Platelet Endothelial Cell Adhesion Molecule-1 (PECAM-1/CD31)

First printed in R&D Systems' 1997 Catalog.


Among the molecules known to (secondarily) activate the integrins are the chemokines and platelet-endothelial cell adhesion molecule-1 (PECAM-1, CD31 or endoCAM). Once a leukocyte binds to endothelium via an integrin-immunoglobulin (Ig) superfamily (IgSF) interaction, it "searches" for junctions between endothelial cells, first squeezing between these potential discontinuities, and then penetrating the underlying basement membrane to reach the tissues.1, 2

Figure 1. PECAM-1/CD31 in inflamation.

Structural Information

Human CD31 is a 130 kDa, type I (extracellular N-terminus) transmembrane glycoprotein that belongs to the cell adhesion molecule (CAM) or C2-like subgroup of the IgSFl.3-6 The mature molecule is 711 amino acid (aa) residues in length and contains a 574 aa residue extracellular region, a 19 aa residue transmembrane segment, and a 118 aa residue cytoplasmic tail.1, 6 In the extracellular region, there are nine potential N-linked glycosylation sites, and, with a predicted molecular weight of 80 kDa, it appears many of these sites are occupied. The most striking feature of the extracellular region is the presence of six Ig-homology units that resemble the C2 domains of the IgSF.6-8 Although they vary in number, the presence of these modules is a common feature of all IgSF adhesion molecules (ICAM-1, 2, 3 & VCAM-1).9-12 Unlike all other IgSF adhesion molecules, CD31 does not display a highly conserved integrin-binding motif that is found in the other adhesion molecules' first Ig-like domains.13, 14 Thus, its functional relationship to members of the integrin family appears to be fundamentally different from that of all other IgSF adhesion molecules. Differences between CD31 and the ICAMs/VCAM also exist in the organization of their genes. In all these molecules, a distinct exon exists that codes for each Ig-like domain. In the transmembrane and cytoplasmic segments, however, differences exist. Seven exons (numbers 10-16) code for the cytoplasmic tail of CD31 and one exon (number 9) codes for the transmembrane segment.15 This contrasts sharply with the genes for the ICAMs/VCAM where only one exon codes for both the cytoplasmic tail and the transmembrane segment.16-18 This marked difference in the PECAM gene is known to account for at least three alternatively spliced forms, two of which effect the cytoplasmic tail, and one which impacts the transmembrane region. The two cytoplasmic variants are believed to give rise to molecules that differ in the configuration of their ligand-binding domains. This suggests that the cytoplasmic tail may regulate ligand-specificity and binding.19, 20 In addition, alternative splicing of the transmembrane region is known to yield a soluble form of CD31.21

Mouse CD31 has also been isolated and found to be 63% identical at the aa level to human CD31.22 While monoclonal antibodies to CD31 are reportedly species specific,22 polyclonal antibodies to CD31 seem to show species cross-reactivity.23 It is currently unclear if human and mouse PECAM are active across species.13 Cells known to express CD31 include endothelial cells,3, 19, 24 basophils,25 neutrophils,26, 27 platelets,28 megakaryocytes,29, 30 CD34+ hematopoietic progenitors,31, 32 monocytes,26, 27 B cells,33 NK cells,32, 34, 35 gd T cells,34 CD4+CD45RA+ CD29lo (naive) T cells,27, 33 plus Kaposi's sarcoma (or endothelial) cells36 and macrophages from both normal and rheumatoid synovium.37 Cell lines known to express CD31 include U937 and THP-1 monocytes,26, 38 MIKALL lymphocytes,38 HL-60 promyelocytes26 and KG-1 myeloid cells.26


One molecule that has been positively identified as being a co- or counter-receptor for CD31 is CD31 itself.4, 14, 38, 39 Homophilic interactions are well established, and appear to be mediated by the first and second C2 domains of the CD31 molecule.14, 40, 41 Additional contributions from domains number 3, 5 and 6 have been proposed, but it is not clear if any one secondary site is more important than any other.38 Although domain number 6 appears not to be involved in direct CD31-CD31 binding, this area, in conjunction with the membrane-proximal cytoplasmic region, seems to modulate overall binding affinity.14, 40, 42

Consistent with other IgSF/CAMs, there are also reports of integrin binding to CD31. In particular, the alphavbeta3 (or CD51/CD61) integrin has been suggested to be a CD31 co-receptor.43, 44 Although the integrin is expressed on T cells, NK cells, mast cells and endothelium,43 it is somewhat unusual as an IgSF ligand because all other ICAM/VCAM receptors are either beta1, beta2 or beta7 integrin types.2 This may prove to be reasonable, since CD31 lacks a consensus integrin-binding motif. As with CD31 homophilic binding of CD31 to alphavbeta3 is believed to involve the first two C2 domains of CD31,44 and cytoplasmic variants of CD31 seem to affect its heterophilic binding ability.40 Finally, CD31 also binds to glycosaminoglycans (GAGs).45 This binding occurs via a consensus binding site in the second Ig-like domain and can be inhibited by heparin or chondroitin sulfate.45

Biological Functions

Most, if not all, activities associated with CD31 can be attributed to homophilic CD31-CD31 interactions. Among these are leukocyte extravasation, bone marrow hematopoiesis, and vascular development. During the events surrounding leukocyte extravasation, CD31 seems to be most important during the later stages. In particular, CD31 appears to facilitate the binding of various leukocyte-associated integrins to ICAM/VCAM IgSF members and assist peripheral blood leukocytes in their transit across endothelial barriers. Facilitation of CAM-IgSF binding appears to be due to a general upregulation of beta1 and beta2 integrin activity following leukocyte CD31-endothelial cell CD31 engagement.33, 46-48 For example, on T cells, CD31 ligation will upregulate the adhesive function of both beta1 integrin (which binds to VCAM) and beta2 integrin (which binds to ICAMs).2, 46 Further, on monocytes and neutrophils, beta2 integrin activity is increased following CD31 ligation.47 Finally, on NK cells, CD31 activation will upregulate beta2 integrin expression,35 and induce cell spreading and cytoplasmic rearrangement.49 This is consistent with reports suggesting that CD31 can transduce an intracellular signal, particularly in conjunction with PKC.5, 24, 48

It has been proposed that CD31 can regulate the actual extravasation of leukocytes into the extracellular space. First, CD31 is well known to be concentrated at endothelial cell borders where it engages in cell-to-cell homophilic binding.4, 19, 27, 48, 50 Second, by blocking CD31, transendothelial cell migration is inhibited21, 38, 52, 53 and cytokine downregulation of CD31 expression leads to a reduction in leukocyte migration through endothelial cell monolayers.53 This may seem counter-intuitive, since an increase in vascular permeability (due to decreased CD31-CD31 homophilic binding) could easily be envisioned to assist leukocyte migration between otherwise tightly joined endothelial cells.1, 54 In fact, however, it appears that the presence of the C1 and C2 domains of PECAM are necessary for leukocyte attachment to endothelial cell edges, and the C6 domain is needed for migration through the underlying basement membrane.41

During inflammation, inflammatory cytokines such as IL-1 induce endothelial cells to upregulate their selectins and secrete membrane-associated factors such as PAF (platelet-activating factor). Subsequent leukocyte tethering to endothelial cells, and their activation by PAF, initiates the adhesion process that is amplified by CD31 engagement. At endothelial cell junctions, leukocyte CD31/endothelial cell CD31 (via domain 6) interactions further stimulate an upregulation of beta1 and beta2 integrins and promote the release of neutrophil granule proteases that dissolve the basement membrane and assist the passage of neutrophils into the extracellular space.55 In addition to endothelial cell-leukocyte interactions, CD31 is also suggested to play a role in hematopoiesis27, 30, 32 and vascular development.19, 27, 48


  1. Bevilacqua, M.P. (1993) Annu. Rev. Immunol. 11:767.
  2. Imhof, B.A. & M. Dunon (1995) Adv. Immunol. 58:345.
  3. Newman, P.J. et al. (1990) Science 247:1219.
  4. Albelda, S.M. et al. (1991) J. Cell Biol. 114:1059.
  5. Zehnder, J. L. et al. (1992) J. Biol. Chem. 267:5243.
  6. Stockinger, H. et al. (1990) J. Immunol. 145:3889.
  7. Williams, A.F. (1987) Immunol. Today 8:298.
  8. Williams, A.F. & A.N. Barclay (1988) Annu. Rev. Immunol. 6:381.
  9. Simmons, D. et al. (1988) Nature 331:624.
  10. Staunton, D.E. et al. (1989) Nature 339:61.
  11. Fawcett, J. et al. (1992) Nature 360:481.
  12. Osborn, L. et al. (1989) Cell 59:1203.
  13. Jones, E.Y. et al. (1995) Nature 373:539.
  14. Sun, Q-H. et al. (1996) J. Biol. Chem. 271:11090.
  15. Kirschbaum, N.E. et al. (1994) Blood 84:4028.
  16. Voraberger, G. et al. (1991) J. Immunol. 147:2777.
  17. Xu, H. et al. (1992) J. Immunol. 149:2650.
  18. Cybulsky, M.I. et al. (1991) Proc. Natl. Acad. Sci. USA 88:7859.
  19. Baldwin, H.S. et al. (1994) Development 120:2539.
  20. Yan, H-C. et al. (1995) J. Biol. Chem. 270:23672.
  21. Goldberger, A. et al. (1994) J. Biol. Chem. 269:17183.
  22. Xie, Y. & W.A. Muller (1993) Proc. Natl. Acad. Sci. USA 90:5569.
  23. Murohara, T. et al. (1996) J. Immunol. 156:3550.
  24. Ferrero, E. et al. (1996) Cancer Res. 56:3211.
  25. Fureder, W. et al. (1994) Allergy 49:861.
  26. Ohto, H. et al. (1985) Blood 66:873.
  27. Watt, S.M. et al. (1995) Leuk. Lymphoma 17:229.
  28. Metzelaar, M.J. et al. (1991) Throm. Haemost. 66:700.
  29. Calapso, P. et al. (1992) Pathologica 84:215.
  30. Hagiwara, T. et al. (1996) Exp. Hematol. 24:690.
  31. Turner, M.L. et al. (1995) Stem Cells 13:311.
  32. Marsafy, S.E. et al. (1996) Leukemia 10:1340.
  33. Ashman, L.K. et al. (1991) Tissue Antigens 38:208.
  34. Piali, L. et al. (1994) Adv. Exp. Med. Biol. 355:147.
  35. Berman, M.E. et al. (1996) J. Immunol. 156:1515.
  36. Jones, R.R. et al. (1995) J. Clin. Pathol. 48:1011.
  37. Szekanecz, Z. et al. (1995) Clin. Immunol. Immunopathol. 76:180.
  38. Fawcett, J. et al. (1995) J. Cell Biol. 128:1229.
  39. Muller, W.A. et al. (1992) J. Exp. Med. 175:1409.
  40. Sun, J. et al. (1996) J. Biol. Chem. 271:18561.
  41. Liao, F. et al. (1995) J. Exp. Med. 182:1337.
  42. DeLisser, H.M. et al. (1994) J. Cell Biol. 124:195.
  43. Piali, L. et al. (1995) J. Cell Biol. 130:451.
  44. Buckley, C.D. et al. (1996) J. Cell Sci. 109:437.
  45. DeLisser, H.M. et al. (1993) J. Biol. Chem. 268:16037.
  46. Tanaka, Y. et al. (1992) J. Exp. Med. 176:245.
  47. Berman, M.E. & W.A. Muller (1995) J. Immunol. 154:299.
  48. DeLisser, H.M. et al. (1994) Immunol. Today 15:409.
  49. Poggi, A. et al. (1996) Eur. J. Immunol. 26:817.
  50. Muller, W.A. et al. (1989) J. Exp. Med. 170:399.
  51. Vaporciyan, A.A. et al. (1993) Science 262:1580.
  52. Bogen, S. et al. (1994) J. Exp. Med. 179:1059.
  53. Rival, Y. et al. (1996) J. Immunol. 157:1233.
  54. Ferrero, E. et al. (1995) FEBS Lett. 374:323.
  55. Wakelin, M.W. et al. (1996) J. Exp. Med. 184:229.