1A, B ). In contrast, S1P-controlled migration of Grk6 ^−/− T cells was normal ( Figure 2D ; Supplementary Figures 1A, B ). Based on our findings with neutrophils, we assumed that in a situation of combined GRK deficiency, the phenotype of Grk2 ^−/− T cells would still be reflected in T cells lacking both GRK2 and GRK6. However, we found that 2xGrk ^−/− T cells did not show the same substantial increase in chemotaxis as observed in cells lacking only GRK2. Instead, the additional loss of GRK6 in 2xGrk ^−/− T cells mitigated the lack of desensitization seen in Grk2 ^−/− T cells ( Figure 2D ; Supplementary Figure 1B ), which contrasts the comparison of Grk2 ^−/− versus 4xGrk ^−/− neutrophils. Next, we tested the chemotactic behavior of T cells in response to the CCR7 ligands CCL19 and CCL21, homeostatic chemokines with important roles for T cell trafficking in lymphoid organs. We found that a deficiency of either GRK2 or GRK6 alone had only a weak (CCL19) or no (CCL21) influence on the migratory response of T cells. Surprisingly, however, a combined loss of GRK2 and GRK6 strongly inhibited T cell chemotaxis toward these chemokines ( Figure 2E ; Supplementary Figures 1A, B ). Interestingly, we observed a similar pattern of GRK-mediated regulation, when T cells migrated towards the CXCR4 ligand CXCL12 ( Figure 2F ; Supplementary Figures 1A, B ). Thus, combined depletion of GRK2 and GRK6 in T cells leads to migration deficiencies that are not detected upon depletion of an individual GRK. Moreover, the combined loss of both GRKs has differential effects for S1PR versus CCR7 and CXCR4 function, leading either to increased or reduced chemotactic migration, respectively.

As our in vitro experiments revealed contradicting roles for GRK-mediated T cell migration depending on the specific GPCR type, we were unable to predict how the combined loss of GRK2 and GRK6 would influence physiological T cell trafficking. Therefore, we analyzed T cell numbers and localization patterns in homeostatic spleens of mice. Confirming previous data from Grk2 ^ΔCD4 mice (6), we could not detect altered T cell numbers in Grk2 ^ΔRorc mice in the spleen ( Figure 2G ). As a consequence of impaired S1PR-desensitization, Grk2-deficient T cells populate only sparsely the CCL21-rich T-cell zone, but accumulate in the blood- and S1P-rich red pulp ( Figure 2H ). In agreement with only weak phenotypes of Grk6 ^−/− T cells in vitro, splenic T cell numbers and localization were unaffected in Grk6 ^ΔRorc mice ( Figures 2G, H ). Importantly, the analysis of 2xGrk2 ^ΔRorc mice revealed an overall reduction of splenic T cells numbers ( Figure 2G ), which was reflected in substantially reduced cell numbers in T-cell zones and red pulp ( Figure 2H ). Based on our in vitro characterization of 2xGrk ^−/− T cells, we argue that impaired attraction to CCR7 ligands is a major underlying cause for this T cell trafficking phenotype in the spleen. Next, we investigated the consequences of GRK depletions and misbalanced GPCR signaling for T cell homeostasis in lymph nodes (LNs). T cell numbers in LNs of Grk2 ^ΔRorc mice and 2xGrk2 ^ΔRorc mice were reduced by half ( Figure 2I ) with corresponding reductions of T cells in T-cell zones ( Figure 2J ). In contrast, T cell populations remained unaltered in Grk6 ^ΔRorc mice ( Figures 2I, J ).

Together, our results show that the combined functions of GRK2 and GRK6 play an important role for T cells in balancing the response toward competing chemokines and localization signals. Individual loss of GRK2 leads primarily to an increase in S1PR-mediated chemotaxis. In contrast, the additional loss of GRK6 strongly decreases CCR7-controlled T cell migration. Both conditions, however, result in a defective positioning of T cells to their dedicated compartments in lymphoid organs.

B lymphocytes, key immune cells of the humoral component of the adaptive immune response, share many GPCRs with T lymphocytes, and we next addressed whether the regulation by GRKs differs between those cell types. Similar to T cells, the homeostatic trafficking of naïve B depends on the expression of S1PR, CCR7 and CXCR4. In addition, B cells express CXCR5 to sense CXCL13, the main B-cell follicle chemokine (40). Similar to T cells, B cells also express GRK2 and GRK6 as the major two GRK family members (4, 5). To simultaneously target them in B cells, we generated CD19^CRE Grk2^fl/fl Grk6^fl/fl (2xGrk ^ΔCD19) mice ( Figure 3A ). This approach allowed us to isolate splenic primary B cells lacking both GRKs (2xGrk ^−/−), compare them to single gene deficient Grk2 ^−/− and Grk6 ^−/− B cells and proceed similar to our T cell analysis ( Figure 3B ).

When we analyzed the movement of B cells toward CXCL13, the CXCR5-dependent migration response was not altered by the absence of GRK2 and GRK6 ( Figure 3C ; Supplementary Figures 1C, D ). Next, we studied B cell migration toward S1P ( Figure 3D ; Supplementary Figures 1C, D ). We confirmed that Grk2 ^−/− B cells showed strongly increased chemotaxis, supporting previous reports of GRK2-mediated S1PR-desensitization (6, 13). In contrast, Grk6-deficiency did not alter the chemotactic response to S1P. Thus, Grk2 ^−/− and Grk6 ^−/− B cells showed comparable phenotypes to T cells with the same single GRK deficiency. However, 2xGrk ^−/− T cells and 2xGrk ^−/− B cells had distinct phenotypes ( Figures 2D , 3D ). In contrast to T cells, the additional loss of GRK6 in 2xGrk ^−/− B cells did not mitigate the pro-migratory effect of Grk2 ^−/− B cells ( Figure 3D ; Supplementary Figures 1C, D ). Further experiments investigated the chemotactic behavior of CCR7-expressing B cells to CCL19 and CCL21 ( Figure 3E ; Supplementary Figures 1C, D ). Analogous to T cells, single Grk2-deficiency reduced B cell chemotaxis to CCL19 (6, 13). Unexpectedly, Grk6 ^−/− B cells showed however an increase in chemotaxis, which was further pronounced in 2xGrk ^−/− B cells. Migration to CCL21 showed the same trends as to CCL19, albeit the migration phenotypes were less pronounced ( Figure 3E ; Supplementary Figure 1C ). Together, we demonstrate that the combined depletion of GRK2 and GRK6 has opposite effects for the CCR7-controlled migration of T cells and B cells. While T cell movement deteriorates, the migration of B cells improves. Lastly, we also assessed the CXCR4-dependent B cell response toward CXCL12 and show that combined GRK2 and GRK6 depletion leads to impaired B cell migration ( Figure 3F ; Supplementary Figures 1C, D ), similar to T cells ( Figure 2F ). Together, these experiments further exemplify that complete GRK deficiency causes entirely different migratory outcomes, depending on the specific immune cell subset and GPCR type.

Physiological B cell trafficking in secondary lymphoid organs depends on the sensing and interpretation of multiple chemoattractant and chemokine gradients. As our in vitro analysis revealed GRK mutant B cells with differential responsiveness for homeostatic attractants, we wondered whether and how this influences B cell localization in spleen and LNs. When we performed flow cytometric analysis of the percentages of splenic B cells ( Figure 3G ) or follicular B cells ( Supplementary Figure 1E ) in various B cell-specific GRK mutant mice, we did not detect gross differences between genotypes. Supporting previous work with Mb1^Cre Grk2^fl/fl mice (6, 13), we confirmed that Grk2 ^ΔCD19 mice have substantially altered intra-splenic B cell distribution ( Figure 3H ). Grk2 ^−/− B cells are drawn out of splenic B cell follicles into the red pulp. In contrast, spleens of Grk6 ^ΔCD19 showed no obvious changes in the localization of follicular B cells ( Figure 3H ). Strikingly, the additional lack of GRK6 in 2xGrk ^ΔCD19 mice led to a partial rescue of splenic B cell distribution in comparison to Grk2 ^ΔCD19 mice. B cell follicle sizes were almost restored, and B cell accumulation in the red pulp was not as pronounced as in Grk2 ^ΔCD19 mice ( Figure 3H , 4^th image). Thus, these observations suggest that the lack of S1PR1 desensitization, which is equally seen in Grk2 ^−/− and 2xGrk ^−/− B-cells and drives Grk2 ^−/− B-cells into the red pulp, is partially counter-balanced in 2xGrk ^−/− B cells by their altered responsiveness to CCR7 and CXCR4 ligands.

Besides follicular B cells, we also analyzed the population of marginal zone (MZ) B cells. Both immunofluorescence stainings ( Figure 3I ) and flow cytometric measurements ( Figure 3J ) revealed significantly reduced numbers of MZ B cells in Grk2 ^ΔCD19, Grk6 ^ΔCD19 and 2xGrk ^ΔCD19 mice. These phenotypes could not be explained by changed integrin receptor expression or integrin-mediated adhesion, which remained unaltered ( Supplementary Figures 1F, G ).

Lastly, we also examined B cell numbers and distribution in LNs of mice. Similar to T cells in the LN, we measured strongly reduced B cell numbers ( Figure 3K ) and diminished sizes of B cell follicles ( Figure 3L ) in Grk2 ^ΔCD19 and 2xGrk ^ΔCD19 mice. In addition, we observed a lack of B cells distributed around high-endothelial venules (HEVs) that are located within the T cell area ( Supplementary Figure 1H ).

Overall, our results demonstrate that the GRK-mediated regulation of GPCR function cannot be generalized across immune cell subsets. Although T and B lymphocytes share many general characteristics, including the expression of the major GRK family members (GRK2, GRK6) and the same cell surface GPCRs (S1PR, CCR7, CXCR4), we here show that the GRK-mediated control of GPCRs can be remarkably different between both cell types. In particular, our experiments with complete GRK depletions revealed some of these unexpected cell-type specific roles of GRKs and the need to consider the specific GPCR type in a defined immune cell subset.

We completed our systematic analysis of GRK-mediated regulation of immune cell migration by studying dendritic cells (DCs), important immune cells with a bridging role between the innate and the adaptive immune systems. These cells reside in peripheral and lymphoid tissues where they acquire antigens, which they process and then present to T cells (41). Although it is known that DCs express three members of the GRK family (GRK2, GRK3 and GRK6), we have only limited insight into their function for DC biology. While single gene knockouts of Grk3 or Grk6 were previously used to study BMDC migration in vitro (9, 15), we are completely missing any knowledge on the functional role of GRKs for DC homeostasis in vivo.

This prompted us to simultaneously target all three GRKs by generating CD11c-Cre Grk2^fl/fl Grk3 ^−/− Grk6^fl/fl (3xGrk ^ΔCD11c) mice ( Figure 4A ), which allowed the study of endogenous DCs depleted of all expressed GRKs. In contrast to neutrophils and lymphocytes, which respond to many different chemoattractants and chemokines, the migration of activated DCs is predominantly driven by CCR7 ligands. CCR7-expressing DCs sense gradients of CCL21 at lymphatic vessels and in LNs, which thus determine the migratory route of peripheral DCs from the tissue periphery into the secondary lymphoid organ ( Figure 4B ). The first step of this migratory route, leaving the interstitium and entering skin lymphatic vessels, can be visualized in the ear skin crawl-out assay ( Figure 4B ; Supplementary Figure 2A ). In this experimental setup, the ex vivo culture of ear skin halves over 16 h activates endogenous skin-resident DCs and triggers their migration into lymphatics ( Figure 4C ; Supplementary Figure 2A ). We quantified immunofluorescence stainings of DCs inside and outside of lymphatics in tissue regions of comparable lymphatic vessel density ( Supplementary Figure 2B ) and found that the majority of WT DCs migrated into the lymphatics during this timeframe ( Figures 4C, D ). In striking contrast, only a small fraction of endogenous skin DCs entered lymphatic vessels in the ears of 3xGrk ^ΔCD11c mice ( Figure 4D ), although flow cytometric analysis measured comparable numbers of CD11c⁺ MHCII⁺ cells between these knockout and control mice ( Figure 4E ). This phenotype was only observed in ears of triple GRK knockout mice and could not be detected in ears of mice with single or double GRK knockout combinations ( Figure 4D ). After entering the lymphatics, DCs follow them until they reach the draining LNs. This migration also takes place under homeostasis and the number of endogenous migratory DCs in skin-draining LNs can be assessed based on their very high MHCII-expression levels in comparison to LN-resident DCs ( Figure 4F ). In agreement with our results from the ear skin crawl-out assay, we measured reduced fractions of migratory DCs in the skin-draining LNs of 3xGrk ^ΔCD11c mice. This phenotype was again specific to triple GRK knockout mice and could not be measured in mice deficient for either GRK3 or GRK6, or a combination of GRK2 and GRK6 ( Figure 4F ).

To study DC migration in more detail, we generated bone marrow-derived DCs (BMDCs) from bone marrow precursor cells in GM-CSF culture medium. We used bone marrow from either 3xGrk ^ΔCD11c mice or Vav-iCre Grk2 ^fl/fl Grk3 ^−/− Grk6 ^fl/fl (3xGrk ^ΔVav) mice, which have early Cre activity in hematopoietic cell precursor stages, and found more efficient GRK depletion in BMDCs from 3xGrk ^ΔVav mice (data not shown). GRK depletion in BMDCs from 3xGrk ^ΔVav mice resulted in an almost complete loss of all expressed GRKs (3xGrk ^−/−) ( Figures 4G, H ). Lipopolysaccharide (LPS) stimulation for 24h causes BMDCs to acquire a mature phenotype and develop into DCs that chemotactically respond to CCR7 ligands. Side-by-side comparison of LPS-matured 3xGrk ^−/− BMDCs with control cells in under-agarose chemotaxis assays revealed a pronounced chemotaxis defect of the triple GRK knockout DCs in response to CCL19 gradients ( Figure 4I ; Supplementary Video 1 ). The migration track straightness of 3xGrk ^−/− DCs was severely reduced ( Figure 4J ), which resulted in a drop of total displacement towards the chemokine source ( Figure 4K ). More detailed analysis revealed that this defect is present during the whole migration time, but more pronounced in the late phase (200-700 min) ( Supplementary Figures 2C, D ). The migration speeds of control and 3xGrk ^−/− BMDCs were comparable for both phases ( Supplementary Figure 2E ). Not only did we observe that 3xGrk ^−/− DCs show reduced total displacement, but also that less cells responded to the stimulus and migrated out of the responder hole ( Figure 4L ). We also assessed the migration of BMDCs with single or double Grk-deficiency. Confirming our in vivo findings, Grk3 ^−/− and Grk6 ^−/− BMDCs displaced normally toward CCL19 gradients ( Figure 4K ). A migration-dampening effect in this in vitro setup was observed for Grk2 ^−/− and Grk2/6 ^−/− BMDCs ( Figure 4K ). As we had however not observed phenotypes of these KO combinations in situ and in vivo ( Figures 4D, F ), we decided to focus on the analysis of 3xGrk ^−/− DCs for all following experiments.

Together, our experiments with single and combinatorial GRK depletions in DCs do not support a role for this kinase family in controlling CCR7 desensitization during chemotactic migration. Instead, we find that complete GRK depletion in 3xGrk ^−/− DCs results in defective CCR7-mediated migration in situ and in vivo.

The migration defect of LPS-matured 3xGrk ^−/− BMDCs prompted us to examine these GM-CSF cultured cells in more detail ( Figure 5A ). As CCR7 is critical for the migration of mature DCs, we analyzed its expression in LPS-matured BMDCs. Flow cytometric analysis showed that both CCR7 cell surface expression and CCR7 total expression, including the intracellular pool, were reduced in LPS-stimulated 3xGrk ^−/− BMDCs in comparison to WT cells ( Figure 5B ). To clarify whether this resulted from a generalized DC maturation defect, we measured other typical cell surface molecules of mature DCs. Indeed, several DC maturation markers, including MHCII, CD80, CD86 and CD40, were diminished on the cell surface of LPS-matured 3xGrk ^−/− BMDCs ( Figure 5C ). However, the cell surface expression levels of the classical DC marker CD11c were unaltered ( Figure 5C ). As these results suggested defective DC maturation, we sought to confirm it by performing bulk RNA sequencing on unstimulated and LPS-matured 3xGrk ^−/− and WT BMDCs. To our surprise, these analyses revealed that not only maturation-related, but also many more genes that define a bona fide DC gene signature (42) were downregulated in 3xGrk ^−/− BMDCs ( Figure 5D , petrol). Instead, a macrophage (Mac) gene signature (43) emerged in both unstimulated and LPS-stimulated cells ( Figure 5D , purple).

Previous work revealed that GM-CSF cultured BMDCs are a heterogeneous myeloid cell population consisting of bona fide DCs (termed “GM-DC”) and a minor fraction of macrophage-like CD11c⁺ MHCII⁺ cells (termed “GM-Mac”) (30). When we refined our flow cytometric analysis for distinguishing GM-DCs and GM-Macs in BMDC cultures, we found a clear shift between both cell populations in 3xGrk ^−/− BMDC cultures. Among CD11c⁺ MHCII⁺ cells, the fraction of GM-DCs was decreased, whereas GM-Macs were markedly increased in both unstimulated and LPS-stimulated 3xGrk ^−/− GM-CSF cultures ( Figure 5E ). Importantly, the remaining bona fide GM-DCs, which still developed in 3xGrk ^−/− cultures, also showed reduced expression of CCR7, MHCII and CD40 ( Figure 5F ). Together, these findings provide an explanation for the two observed phenotypes of 3xGrk ^−/− cells in the under-agarose chemotaxis assays; (a) less responding cells, because GM-Macs do not migrate to CCL19 (30), and (b) reduced displacement of the responding cells, because the remaining bona fide GM-DCs remain lowered in CCR7-expression. Additionally, we noticed that not only DCs derived from GM-CSF cultures showed an altered phenotype, but also DC development in the presence of Flt3 ligand was impaired. In particular, the number of CD11b⁻ DCs was strongly reduced, whereas the CD11b⁺ DC population was not significantly altered ( Supplementary Figure 3A ).

In the following experiments, we focused on a more detailed analysis of the development of GM-DCs and GM-Macs in 3xGrk ^−/− cultures. It has been shown that different bone marrow-derived precursors contribute to both cell states: GM-DCs derive from common DC precursors (CDPs), GM-Macs from common monocyte precursors (cMoPs). CDPs and cMoPs derive from macrophage-DC progenitors (MDPs) as common progenitor, which has the highest proliferation potential (30) ( Figure 5G ). Refined flow cytometric analysis of the bone marrow from 3xGrk ^ΔVav mice revealed decreased numbers of MDPs and CDPs, which suggests functional roles of GRKs for early hematopoietic stem cell stages ( Supplementary Figure 3B ). The very limited numbers of MDPs and CDPs in 3xGrk ^ΔVav mice made it impractical to sort these cells for further analyses. Instead, we used for the following experiments cells generated from 3xGrk ^ΔCD11c mice, whose bone marrow precursors do not yet express the Cre recombinase. Therefore, the frequencies of MDPs and CDPs are normal in these mice ( Supplementary Figure 3C ). Importantly, cultures from bone marrows of 3xGrk ^ΔCD11c mice showed the same, albeit mitigated, phenotypic hallmarks as cultures from bone marrow of 3xGrk ^ΔVav mice: defective BMDC displacement in under-agarose assays ( Supplementary Figure 3D ), a trend to reduced expression of DC maturation markers ( Supplementary Figure 3E ) and increase of GM-Macs over GM-DCs ( Supplementary Figure 3F ). To delineate whether the observed misbalance of GM-Macs/GM-DCs could be attributed to a specific differentiation path, we sorted CDP, cMoP and MDP cells from the bone marrow of WT or 3xGrk ^ΔCD11c mice and followed their expansion in GM-CSF culture medium. As expected, cMoPs and MDPs from WT mice were more proliferative than CDPs ( Figure 5H ). In comparison to WT cultures, the output of differentiated cells from CDPs of 3xGrk ^ΔCD11c mice was slightly decreased, whereas cMoPs and MDPs expanded more strongly ( Figure 5H ). Furthermore, MDPs did not differentiate into GM-DCs and GM-Macs at regular proportions. Instead, their increased proliferation generated mainly GM-Macs ( Figure 5I ).

Taken together, these results show that GRKs play a role in BMDC development with subsequent consequences for the migration response. Complete GRK depletion leads to a vastly increased number of CD11c⁺ macrophages in GM-CSF cultures, and those DCs, that still develop, are phenotypically impaired.

Having identified that complete GRK depletion shifts the development of GM-DCs toward GM-Macs in vitro, we next asked whether the development of endogenous DCs might also be altered. In particular, we were interested in the DC populations of secondary lymphoid organs, where several DC subsets can be distinguished. Two main populations of classical DCs (cDC1 and cDC2), which differ in origin and function, have previously been defined. Moreover, plasmacytoid DCs (pDCs) represent another DC subset in lymphoid organs.

First, we used immunofluorescence stainings to detect cDC1 and cDC2 in spleens, where these DC subsets have distinct localization patterns (44). Using the specific markers XCR1 (for cDC1) and 33D1 (for cDC2) we found comparable localization of both DC subsets in WT and 3xGrk ^ΔCD11c animals: cDC1s positioned in the T-cell zone of the white pulp ( Figure 6A ) and cDC2s in the splenic bridging channels ( Figure 6B ). Flow cytometric analysis revealed that the abundance of 3xGrk ^−/− cDC1s and cDC2s in comparison to all other immune cells was slightly decreased ( Figure 6C ), but the overall homeostasis of tissue-resident DCs appeared only mildly disturbed in 3xGrk ^ΔCD11c mice. Immunofluorescence stainings against CD11c revealed however a striking and unexpected phenotype. Massive accumulations of CD11c⁺ cells, mainly localized in the T-cell zones of the white pulp and bridging channels, were visible in the spleens of 3xGrk ^ΔCD11c mice ( Figure 6D ). Similar to cDCs, also pDCs could not account for this observation, as pDC numbers were comparable in spleens of WT and 3xGrk ^ΔCD11c mice ( Supplementary Figure 4A ). As our in vitro results showed increased CD11c⁺ GM-Macs in 3xGrk ^−/− GM-CSF cultures, we next tested several macrophage and monocyte markers to characterize the accumulating CD11c⁺ population in vivo. CD11c⁺ cells in the T cell cortex did not co-express CD169, CD115 or F4/80 ( Figure 6D ; Supplementary Figures 4B, C ). Next, we analyzed the expression of CD68, which is often used as macrophage-marker, but can also be expressed by other cells of myeloid origin (45). While in WT splenic white pulp only discrete, very bright signals were visible, which correspond to cortical macrophages, we detected an additional diffuse CD68 staining in the bridging channels of 3xGrk ^ΔCD11c spleens ( Figure 6E ), suggesting a macrophage-like phenotype of the accumulating CD11c⁺ cells.

Lastly, we analyzed whether these macrophage-like CD11c⁺ cells were only present in spleens. Strikingly, we also detected an accumulation of CD11c⁺ cells in LNs of 3xGrk ^ΔCD11c mice, where these cells localized to the peripheral parts of the T-cell zone or distributed in smaller patches within the T-cell zone ( Figure 7A ). At the same time, the numbers of cDC1 and cDC2 in LNs were unchanged ( Figure 7B ) and the number of pDCs in LNs reduced in 3xGrk ^ΔCD11c mice ( Supplementary Figure 4A ). Analogous to the splenic population, also this LN population was CD68⁺ and additionally expressed MerTK ( Figure 7C ), a cell surface receptor expressed by many macrophage subsets.

In summary, 3xGrk ^ΔCD11c animals exhibit an accumulation of CD11c⁺ cells in secondary lymphoid organs, which did not resemble any conventional DC or macrophage population. Based on the expression of CD11c, CD68 and MerTK, we presume that these cells are of myeloid origin and phenotypically “in-between” macrophages and DCs. Presumably, they are the in vivo equivalent of the GM-Macs, which develop excessively in 3xGrk ^−/− GM-CSF cultures. Thus, complete GRK depletion in DCs has major consequences for DC development in vitro and in vivo.