PerspectiveCell Migration

Signals on the Move: Chemokine Receptors and Organogenesis in Zebrafish

+ See all authors and affiliations

Science's STKE  21 Aug 2007:
Vol. 2007, Issue 400, pp. pe45
DOI: 10.1126/stke.4002007pe45


The chemokine SDF1 (stromal cell–derived factor 1) directs cell migration in many different contexts, ranging from embryogenesis to inflammation. SDF1a is the guidance cue for the zebrafish lateral line primordium, a tissue that moves along the flank of the embryo and deposits cells that form mechanosensory organs. The SDF1a receptor CXCR4b acts in cells at the leading edge of the primordium to direct its migration. Two new studies show that a second SDF1 receptor, CXCR7, is required only in the trailing cells of the primordium, and they explore how these two receptors orchestrate migration of the primordium. CXCR4b and CXCR7 are expressed in complementary domains, possibly through mutual repression in which each receptor inhibits expression of the other. These studies illustrate how the entire primordium can respond to a single signal, yet generate cell type–specific responses by using different receptors.

The orchestration of cell movements is essential for a vast array of processes, ranging from organogenesis to the immune response. In many instances, cell migration is directed by chemokines: small secreted proteins that signal through seven-pass transmembrane heterotrimeric guanine nucleotide–binding protein (G protein)–coupled receptors (GPCRs) to elicit various downstream responses (11). Chemokine and chemokine receptor signaling pairs were initially studied in the context of leukocyte recruitment during inflammatory response, but have since been shown to be involved in a wide range of biological processes.

Because of its role as a co-receptor for human immunodeficiency virus (HIV), CXCR4 is one of the best-studied chemokine receptors (1, 2). CXCR4 and its sole ligand, stromal cell–derived factor 1 (SDF1), together govern diverse aspects of angiogenesis, lymphocyte trafficking, and nervous system development (1, 36). A search for receptors with homology to conserved GPCR transmembrane domains uncovered a new chemokine receptor, CXCR7 (12). This receptor binds with high affinity to SDF1, which was initially posited to signal only through CXCR4 (7, 8). Initial studies of CXCR7 yielded conflicting results with regard to its specific function, as well as its potential interactions with CXCR4. One group found that CXCR7 could mediate SDF1-induced chemotaxis and that blocking both CXCR4 and CXCR7 leads to an additive inhibitory effect on T cell migration mediated by SDF1 (7). In contrast, another group found that CXCR7 did not stimulate cell migration or Ca+2 mobilization, two responses that are characteristically observed after the activation of C-X-C–motif GPCRs; this group instead proposed that CXCR7 regulates other cellular activities, including cell adhesion and survival (8). CXCR4 and CXCR7 may activate distinct downstream effectors; alternatively, CXCR7 could function as a decoy receptor or sink that controls the amount of SDF1 that is available to CXCR4 (8, 9). Two recent papers have explored how SDF1 and its two receptors orchestrate the development of cells within a coherent migrating tissue, the lateral line primordium in zebrafish (9, 10).

Transparent zebrafish embryos are an ideal system in which to investigate cell and tissue migration in vivo. SDF1-CXCR4 chemokine signaling is used in many contexts in the developing zebrafish embryo. Because of a genome duplication in the ray-fin fish lineage, the zebrafish has two genes encoding the CXCR4 receptor (cxcr4a and cxcr4b) and two for the SDF1 ligand (sdf1a and sdf1b) (13). SDF1-CXCR4 signaling functions in several distinct ways during zebrafish development. SDF1a acts as a short-range cue for primordial germ cells (PGCs), guiding individual CXCR4-expressing PGCs over long distances in the embryo to reach their eventual target in the gonad. During PGC migration, sdf1a expression changes dynamically to progressively attract the germ cells across the embryo (14). In contrast to germ cells, sensory neurons can respond to SDF1 signals at a distance. In this case, a posterior source of SDF1a and SDF1b acts as a long-range cue to attract anterior neurons to the position of the trigeminal sensory ganglia. After the neurons reach their destination, SDF1 acts as a short-range cue to retain these cells in the ganglion (15). In yet another example, SDF1a guides migration of a coherent group of cells that constitute the primordium of the lateral line system (16, 17).

The lateral line is a mechanosensory organ that senses water currents. Lateral line sensory organs, known as neuromasts, contain mechanosensitive hair cells similar to those of the ear (18). The lateral line primordium migrates from a placode posterior to the ear along the length of the fish; during this process, groups of cells periodically split off from the primordium to differentiate as neuromasts (Fig. 1) (19). The population of migrating cells is not uniform, but rather is polarized. Cells at the leading edge of the migrating primordium extend protrusions and display characteristics of actively pathfinding cells. In contrast, cells at the trailing edge do not extend processes, and some are periodically deposited in the wake of the migrating primordium (17). sdf1a is expressed in a stripe along the length of the fish and guides migration of the lateral line primordium (Fig. 1). Eliminating SDF1a specifically along the migration path misroutes the primordium toward other endogenous sources of the attractant, whereas in the absence of SDF1a, migration stalls (13, 16). cxcr4b is expressed predominantly at the leading edge of the lateral line primordium, and migration of the primordium progresses poorly in mutants that lack functional CXCR4b (cxcr4b mutants) (13). Analysis of chimeric embryos comprising wild-type and cxcr4b mutant cells showed that CXCR4b is required only in cells at the leading edge of the migrating primordium, even though migration of the entire primordium is disrupted in cxcr4b mutants (17). These results show that SDF1a and CXCR4b are essential for guidance but do not explain what controls the migration of cells at the trailing edge of the primordium.

Fig. 1.

(A) Development of the posterior lateral line system shown at about 35 hours after fertilization. The primordium (blue/red) migrates along a stripe of sdf1a expression (light blue) (16). Cells periodically split off from the trailing edge of the primordium and differentiate into neuromasts (red circles) (18). (B) Model of lateral line primordium guidance (9, 10, 17). As cells of the lateral line primordium first migrate away from the ganglion, they are exposed to high concentrations of SDF1a (light blue) and express cxcr4b (dark blue). At slightly later stages, trailing cells and cells left in the wake of the primordium express cxcr7 (red). Axons that innervate the neuromasts are towed by the primordium, and the axons in turn direct the migration of accompanying Schwann cells (2224).

Recent work has shown that the leading and trailing cells within the primordium express different SDF1a receptors. Two groups found that cxcr7 is expressed in the trailing cells of the primordium and that cxcr4b and cxcr7 are expressed in complementary domains (Fig. 1) (9, 10). Reducing CXCR7 function with antisense morpholino oligonucleotides impeded primordium migration but did not alter its direction. When CXCR7 was knocked down, trailing cells displayed protrusions, reminiscent of the actively pathfinding leading cells (10). Trailing cells were also apparently unable to bring up their rear edges, leading to an elongated primordium with cells of twice the normal length. In a subset of fish in which CXCR7 was knocked down, the primordium split so that the cells at the leading edge became uncoupled from those following. Migration failed to progress in either half of the split primordium, indicating that successful migration depends on the proper function of both leading and following cells. Whereas CXCR4b is required only in leading cells, CXCR7 is required only in trailing-edge cells (10). Wild-type cells can rescue a primordium lacking CXCR7 only if the cells are incorporated into the trailing end of the primordium. Strikingly, once trailing wild-type cells split off from the rescued primordium to form neuromasts, forward movement of the primordium immediately stops. Thus, CXCR7 is required in trailing cells for proper forward migration of the entire tissue (10).

Valentin et al. (10) propose that the two receptors act independently in leading and trailing cells to control migration of the entire primordium. By injecting CXCR7 morpholinos into cxcr4b mutants, Valentin et al. found that migration is more severely affected in embryos lacking both receptors than in embryos retaining the function of either CXCR4b or CXCR7. Moreover, the severe phenotype resulting from the loss of both receptors resembles that of mutants lacking sdf1a (10). Thus, these data suggest that the two receptors are independently required to coordinate the migratory responses of different populations of cells within the moving primordium. Dambly-Chaudiere et al. (9) reached a different conclusion from studies in which both receptors were inactivated simultaneously by morpholino injection. Morpholino experiments often involve complications from nonspecific effects and partial loss-of-function phenotypes, and the studies using the cxcr4b mutant may therefore be more reliable (10). The generation of a cxcr7 mutant will be required to definitively address the functional interactions between CXCR4b and CXCR7.

Dambly-Chaudiere et al. (9) propose that the differences between leading and trailing cells arise from a combination of asymmetric expression of sdf1a in the region through which the primordium initially migrates, differential affinities of CXCR4b and CXCR7 for SDF1a, and a mutual repression in which each receptor inhibits expression of the other. According to this view, as leading cells first encounter high levels of SDF1a, CXCR4b binds the ligand, leaving less available for trailing cells. The model further postulates that CXCR7 has higher affinity for SDF1a and can therefore respond to lower concentrations of the signal by repressing cxcr4b expression in trailing cells. In support of this model, biochemical studies of the mammalian proteins show that CXCR7 binds SDF1 with substantially greater affinity than does CXCR4 (7). Furthermore, morpholino injection experiments reported by this group (9) show that CXCR7 is required to repress cxcr4b expression in trailing cells, and vice versa, which is consistent with the proposed mutual antagonism between the receptors. In contrast, Valentin et al. (10) reported that cxcr4b expression is normal in embryos injected with cxcr7 morpholinos and that cxcr7 expression is normal in cxcr4b mutants, which would indicate that neither receptor is required to repress the other. Future work will address this discrepancy and determine whether antagonism between cxcr4b and cxcr7 is required to organize asymmetry within the migrating primordium.

The recent work on CXCR7 illustrates how the entire primordium tissue can respond to a single signal but generate cell type–specific responses by using different receptors. CXCR4b is required in leading cells for migration in response to SDF1a. CXCR7 in the trailing edge is necessary for forward progression of the trailing cells in response to SDF1a and perhaps for other trailing-cell functions such as neuromast development. This work raises many intriguing questions for further investigation. It will be interesting to determine whether CXCR4 and CXCR7 act through different signaling pathways to bring about different cellular responses. It will also be interesting to investigate additional interactions between leading and trailing cells, which may explain why the loss of SDF1a-responsiveness in one subpopulation impedes migration of the entire primordium. It is likely that CXCR4 and CXCR7 control the response to SDF1 in other contexts. For example, CXCR4a is required for fast-muscle development, and CXCR7 is expressed in slow-muscle precursors, suggesting that these receptors might elicit different responses to SDF1 in muscle development (20, 21). It will be particularly interesting to see whether CXCR4 and CXCR7 are used to coordinate SDF1 signaling in other tissues.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
View Abstract

Navigate This Article