Altered CXCR4 dynamics at the cell membrane impairs directed cell migration in WHIM syndrome patients

Altered CXCR4 dynamics at the cell membrane impairs directed cell migration in WHIM syndrome patients Eva M. Garc ıa-Cuesta a ,Jos  e Miguel Rodr  ıguez-Frade a ,Sof  ıa R. Gardeta a,b , Gianluca D’Agostino a , Pablo Mart ınez a , Blanca Soler Palacios a , Graciela Cascio c , Tobias Wolf d , Nicolas Mateos e , Rosa Ayala-Bueno a ,C  esar A. Santiago f , Pilar Lucas a , Lucia Llorente g , Luis M. Allende g , Luis Ignacio Gonz alez-Granado g,h , Noa Mart  ın-C ofreces i,j,k , Pedro Roda-Navarro g,l , Federica Sallusto d,m , Francisco S anchez-Madrid i,j,k , Mar  ıa F. Garc ıa-Parajo e,n , Laura Mart  ınez-Mu~ noz o,p , and Mario Mellado a,1 Edited by Michael Dustin, University of Oxford, Oxford, United Kingdom; received October 29, 2021; accepted February 16, 2022 Chemokine receptor nanoscale organization at the cell membrane is orchestrated by the actin cytoskeleton and influences cell responses. Using single-particle tracking analysis we show that CXCR4 R334X , a truncated mutant chemokine receptor linked to WHIM syndrome (warts, hypogammaglobulinemia, infections, myelokathexis), fails to nanoclusterize after CXCL12 stimulation, and alters the lateral mobility and spatial organization of CXCR4 when coexpressed. These findings correlate with multiple phalloidin-positive protrusions in cells expressing CXCR4 R334X , and their inability to correctly sense chemokine gradients. The underlying mechanisms involve inappropriate actin cytoskeleton remodeling due to the inadequate β-arrestin1 activation by CXCR4 R334X ,whichdisrupts the equilibrium between activated and deactivated cofilin. Overall, we provide insights into the molecular mechanisms governing CXCR4 nanoclustering, signaling and cell function, and highlight the essential scaffold role of β-arrestin1 to support CXCL12mediated actin reorganization and receptor clustering. These defects associated with CXCR4 R334X expression might contribute to the severe immunological symptoms associated with WHIM syndrome. cell migration jchemokine receptors jWHIM syndrome WHIM syndrome (an acronym for warts, hypogammaglobulinemia, recurrent bacterial infections, and myelokathexis) is a rare combined immunodeficiency disorder that is linked to inherited, heterozygous autosomal-dominant mutations in the gene for the chemokine receptor CXCR4 (1). Typically, mutations cause premature termination or a frameshift in the region that encodes the cytoplasmic tail of the protein, which is important for signaling (2, 3). Indeed, elevated G protein-dependent signaling has been reported in leukocytes from patients with WHIM syndrome and in cell lines expressing CXCR4 WHIM mutants, which correlates with impaired CXCR4 desensitization and internalization in response to the receptor ligand CXCL12 (2). This can be mechanistically linked to an inability of mutant CXCR4 to activate negative feedback loops, including the induction of G protein-coupled receptor kinase and protein kinase C-mediated phosphorylation of the C-terminus domain of the receptor for coupling to β-arrestins (4, 5). Accordingly, CXCR4 WHIM mutant receptors show enhanced signaling activation in the presence of CXCL12, and are therefore considered as gain-offunction mutants (6). The most common and best-studied WHIM mutation is CXCR4 R334X , which lacks the C-terminal 19 amino acids (7). Mice and zebrafish expressing CXCR4 R334X recapitulate the neutropenia and myelokathexis observed in patients with WHIM syndrome (8, 9). In addition to explaining the loss of receptor internalization (10), the partial deletion of the cytoplasmic tail could also lead to changes in its interactions with the actin cytoskeleton (11). Chemokines are known to regulate actin dynamics (12), which coordinate cell responses (13). Among its physiological roles, the actin cytoskeleton regulates membrane diffusion, protein compartmentalization and clustering, and controls receptor signaling (14). Recent data suggest that the actin cytoskeleton is essential for ligand-mediated CXCR4 nanoclustering, which is necessary for achieving complete receptor functions (15). The C-terminal end of CXCR4 is implicated in the recruitment of β-arrestin adaptor proteins, which link CXCR4 to a clathrin lattice and participate in coupling the receptor to the actin cytoskeleton (11). Previous studies have shown that G proteindependent actin reorganization is mediated by mechanisms involving members of the Rho family of small GTPases and cofilin activation (16, 17). This signaling cascade contributes to the Rac1-mediated generation of free actin barbed ends at the leading edge of motile cells (18). β-Arrestins have been postulated to spatially regulate actin polymerization by scaffolding cofilin at the leading edge (19, 20). Complexes between Significance New imaging-based approaches are incorporating new concepts to our knowledge of biological processes. The analysis of receptor dynamics involved in cell movement using single-particle tracking demonstrates that cells require chemokine-mediated receptor clustering to sense appropriately chemoattractant gradients. Here, we report that this process does not occur in T cells expressing CXCR4 R334X ,a mutant form of CXCR4 linked to WHIM syndrome (warts, hypogammaglobulinemia, infections, myelokathexis). The underlaying molecular mechanism involves inappropriate actin cytoskeleton remodeling due to the inadequate β-arrestin1 activation by CXCR4 R334X ,which alters its lateral mobility and spatial organization. These defects, associated to CXCR4 R334X expression, contribute to the retention of hematopoietic precursors in bone marrow niches and explain the severe immunological symptoms associated with WHIM syndrome. The authors declare no competing interest. This article is a PNAS Direct Submission. Copyright © 2022 the Author(s). Published by PNAS. This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND). 1 To whom correspondence may be addressed. Email: [email protected]. This article contains supporting information online at http://www.pnas.org/lookup/suppl/doi:10.1073/pnas. 2119483119/-/DCSupplemental. Published May 19, 2022. PNAS 2022 Vol. 119 No. 21 e2119483119 https://doi.org/10.1073/pnas.2119483119 1of12 RESEARCH ARTICLE | IMMUNOLOGY AND INFLAMMATION OPEN ACCESS Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238. β-arrestins and the actin-binding protein filamin A have been identified by proteomics analysis (21) and have been shown to functionally cooperate to regulate ERK activation and actin cytoskeleton reorganization (22). Using quantitative single-molecule spatio-dynamic imaging, we show here that CXCR4 R334X is unable to form large nanoclusters in response to CXCL12. Similar behavior was observed for CXCR4 in cells lacking β-arrestin1. These results suggest that— due to the lack of negative feedback loops—CXCR4 R334X triggers continuous G protein activation in the presence of CXCL12, which in turn distorts the balance between active and inactive cofilin, perturbing actin dynamics and ligand-mediated receptor nanoclustering in migrating cells. This phenotype concours with the formation of multiple lamellae in CXCR4 R334X - expressing cells through mechanisms involving sustained cofilin activation, which might also contribute to the inability of these cells to correctly sense chemokine gradients. Our results also indicate that by altering the activity of the actin-binding protein cofilin, β-arrestin–mediated signaling pathways influence the lateral partitioning of CXCR4 and the ability of cells to correctly sense chemoattractant gradients. Results Gain-of-Function CXCR4 R334X Does Not Enhance Cell Migration in a Lipid-Bilayer System. CXCR4 R334X behavesasagain-offunction mutant because of its inability to be desensitized (6). We tested the capacity of CXCR4 R334X to promote cell migration in transwell assays using transiently transfected Jurkat (JK) cells (cells expressing CXCR4 endogenously) or JKX4 / cells (CXCR4 deficient cells) (SI Appendix,Fig.S1Aand B). As expected, cell migration of GFP + cells to CXCL12 was higher in JKX4 / -R334X cells (CXCR4-deficient cells that transiently express CXCR4 R334X - AcGFP) than in JKX4 / -X4 cells (CXCR4-deficient cells that transiently express CXCR4-AcGFP), at all concentrations tested (Fig. 1A). In agreement with the dominant effect of mutant CXCR4 receptors when coexpressed in heterozygosis with wildtype CXCR4 (1), we found that ligand-mediated cell migration was higher in JK-R334X cells than in JK-X4 cells (Fig. 1A). We also found that the WHIM mutation promoted a significant reduction of CXCL12-induced receptor internalization in both JKX4 / -R334X and JK-R334X cells (Fig. 1 Band C), confirming previous work (6). As transwell experiments reflect the contribution of only some molecular components involved in migration, we tested the behavior of cells using another experimental model where cell migration depends not only on the expression of chemokine receptors, but also on integrin activation. Using a twodimensional lipid bilayer system with embedded intercellular adhesion molecule 1 (ICAM-1) and CXCL12, we found that the percentage of migrating (GFP + ) cells was significantly lower for JKX4 / -R334X and JK-R334X cells than for JKX4 / -X4 or JK-X4 cells (Fig. 1D). Although many factors can affect cell migration (23), we discarded significant differences in CXCL12-mediated cell adhesion to ICAM-1 between the different cells used (Fig. 1E). The reduction in cell migration in A none 50 40 30 20 10 5 nM 0 15 nM % Migration CXCL12 ** ** DE % cell adhesion 0 20 40 60 n.s. n.s. -+-+ CXCL12 -+-+ 0 10 20 30 40 % migrating cells JKX4-/--X4 JKX4-/--R334X JK-X4 JK-R334X JKX4-/--X4 JKX4-/--R334X JK-X4 JK-R334X * -+-+ ** CXCL12 -+ -+ B JKX4-/--X4 JKX4-/--R334X 120 100 80 60 40 20 time (min) 0 % of cell surface CXCR4 % of cell surface CXCR4 0 1 5 153045 JKX4-/--X4 JKX4-/--R334X C JK-X4 JK-R334X 120 100 80 60 40 20 time (min) 0 0 1 5 153045 JK-X4 JK-R334X Fig. 1. CXCR4 R334X expressed in both homoand heterozygosis alters CXCL12-mediated responses. (A) JK-X4, JK-R334X, JKX4 / -X4, and JKX4 / -R334X cell migration in Boyden chambers in response to CXCL12. Data are shown as the mean percentage (plus SD) of input cells that migrate (n=5; **P≤0.01). (B) Cell surface expression of CXCR4 and CXCR4 R334X in JK-X4 and JK-R334X cells after stimulation with CXCL12 (40 nM) at different time points and analyzed by flow cytometry using an anti-CXCR4 antibody in nonpermeabilized cells. Results show mean ±SEM of the percentage of CXCR4 expression at the cell surface (n=4). (C) Surface receptor expression in JKX4 / -X4 and JKX4 / -R334X cells was analyzed as in B. Results show mean ±SEM of the percentage of CXCR4 expression at the cell surface (n=3). (D) Migration frequency of JKX4 / or JK cells transiently transfected with CXCR4 wild-type (-X4) or with CXCR4 R334X (-R334X) on ICAM-1-containing lipid bilayers alone or together with CXCL12 (mean ±SD, n=3; *P≤0.05, **P≤0.01). (E) Cell adhesion frequency of cells as in Don ICAM-1–containing lipid bilayers alone or together with CXCL12 (mean ±SD, n=3; n.s., not significant). 2of12 https://doi.org/10.1073/pnas.2119483119 pnas.org Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238. the absence of differences in cell adhesion strongly suggests that CXCR4 R334X functions less efficiently than CXCR4 in experiments mimicking physiological cell migration. CXCL12 Fails to Promote Larger CXCR4 R334X Nanoclusters. By acting as a physical barrier, the actin cytoskeleton modulates plasma membrane compartmentalization and membrane protein dynamics (24, 25). Hence, actin dynamics not only define the cell shape during migration, but also have an essential role in coordinating chemokine receptor signaling (26). We previously found that abrogating actin polymerization with the actin monomer sequestering drug latrunculin A abolishes CXCL12mediated CXCR4 nanoclustering (15). We next used singleparticle tracking (SPT) in total internal reflection fluorescence (TIRF) mode to examine CXCR4 and CXCR4 R334X dynamics in JKX4 / cells. (Movies S1–S4). This technique allows the detection of individual molecules close to the plasma membrane with great signal-to-background ratio, and has been used to capture the dynamics of individual receptors or complexes as well as determining their stoichiometries (see SI Appendix, Supplementary Methods for more discussion) (27). Using cells expressing the receptors in homozygosis, we observed that CXCR4 and CXCR4 R334X dynamics were very similar in steady state; in both cases the highest proportion of CXCR4 particles corresponded to mobile particles (∼93% vs. ∼92%) (Fig. 2A). The median value of the short time-lag diffusion coefficient (D 1–4 ) for both CXCR4 and the mutant CXCR4 R334X was also similar (0.027 μm 2 /s for CXCR4 and 0.023 μm 2 /s for the mutant) (Fig. 2B). CXCL12 promoted a significant reduction in overall receptor diffusivity (basal, median D 1–4 =0.027 μm 2 s 1 ; CXCL12, median D 1–4 = 0.011 μm 2 s 1 ) and increased the percentage of immobile particles from ∼7% (basal) to ∼20% (CXCL12) in JKX4 / -X4 cells but not in JKX4 / -R334X cells (basal, median D 1–4 = 0.023 μm 2 s 1 ; CXCL12, median D 1–4 =0.037 μm 2 s 1 ). In the latter case, we also detected a similar percentage of immobile particles independently of ligand activation (∼8% basal vs. ∼6% CXCL12) (Fig. 2 Aand B). Both CXCR4 and CXCR4 R334X were found as predominantly monomers and dimers in steady state (∼90% for CXCR4 vs. ∼81% for CXCR4 R334X )(Fig.2C–E). Accordingly, basal intensity distribution was comparable for both receptor types (1,255 arbitrary units for CXCR4 vs. 1,521 arbitrary units for CXCR4 R334X ) (Fig. 2C). We observed an increase in the number of larger CXCR4 nanoclusters at the membrane of JKX4 / -X4 cells upon CXCL12 activation (∼64% of nanoclusters of ≥3 receptors), but this was not evident for CXCR4 R334X in CXCL12-activated JKX4 / -R334X cells (∼17%) (Fig. 2 Dand E). B D1-4(μm2/s) JKX4-/--X4 JKX4-/--R334X CXCL12 ++-- **** **** **** n.s. 0.00 0.02 0.04 0.06 0.08 0.10 0.2 0.4 04080 85 90 95 100 %particles mobile immobile CXCL12 - + - + A JKX4-/--X4 JKX4-/--R334X C Intensity (a.u.) 0 104 5x103 **** n.s. **** JKX4-/--X4 JKX4-/--R334X CXCL12 ++-- %particles D 123456789101112 0 20 40 60 80 0 20 40 60 80 100 0 20 40 60 80 100 JKX4-/--X4 JKX4-/--X4 + CXCL12 receptors/particle % particles ≤2 ≥3 JKX4 -/- -X4 CXCL12 +- JKX4-/--R334X JKX4-/--R334X + CXCL12 %parti cles E 123456789101112 0 20 40 60 receptors/particle %particles ≤2 ≥3 JKX4 -/- -R334X CXCL12 +- Fig. 2. CXCL12 does not influence CXCR4 R334X dynamics and nanoclustering. SPT analysis of CXCR4-AcGFP and CXCR4 R334X -AcGFP in JKX4 / cells on fibronectin (FN) or FN+CXCL12-coated coverslips (697 particles in 64 cells on FN; 1,221 in 66 cells on FN+CXCL12 in JKX4 / -X4 cells; 461 in 54 cells on FN; 775 in 72 cells on FN+CXCL12 in JK ×4 / -R334X cells; n=4). (A) Percentage of mobile and immobile CXCR4and CXCR4 R334X -AcGFP particles at the cell membrane. (B) Diffusion coefficients (D 1–4 ) of mobile single trajectories, with median (black line) corresponding to JKX4 / -X4 and JKX4 / -R334X cells as in A. (n.s., not significant, ****P ≤0.0001). (C) Intensity distribution (arbitrary units, a.u.) from individual CXCR4and CXCR4 R334X -AcGFP trajectories on unstimulated and CXCL12-stimulated JKX4 / transfected cells, mean is indicated (red) (n=3; n.s., not significant; ****P ≤0.0001). (Dand E) Frequency of CXCR4-AcGFP particles containing different number of receptors expressed as a histogram of unstimulated and CXCL12-stimulated JKX4 / -X4 (D) and JKX4 / -R334X (E) cells, calculated from MSI values of each particle as compared with the MSI value of monomeric CD86-AcGFP. The frequency of particles expressing monomers plus dimers (≤2) or nanoclusters (≥3) in both cell types is also shown (Right). PNAS 2022 Vol. 119 No. 21 e2119483119 https://doi.org/10.1073/pnas.2119483119 3of12 Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238. Essentially similar results were obtained when CXCR4 R334X dynamics were evaluated in heterozygosis (JK-R334X) (Fig. 3 and SI Appendix,Fig.S2), although in this case the endogenous expression of CXCR4 was associated with a reduction of CXCR4 R334X nanoclusters both in steady state and after CXCL12 stimulation (SI Appendix,Fig.S2). In addition, the expression of CXCR4 R334X altered the typical behavior of CXCR4 in response to CXCL12 when analyzed in JKX4 / cells stably expressing CXCR4 R334X (JKX4 / R334X + ) and transiently transfected with CXCR4-AcGFP (35.1% of nanoclusters in JKX4 / R334X + -X4 vs. 51.2% in JK-X4) (Fig. 3B). These results confirm a dominant effect of CXCR4 R334X on wild-type CXCR4 and concur with the ability of the mutant receptor to heterodimerize with CXCR4 even in the absence of ligand stimulation (28) (SI Appendix,Fig.S3). Altogether, the data indicate that CXCL12 does not trigger CXCR4 R334X nanoclustering nor does it increase the percentage of immobile particles; instead, it increases the diffusion of the mutant receptor at the cell membrane. These observations might correlate with the defects observed in the migration of JK-R334X and JKX4 / -R334X cells on lipid bilayers. CXCR4 R334X Abrogates Directed Cell Migration. Receptor nanoclustering influences several CXCL12-mediated responses, including ligand-induced directed cell migration (15). We utilized fibronectin-coated chemotaxis chambers to assess the ability of JKX4 / -R334X cells to migrate toward CXCL12 gradients. Results showed that whereas JKX4 / -X4 cells sensed the gradient, JKX4 / -R334X cells did not (Fig. 4A and Movies S5–S8). Quantitation of the results indicated that, compared with JKX4 / -X4 cells, CXCL12 exposure failed to increase the forward migration index and track straightness in JKX4 / -R334X cells (Fig. 4 Band C). Similarly, JK cells coexpressing both receptors, JKX4 / R334X + -X4 (heterozygosis), were unable to sense CXCL12 gradients (Fig. 4 D–F). These data show that the ability to sense CXCL12 gradients is blocked in cells expressing CXCR4 R334X . Actin cytoskeleton dynamics not only regulate receptor compartmentalization, but are also important for maintaining directional migration by forming and stabilizing protrusions or lamellipodia at the leading edge of motile cells (29), a key element in the response to chemoattractant gradients. We thus tested the ability of CXCL12 to promote lamellipodia formation. JKX4 / -X4 and JKX4 / -R334X cells were activated with CXCL12, and phalloidin staining was evaluated in fixed cells by confocal microscopy. Quantitation of the number of protrusions demonstrated that both types of cells showed a spherical phenotype in steady-state conditions, with a weak phalloidinstaining pattern around the cell. CXCL12 promoted the rapid polarization of JKX4 / -X4 cells and phalloidin staining concentrated mainly in a unique protrusion (lamellipodium) (Fig. 5 A–C). In contrast, phalloidin staining in CXCL12-stimulated JKX4 / -R334X cells was randomly distributed in the cell in multiple protrusions (Fig. 5 A–C). Comparable results were obtained when we compared primary CD3 + T cells from WHIM patients (CXCR4 R334X ) with those of healthy donors. Whereas CXCL12 triggered multiple phalloidin-enriched protrusions in cells from WHIM patients, a unique polarized lamellipodium was evident in cells from healthy donors (Fig. 5 D–F). These results suggest that defects in actin cytoskeleton reorganization might not only affect the migration of cells expressing CXCR4 R334X but might also affect CXCL12-mediated receptor nanoclustering. The absence of negative feedback mechanisms for CXCR4 R334X facilitates the continuous activation of Gα i -mediated signaling. To determine whether this hyperactivated pathway might be linked to actin dynamics, we studied the actin-binding protein cofilin, which is activated in response to CXCL12 (30). Cofilin is inactivated by LIM kinase (LIMK)-mediated phosphorylation, which inhibits its binding to actin, and is reactivated by the phosphatase slingshot homolog 1 (SSH1), which enables actin filament depolymerization (31). We found that whereas CXCL12 mediated the rapid dephosphorylation/phosphorylation cycle of cofilininJKX4 / -X4 cells, dephosphorylation of cofilin was sustained in JKX4 / -R334X cells (Fig. 6A). However, other ligand-mediated signaling pathways, such as AKT phosphorylation, were normally activated (Fig. 6A). Again, heterozygous expression of CXCR4 R334X mimics the results obtained in JKX4 / -R334X. In this heterozygous model, we detected prolonged cofilin activation without a significant effect on AKT phosphorylation (Fig. 6B). A B JK-X4 D1-4 (μm2/s) CXCL12 ++-- JKX4-/-R334X+-X4 JKX4-/-R334X+-X4 0.00 0.02 0.04 0.06 0.06 0.08 0.10 0.12 **** **** **** n.s. % particles ≤ 2 ≥ 3 JK-X4 CXCL12 ++-- 0 20 40 60 80 100 Fig. 3. Heterozygous expression of CXCR4 R334X abrogates CXCL12-mediated CXCR4 dynamics and nanoclustering. SPT analysis of CXCR4-AcGFP in JKX4 / -X4 and JKX4 / -R334X + -X4 cells on fibronectin (FN) or FN+CXCL12-coated coverslips (1,176 particles in 60 cells on FN; 3,037 in 66 cells on FN+CXCL12 in JK-X4 cells; 867 in 72 cells on FN; 1,525 in 75 cells on FN+CXCL12 in JKP ×4 / R334X + -X4 cells; n=3). (A) Diffusion coefficients (D 1–4 ) of mobile single trajectories, with median (black line) corresponding to JK-X4 and JKX4 / -R334X + X4 cells (n.s., not significant; ****P ≤0.0001). (B) Frequency of CXCR4-AcGFP particles containing the same number of receptors [monomers plus dimers (≤2) or nanoclusters (≥3) in JK and JKX4 / R334X + cells, calculated from MSI values of each particle as compared with the MSI value of monomeric CD86-AcGFP. 4of12 https://doi.org/10.1073/pnas.2119483119 pnas.org Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238. To determine whether a disequilibrium in the dynamics of active/inactive cofilin might explain the inability of CXCL12 to mediate directed cell migration in JKX4 / -R334X cells, we blocked cofilin in its dephosphorylated state using BMS-3, a chemical inhibitor of LIMK1/2 activity (32), or in JK cells transiently transfected with GFP-SSH1 or GFP alone (as a control). As expected, cofilin was constitutively active independently of ligand stimulus in both JK cells treated with BMS-3 and in SSH1-transfected cells (Fig. 6 Cand D). Neither BMS-3 treatment nor SSH1 overexpression affected other signaling pathways, such as Akt phosphorylation (Fig. 6 Cand D). In both cases, CXCL12 failed to trigger direct cell migration (Fig. 6 E–H). These data support a role for cofilin in promoting directed cell migration and underscore the importance of CXCL12 controlling the function of this actin-binding protein through regulation of thekinaseLIMK1andthephosphataseSSH1. Our results thus far led us to hypothesize that in the absence of negative feedback, due to impaired β-arrestin activation mediated by CXCR4 R334X , CXCL12 promotes continuous Gα i -mediated signaling and alters the equilibrium of active/ inactive cofilin and, consequently, the actin dynamics. In turn, this limits the ability of CXCL12 to promote correct cell polarization and to trigger large receptor nanoclusters. β-Arrestin1 Links CXCR4 with Actin Cytoskeleton Dynamics and Regulates CXCL12-Mediated CXCR4 Nanoclustering. The scaffolding role of the β-arrestins has been implicated both in receptor desensitization and in the actin assembly events needed for the formation of gradient-sensing filopodia and lamellipodia at the leading edge of motile cells (20). Impaired chemokine receptor desensitization has been also associated with random rather than directed migration (33). Additionally, although some reports implicate β-arrestins in chemokine-mediated cell migration, only β-arrestin1 has been associated to CXCL12induced directed cell migration (34, 35). We thus questioned whether defects in β-arrestin activation might explain the lack CJKX4-/--R334XJKX4-/--X4 -0.5 **** 0.5 Forward Migration Index n.s. Track straightness (a.u.) JKX4-/--R334XJKX4-/--X4 n.s. **** 0.0 0.2 0.4 0.6 0.8 F EB AJKX4-/--R334XJKX4-/--X4 CXCL12 none y axis (μm) 400 200 -200 -400 0 0 400 400 200 200 -200 -200 -400 -400 0 400 200-200 -400 0 y axis (μm) 50 nM 0 50 nM 0 DJK-X4 CXCL12 none x axis (μm) 300 100 -300 -100 0 300 300 -300 -100 0 0 300 100 -300 -100 0 300 100 -300 -100 x axis (μm) JKX4-/-R334X+-X4 JKX4-/-R334X+-X4JK-X4 -0.4 CXCL12 -+ - +CXCL12-+ -+ CXCL12-+ - +CXCL12-+-+ 0.4 0.2 0.0 -0.2 Forward Migration Index **** n.s. Track straightness (a.u.) JKX4-/-R334X+-X4 JK-X4 0.0 0.2 0.4 0.6 0.8 **** n.s. Fig. 4. CXCL12 does not trigger directed migration of cells expressing CXCR4 R334X both in homoor heterozygosis. Migration of JKX4 / transiently transfected with CXCR4-AcGFP (JKX4 / -X4) or CXCR4 R334X -AcGFP (JKX4 / -R334X) cells (homozygosis) (A–C) or JK-expressing CXCR4-AcGFP (JK-X4) or JKX4 / R334X + cells transiently transfected with CXCR4-AcGFP (JKX4 / -R334X + -X4) (heterozygosis) (Dand E)onfibronectin-coated μ-slide chemotaxis chambers in response to the indicated gradient of CXCL12 concentration for 18 h. (Aand D) Representative spider plots showing the trajectories of tracked cells migrating along the gradient. Dots in the plots represent the final position of each single tracked cell. Gray triangle indicates CXCL12 gradient. (B,C,E, and F) Quantitative evaluation of forward migration index (Band E) and track straightness (Cand F). These graphs show the data of individual cells, with the mean indicated (red) (n=3; n.s., not significant; ****P ≤0.0001). PNAS 2022 Vol. 119 No. 21 e2119483119 https://doi.org/10.1073/pnas.2119483119 5of12 Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238. of ligand-mediated CXCR4 R334X nanoclustering and directed cell migration. To do this, we transiently transfected CXCR4-AcGFP into JKX4 / β-arrestin1 knockout cells (JKX4 / βarr / )(SI Appendix,Fig.S1C) to produce JKX4 / βarr / -X4 cells. We found that CXCL12 failed to induce cofilin phosphorylation (Fig. 7A) and to trigger directed cell migration (Fig. 7Band Movies S9 and S10), as demonstrated by forward migration index and track straightness quantitation (Fig. 7 Cand D). In contrast, other CXCL12-mediated signaling pathways, such as Akt phosphorylation, were unaffected (Fig. 7A). To evaluate whether defects in β-arrestin1 activation might be involved in the lack of CXCL12-mediated CXCR4 R334X nanoclustering, we next analyzed the consequences of β-arrestin1 deficiency for CXCR4-AcGFP dynamics. In steady state, CXCR4 behaved similarly in both JKX4 / -X4 and JKX4 / βarr / -X4 cells with respect to receptor nanocluster distribution (∼4% in JKX4 / -X4 vs. ∼9% JKX4 / βarr / -X4 of complexes of ≥3 receptors) (Fig. 7Eand Movies S11 and S12) and the diffusion coefficient (D 1–4 values of 0.024 μm 2 s 1 and 0.025 μm 2 s 1 , respectively) (Fig. 7F). In contrast, β-arrestin1 deficiency abrogated ligand-mediated receptor nanoclustering (∼54% in JKX4 / -X4 vs. ∼6% in JKX4 / βarr / -X4 of complexes of ≥3 receptors) (Fig. 7E). We also detected a 2-fold decrease in the diffusion in JKX4 / -X4 cells, while in JKX4 / βarr / -X4 the decrease was only 1.5-fold (0.011 μm 2 s 1 in JKX4 / -X4 vs. 0.015 μm 2 s 1 in JKX4 / βarr / -X4) (Fig. 7F). Among other signaling pathways that can be involved, our data implicate β-arrestin1 in CXCL12-mediated CXCR4 nanoclustering and dynamics and indicate that defects in its activation through CXCR4 R334X promote deficiencies in lateral partitioning CXCR4 JKX4-/--X4 JKX4-/--R334X Phalloidin 334X Phalloidin FNFN + CXCL12 A % cells with protrussions % cells with >1protrussions BC %cellswith protrussions E F % cells with >1protrussions **** **** 0 40 20 80 60 100 WHIMHD *** WHIMHD 0 20 40 60 80 100 0 20 40 60 80 100 JKX4-/--R334XJKX4-/--X4 **** JKX4-/--R334XJKX4-/--X4 0 40 80 120 **** **** FNFN + CXCL12FNFN + CXCL12 CD3 Phalloidin ICAM3 WHIM patient (WHIM) Healthy Donor (HD) D CXCL12-+-+ CXCL12-+ -+ Fig. 5. CXCL12 induces multiple lamellipodia through interaction with mutant CXCR4 R334X .(A–C) F-actin (phalloidin-TRITC, red) and CXCR4 (AcGFP, green) visualized by confocal microscopy in JKX4 / -X4 and JKX4 / -R334X cells adhered to fibronectin and treated or not with CXCL12 (100 nM), as indicated (n=2, more than 150 cells analyzed). (A) A representative cell type of each condition is shown. Original magnification 63×.(B) Percentage of cells with positive phalloidin staining (mean ±SEM, ****P ≤0.0001), and (C) percentage of cells showing >1 phalloidin + protussion/cell (mean ±SEM, ****P ≤0.0001). (D–F) F-actin (phalloidinTRITC, red), anti-CD3 (white) and anti-ICAM3 (green) visualized by confocal microscopy in PBMCs isolated from blood of patients with WHIM (R334X) andof healthy controls, adhered to fibronectin and treated or not with CXCL12 (100 nM), as indicated (n=2, more than 150 cells analyzed of each condition). (D)Arepresentative cell type of each condition is shown. Original magnification 63×.(E) Percentage of PBMCs with positive phalloidin staining (mean ±SEM, ****P ≤ 0.0001) and (F) percentage of PBMCs showing more than one phalloidin + protrusion per cell (mean ±SEM, ***P ≤0.001). Images were acquired using a planapochromat 63 ×/1.4 NA Oil DIC objective and without zoom. Resolution of acquired images is 1024 ×1024, although, for the figure, cells were cropped. 6of12 https://doi.org/10.1073/pnas.2119483119 pnas.org Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238. BMS-3 JK DMSO H none CXCL12 none CXCL12 **** n.s. -0.5 0.0 0.5 Forward Migration Index JK-SSH1JK-GFP F -0.2 0.0 0.2 0.4 *** Forward Migration Index none CXCL12 none CXCL12 n.s. JK-SSH1 JK-GFP C 01’ 5’ 15’ 30’ 01’5’15’30’ 45’ 45’ pAkt pCofilin Cofilin D pCofilin pAkt Cofilin 0 1’ 5’ 15’ 30’ 45’ 0 1’ 5’ 15’ 30’ 45’ JK-DMSO JK-BMS3 A pCofilin pAkt Cofilin 0 1’ 5’ 15’ 30’ 0 1’ 5’ 15’ 30’ JKX4-/--X4 JKX4-/--R334X 45’45’ BMS-3-treated JKDMSO-treated JK JK-SSH1JK-GFP CXCL12 none x axis (μm) y axis (μm) 400 200 -200 -400 0 0 400 400 200 200 -200 -200 -400 -400 0 400200-200-400 0 50 nM 0 CXCL12 none x axis (μm) y axis (μm) 400 200 -200 -400 0 0 400 400 200 200 -200 -200 -400 -400 0 400200-200-400 0 50 nM 0 EG Cofilin pAkt pCofilin B 0 1’ 5’ 15’ 30’ 0 1’ 5’ 15’ 30’ JK-X4 JKX4-/-R334X+-X4 45’45’ Fig. 6. CXCR4 R334X triggers sustained cofilin activation in response to CXCL12. Western blot analysis of cofilin phosphorylation in (A) JKX4 / -X4 or JKX4 / -R334X cells (homozygosis), (B) JK-X4 or JKX4 / R334X + -X4 cells (heterozygosis), (C) JK cells transiently transfected with GFP-SSH1 (JK-SSH1) or with GFP (JK-GFP) and (D) JK cells treated with BMS-3 or DMSO. As loading control, the membrane was reblotted with an anticofilin antibody. As control of cell activation, phospho-Akt was assessed in all cells used in A–D, which were treated with CXCL12 at the indicated time points (n=3). (E–H) Cells in Cand D were plated on fibronectin-coated μ-slide chemotaxis chambers and allowed to migrate in response to the indicated gradient of CXCL12 concentration for 18 h. (Eand G) Representative spider plots showing the trajectories of tracked JK-GFP–, JK-SSH1–, BMS-3–treated JK or DMSO-treated JK migrating along the gradient (n=3, in duplicate). Dots in the plots represent the final position of each single tracked cell. Gray triangle indicates CXCL12 gradient. (Fand H) Quantitative evaluation of the forward migration index of experiments performed in Eand G, respectively. Figures show the data of individual cells, with the mean indicated (red) (n=3; n.s., not significant; ***P ≤0.001; ****P ≤0.0001). PNAS 2022 Vol. 119 No. 21 e2119483119 https://doi.org/10.1073/pnas.2119483119 7of12 Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238. pAkt pCofilin Cofilin AJKX4-/--X4 JKX4-/-βarr1-/--X4 0 1’ 5’ 15’ 30’ 0 1’ 5’ 15’ 30’ 45’45’ E %particles ≤ 2 ≥3 CXCL12 ++-- 0 20 40 60 80 100 JKX4-/--X4 JKX4-/-βarr1-/-X4C -0.2 -0.1 0.0 0.1 0.2 0.3 JKX4-/--X4 JKX4-/-βarr1-/--X4 Forward Migration Index none **** CXCL12 none CXCL12 n.s. n.s. D JKX4-/--X4 JKX4-/-βarr1-/-X4none CXCL12 none CXCL12 *** 0.0 0.2 0.4 Track straightness (a.u.) BJKX4-/-βarr1-/--X4JKX4-/--X4 CXCL12 none x axis (μm) y axis (μm) 400 200 -200 -400 0 0 400 400 200 200 -200 -200 -400 -400 0 400200-200-400 0 50 nM 0 F D1-4 (μm2/s) CXCL12 -- ++ JKX4-/--X4 JKX4-/--βarr1-/-X4- ** n.s. **** ** 0.00 0.02 0.04 0.06 0.08 0.10 0.2 0.3 Fig. 7. β-Arrestin1 regulates the formation of large CXCR4 nanoclusters and directional cell migration in response to CXCL12. (A) Western blot analysis of cofilin phosphorylation in JKX4 / -X4 and JKX4 / βarr / -X4 cells treated with CXCL12 at the indicated time points (Top). As loading control, the membrane was reblotted with an anti-cofilin antibody (Bottom). As control of cell activation, phospho-Akt was assessed in JKX4 / -X4 and JKX4 / βarr / -X4 cells treated with CXCL12 at the indicated time points (Middle)(n=3). (B–D) JKX4 / -X4 and JKX4 / βarr / -X4 cells were plated on fibronectin-coated μ-slide chemotaxis chambers and allowed to migrate in response to the indicated gradient of CXCL12 concentration for 18 h. (B) Representative spider plots showing the trajectories of tracked JKX4 / -X4 and JKX4 / βarr / -X4 cells migrating along the gradient (n=3, in duplicate). Dots in the plots represent the final position of each single tracked cell. Gray triangle indicates CXCL12 gradient. (C) Quantitative evaluation of the forward migration index and (D) track straightness of experiments performed as in B. Figures show the data of individual cells, with the mean indicated (red) (n.s., not significant; ***P ≤0.001; ****P ≤0.0001). (Eand F) SPT analysis of CXCR4-AcGFP in JKX4 / -X4 and JKX4 / βarr / -X4 cells on CXCL12-coated coverslips with or without coated FN (JKX4 / -X4, 531 trajectories in 50 cells plated on FN, 634 trajectories in 50 cells on FN+CXCL12; JKX4 / βarr / -X4, 551 trajectories in 78 cells plated on FN, 587 trajectories in 70 cells on FN+CXCL12, n=3). (E) Frequency of CXCR4-AcGFP particles containing the same number of receptors (monomers plus dimers [≤2] or nanoclusters [≥3]), calculated from MSI values of each particle as compared with the MSI value of monomeric CD86-AcGFP. (F) Diffusion coefficients (D 1–4 ) of mobile single trajectories, with median (black line), corresponding to JKX4 / -X4 and JKX4 / βarr / -X4 cells (n.s., not significant; **P ≤0.01; ****P ≤0.0001). 8of12 https://doi.org/10.1073/pnas.2119483119 pnas.org Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238. of this mutant receptor. The mechanism might involve, through regulation of LIMK/SSH1 balance, the spatial control of cofilin activity, which was lost in cells lacking β-arrestin1. Indeed, we detected a continuous activation of cofilin, which in turn might affect actin reorganization in these cells. As actin polymerization also regulates CXCR4 nanoclustering, our data might also explain the evident differences in the lateral partitioning of CXCR4 and CXCR4 R334X receptors after ligand activation. Discussion Cell migration is a complex process involving myriad signaling proteins and receptors that act coordinately to activate intracellular pathways and promote polarized cell states and directional migration. To migrate directionally in response to external stimuli, the internal machinery of cells needs to be spatially organized, which involves the integration of biochemical and mechanical factors to generate force in a specific direction to move the cell forward. The classic image of migration on substrates is one of actin-driven protrusions occurring primarily at the front, and myosin-driven contractile forces generated at the rear, causing the cell to detach and move forward (36). Recent evidence indicates that the actin cytoskeleton also governs the nanoscale organization and lateral dynamics of chemokine receptors, which is essential for correct receptor function to allow cells to sense chemoattractant gradients (15). Due to the absence of desensitization mechanisms, CXCR4 R334X and other mutant CXCR4 proteins causing WHIM syndrome have been considered as gain-of-function receptors (1, 37). But, in agreement with the observed retention of mature neutrophils in the bone marrow of patients with WHIM, resulting in neutropenia and panleukopenia (7), we found that JK cells expressing CXCR4 R334X failed to exhibit enhanced migration on lipid bilayers with embedded ICAM-1 and CXCL12, and were unable to sense CXCL12 gradients in fibronectin-coated μ-chambers. The results argue against differences in CXCL12-mediated adhesion to ICAM-1 of these cells compared with controls, although the cells showed higher adhesion in steady state. Chemokine receptors are found concentrated at the leading edge of motile cells, on the flattened cell-substratum contact area (38). There, they act as a sensor mechanism for the directed migration of leukocytes through a chemoattractant gradient, as was elegantly shown using a chimeric receptor between rhodopsin and CXCR4 in T cells (39). Using SPT in TIRF-M mode, we observed that, at steady state, CXCR4 and CXCR4 R334X exhibited comparable spatial organization at the cell membrane, preferentially forming nonclustered entities (monomers and dimers) and a small fraction of nanoclusters (groups of ≥3 receptors) that are CXCR4 aggregates present at the cell membrane that facilitate some chemokine-mediated signaling events (15). However, we observed some variability in the percentage of nanoclusters between the cell types used in this study, specifically when CXCR4 R334X was expressed in heterozygosis versus homozygosis. Although further work is needed to clarify this effect, differences in the CXCR4/CXCR4 R334X ratio at the cell surface might affect dynamic determinations, as unlabeled receptors are not detected in our technical approach. Surprisingly, whereas CXCL12 stimulated the formation of large receptor nanoclusters on CXCR4-expressing cells, reduced their diffusion coefficient, and increased the percentage of immobile particles at the cell membrane, this did not occur in cells expressing CXCR4 R334X alone or coexpressed with CXCR4. We also observed that CXCR4 R334X coexpression altered CXCL12-mediated CXCR4 nanoclustering, consistent with the known dominant characteristics of the mutant receptor detected in patients. In primary cells, WHIM receptors are always expressed in heterozygosis and are therefore coexpressed with wild-type CXCR4 (40). Analysis of FRET data confirmed the heterodimerization between CXCR4 R334X and CXCR4 (28), which is a possible mechanism to explain the observed dominant effect of WHIM mutant receptors. Receptor clustering increases cell sensitivity to external stimuli (41), and is also a means for efficient cell signal propagation (42), increasing the robustness of signaling systems (43). The present data suggest that the structural differences between CXCR4 and CXCR4 R334X affect the dynamics of these receptors at the cell surface and, consequently, the migratory behavior of the cells expressing them. The results also point to the relevance of the C-terminal tail of CXCR4 and of its coupling to β-arrestin1 for these processes. The actin cytoskeleton has a key role in regulating membrane diffusion, protein compartmentalization and clustering, and in controlling receptor signaling (14). Previous data from our group indicate that the actin cytoskeleton is also essential for ligand-mediated CXCR4 nanoclustering, as latrunculin A treatment interferes with CXCR4 nanoclustering and abrogates CXCL12-mediated cell migration (15). Analysis of phalloidin immunofluorescence images revealed clear differences between JKX4 / -R334X and control (JKX4 / -X4) cells in actin dynamics. In cells expressing CXCR4, CXCL12 treatment concentrated phalloidin staining in a unique cellular localization, the lamellipodium, whereas phalloidin staining was randomly distributed in cells expressing CXCR4 R334X , and several protrusions were detected around the cell. Notably, a comparable phenotype was observed when we compared phalloidin staining on CXCL12-stimulated CD3 + T cells from healthy donors and from patients with WHIM. β-Arrestins are involved in chemokine receptor desensitization and internalization (44), but they also have a scaffolding role for a number of signaling molecules (45, 46) and participate in actin reorganization and chemotaxis processes (28, 47, 48). In vivo studies have demonstrated the involvement of β-arrestins in tumor cell migration and metastasis (49), and in the recruitment of immune cells to sites of inflammation (50), although only β-arrestin1 participates in chemokine-mediated directed cell migration (35). There is also evidence supporting a role for β-arrestins in the spatial control of actin assembly events at the leading edge of primary leukocytes and cultured cells (19). β-Arrestins are required for PAR-2–dependent activation of the actin-binding protein cofilin, which binds and destabilizes actin filaments, promoting actin severing (51). Through a process involving Ser/Thr phosphorylation of specific residues by G protein-coupled receptor kinase, the C-tail of CXCR4 is essential for β-arrestin association and activation (52). Interestingly, a recent study reported that the phosphorylation pattern of the C-terminal end of G protein-coupled receptors defines not only the binding of β-arrestins, but also their spatial conformation and scaffolding role (53). CXCR4 R334X has a truncated C terminal, and cells expressing this receptor show defects not only in desensitization processes (54), but also in those pathways related to the scaffold role of β-arrestin. Our findings show that in cells expressing CXCR4, CXCL12 triggered rapid activation/inactivation of cofilin, allowing a correct balance to regulate actin dynamics. In contrast, in cells expressing CXCR4 R334X ,bothin homoand heterozygosis, CXCL12 triggered sustained cofilin PNAS 2022 Vol. 119 No. 21 e2119483119 https://doi.org/10.1073/pnas.2119483119 9of12 Downloaded from https://www.pnas.org by UNIVERSIDAD DE SEVILLA on April 17, 2023 from IP address 150.214.182.238.