Class 3 Semaphorin Signaling: The End of a Dogma

See allHide authors and affiliations

Science's STKE  24 May 2005:
Vol. 2005, Issue 285, pp. pe24
DOI: 10.1126/stke.2852005pe24


Semaphorins—a family of secreted, membrane-bound, and transmembrane proteins—play an important role in the development of various organs, as well as in axonal pathfinding, angiogenesis, tumorigenesis, and the immunological response. Neuropilins 1 and 2 (NRP1 and 2) are receptors for the class 3 secreted semaphorins (SEMA3s) but not for the other classes of semaphorins. NRPs are also coreceptors for vascular endothelial growth factor 165 (VEGF165), suggesting that SEMA3s could inhibit the VEGF165-VEGF receptor (VEGFR) pathway during angiogenesis. Until recently, it was believed that binding of SEMA3s to neuropilins was necessary to initiate signaling from plexins, the active players in semaphorin signal transduction. However, Gu and colleagues have recently described an exception: Their research suggests that SEMA3E signal transduction may be neuropilin independent. This Perspective focuses on this recent finding in the context of semaphorin signaling outside the nervous system.

The discovery of protein-protein interactions was a major breakthrough in the biological sciences. Decades later, hundreds of protein ligands and their receptors have been identified, including the class 3 semaphorins (SEMA3s), which bind to neuropilin (NRP)-plexin receptor complexes. The overwhelming importance of these semaphorins in development and cancer has mostly been attributed to the interaction of NRPs with vascular endothelial growth factor receptors (VEGFRs). However, recent evidence suggests that the plexins themselves may hold the key to the role of the SEMA3s in development, even without the help of the neuropilins. In a recent study, Gu and colleagues (1) describe for the first time the direct binding of a SEMA3 to a plexin, and demonstrate a critical role for this interaction in vascular development.

Semaphorins constitute a family of secreted, membrane-bound, and transmembrane proteins (2) with a wide range of normal functions—from axon guidance to organogenesis, angiogenesis, and the immunological response. The transmembrane semaphorins (classes 1, 4, 5, and 6) bind directly to the plexins—the group of proteins that mediate their signals (3). On the other hand, the secreted class 3 semaphorins require an initial interaction with neuropilins (NRP1 or 2) to transduce the signal to the plexins (4) (Fig. 1B). Thus, neuropilins are receptors only for semaphorins from class 3. In vertebrates, nine plexins (A1 to A4, B1 to B3, C1, and D1) have been characterized. Even if specific semaphorins could be matched to different subsets of plexins (5), the complexity of the signaling pathways used by these proteins would need to be further explored. For instance, the existence of a retrograde signal mediated by the transmembrane semaphorins underlines the difficulty of sorting out all of the pathways involved.

Fig. 1.

Signaling by class 3–secreted semaphorins. (A) By competing for binding to neuropilin 1 (NRP1), SEMA3A can antagonize VEGF165-induced motility in endothelial cells. [Model based in part on (8)] (B) SEMA3s usually bind to a receptor complex composed of NRPs and plexins, which transduce their effects through cytoplasmic effectors (not shown). Here, SEMA3F mediates repulsion of breast cancer cells. [Model based on (25, 36)] (C) New mechanism shown by Gu and colleagues (1): SEMA3E interacts directly with plexin-D1 (interaction domains are not known; shown as "?") to mediate repulsion of embryonic vascular cells. a1/a2, complement components C1r and C1s homology domains; b1/b2, coagulation factors VIII and V homology domains; cyto, cytoplasmic domain; HBD, heparin binding domain; Ig, immunoglobulin-like domain; MAM, Meprin-A5-μ domain; Nter, N terminus; PSI, plexin-semaphorin-integrin domain; SEMA, Sema domain.

Both NRPs and plexins participate in signaling complexes with various additional proteins. Two research teams have shown that NRP1 also acts as a coreceptor for placental growth factor 2 (PlGF2) and VEGF165 (6, 7). These findings led researchers to investigate the possibility that SEMA3s might compete for—and thereby inhibit—VEGF165 binding to NRPs, a process that is critical to developmental angiogenesis as well as tumorigenesis (8) (Fig. 1A). Semaphorin function also extends to the immune system, by binding with the receptors C-type lectin CD72, for SEMA4D, or the T cell immunoglobulin and mucin domain (Tim-2), for SEMA4A (9, 10) [for a review, see (11, 12)]. Furthermore, the hepatocyte growth factor receptor (HGFR), Met, and its homolog, Ron, interact with plexin-B1. This association results in the phosphorylation of both plexin-B1 and the HGFR following binding of the class 4 semaphorin SEMA4D. The consequent signal drives an invasive growth program in epithelial and COS-7 cells (13, 14) and stimulates angiogenesis by human umbilical vascular endothelial cells (HUVECs) (15, 16). Even an epidermal growth factor receptor, ErbB-2, can associate with plexin-B1 in transfected human embryonic kidney (HEK) 293 cells (17). However, SEMA4D-induced activation of plexin-B1 down-regulates the activity of the small guanosine triphosphatase (GTPase) R-Ras (18) (see below), unraveling the duality of semaphorin signaling, which, depending on the specific circumstances, can lead to opposite effects. Similarly, during cardiac morphogenesis, SEMA6D signals through plexin-A1 either in a complex with VEGFR2 to promote migration or in a complex with Off-track to inhibit it (19). SEMA6D is even known to signal backward through its intracellular domain by activating the Abl cytoplasmic kinase (20) (Table 1).

Table 1. Semaphorins and their receptors outside the nervous system. Semaphorins bind and signal through different but redundant receptor complexes. The resulting effects can vary depending on what receptor is used, even for the same semaphorin. SEMA, semaphorins; PAEC, porcine aortic endothelial cells; HUVECs, human umbilical vascular endothelial cells; HEK293, human embryonic kidney 293 cells. *Data concerning plexin-A1 are from (36).

Recent research has investigated signaling downstream of the plexins. Because the sequence of the plexin-B1 intracellular C terminus resembles that of the GAPs (GTPase activating proteins) for the Ras superfamily, one group of investigators focused on the ability of plexin-B1 to stimulate the intrinsic GTPase activity of R-Ras. Indeed, they found that plexin-B1 behaves like an R-Ras GAP following stimulation by its ligand, SEMA4D. Even though they described this molecular event as responsible for axon collapse, this function may not be restricted to neurons. They pointed out that another plexin, plexin-A1, can act analogously to inhibit R-Ras with SEMA3A.Unlike other Ras proteins, R-Ras does not use the mitogen-activated protein kinase (MAPK) pathway, but rather regulates integrin activation (21). Strikingly, two studies reported independently that SEMA3A and 3F and SEMA4D modulate integrin activation and function in HUVECs and NIH3T3 fibroblasts, respectively (22, 23). Moreover, SEMA3F inhibits integrin activation in the H157 lung cancer cell line (24). Whereas previous studies presented SEMA3s solely as inhibitors of VEGF165 (and the migration that it stimulates), these works reveal that semaphorins have potent and direct effects on integrins, a class of proteins with roles in diverse processes ranging from cardiovascular development to tumor metastasis. Consistent with these findings, a recent study (23) demonstrated a requirement for plexins, but not neuropilins or VEGFRs, in the regulation of adhesion by SEMA4D.

Gu, Yoshida, and co-workers have now investigated the details of signaling mediated by SEMA3E, a class 3 semaphorin about which relatively little is known. They focused on vascularization in developing somites, where SEMA3E mRNA is found in a specific pattern. Surprisingly, they report that both the binding and function of SEMA3E are independent of neuropilins, which were previously presumed necessary for SEMA3 activity. Instead, SEMA3E bound plexin-D1 with high affinity, as measured through Scatchard analysis, regardless of the presence of NRP1 or 2. In vivo, SEMA3E binds to the intersomitic regions of mouse embryos, where plexin-D1 is expressed. Embryonic zones engineered to ectopically overproduce SEMA3E repulse vascularization and SEMA3E collapses the lamellipodia of COS-7 cells transfected with plexin-D1. Inhibition of SEMA3E resulted in the development of disorganized intersomitic blood vessels, a phenotype similar to that of plexin-D1 knockout. SEMA3E might therefore function to sequester plexin-D1–positive vessels in the intersomitic areas. In contrast, neither NRP1sema−/−, which is an NRP1 mutant that is defective in binding SEMA3s but not VEGF165, nor NRP2 knockout exhibited the same phenotype. Thus, SEMA3E function in vivo is not related to the neuropilins.

Moreover, these data describe for the first time the direct binding of a class 3 semaphorin to a class D plexin. Not only does this research contradict the generally accepted "SEMA3–NRP–plexin-A" dogma and offer new opportunities for investigation of the molecular details of SEMA3 signaling, it also demonstrates that the role of SEMA3s in vascular patterning goes far beyond the expectations for a simple VEGF165 inhibitor. Here, SEMA3E repelled embryonic vessel migration through the plexin-D1 receptor alone. In a context where SEMA3E could not act by antagonizing VEGF165 binding to NRP1 (NRP1sema−/−), SEMA3E retained its full activity. Therefore, SEMA3E controls embryonic vascular patterning in a paracrine, VEGFR-NRP–independent and plexin-D1–dependent manner. In summary, these results strongly point toward a direct function of SEMA3E acting through a single plexin, an exception to the usual mechanisms for SEMA3 signaling (Fig. 1C).

Although the SEMA3E-plexinD1 interaction is unique in that SEMA3s are secreted ligands whose activity usually depends on the presence of neuropilins, their transmembrane counterparts (the SEMAs 1, 4, 5, and 6) already highlighted the pluripotency of semaphorin signaling in various cell types. However, over the past few years, we have minimized the role of semaphorins outside the nervous system as first and foremost activators of plexin signaling. These findings remind us that, more often than not, plexins mediate the critical semaphorin responses, even when in association with distinct growth factor receptors. For instance, plexin-B1 complexes with ErbB-2, as indicated above (17), but thereby enhances its own intrinsic capacity to activate RhoA, a previously known semaphorin cytoplasmic effector. Years after the first plexins were identified, we still do not fully understand their true function and specificity. Despite a growing number of studies, the tyrosine phosphorylation sites within the cytoplasmic domain are not yet mapped. Understanding how different semaphorins—or even the same ones—can exert such opposite effects may be of therapeutic interest, especially in cancer. Future studies will have to focus on the mechanisms through which plexins transduce semaphorin signals in tumors, with the hope that it would be possible to design targeted drugs aimed at specifically inhibiting protumorigenic signals triggered by plexins while simultaneously not affecting—or up-regulating—their antitumorigenic properties. Unearthing the details of the plexins’ biological complexity could greatly help in this quest.


  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.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.

Stay Connected to Science Signaling

Navigate This Article