Chemokines III

First printed in R&D Systems' 2001 Catalog.

This article is the third in a series of chemokine reviews that have appeared within the R&D Systems catalog.



Chemokines are a superfamily of mostly small, secreted proteins that function in leukocyte trafficking, recruiting, and recirculation. They also play a critical role in many pathophysiological processes such as allergic responses, infectious and autoimmune diseases, angiogenesis, inflammation, tumor growth, and hematopoietic development. Approximately 80 percent of these proteins have from 66 to 78 amino acids (aa) in their mature form. The remainder are larger with additional aa occurring upstream of the protein core or as part of an extended C-terminal segment. All chemokines signal through seven transmembrane domain G-protein coupled receptors. There are at least seventeen known chemokine receptors, and many of these receptors exhibit promiscuous binding properties whereby several different chemokines can signal through the same receptor.

Chemokines are divided into subfamilies based on conserved aa sequence motifs. Most family members have at least four conserved cysteine residues that form two intramolecular disulfide bonds. The subfamilies are defined by the position of the first two cysteine residues:

  • The alpha (α) subfamily, also called the CXC chemokines, have one aa separating the first two cysteine residues. This group can be further subdivided based on the presence or absence of a glu-leu-arg (ELR) aa motif immediately preceding the first cysteine residue. There are currently five CXC-specific receptors and they are designated CXCR1 to CXCR5. The ELR chemokines bind to CXCR2 and generally act as neutrophil chemoattractants and activators. The ELR- chemokines bind CXCR3 to -5 and act primarily on lymphocytes. At the time of this writing, 14 different human genes encoding CXC chemokines have been reported in the scientific literature with some additional diversity contributed by alternative splicing.
  • In the beta (ß) subfamily, also called the CC chemokines, the first two cysteines are adjacent to one another with no intervening aa. There are currently 24 distinct human ß subfamily members. The receptors for this group are designated CCR1 to CCR11. Target cells for different CC family members include most types of leukocytes.
  • There are two known proteins with chemokine homology that fall outside of the α and ß subfamilies. Lymphotactin is the lone member of the gamma (γ) class (C chemokine) which has lost the first and third cysteines. The lymphotactin receptor is designated XCR1. Fractalkine, the only known member of the delta (δ) class (CXC chemokine), has three intervening aa between the first two cysteine residues. This molecule is unique among chemokines in that it is a transmembrane protein with the N-terminal chemokine domain fused to a long mucin-like stalk. The fractalkine receptor is known as CXCR1.

A variety of approaches have been used to identify chemokines. The earliest discoveries of chemokines were made as a result of their biological activity or through studies that sought to identify proteins that are upregulated following cell activation or differentially expressed in selected cell types. Most of the recently reported chemokines, however, were identified through bioinformatics. EST (Expressed Sequence Tags) databases contain the sequences of a large number of cDNA fragments from a variety of tissues and organisms. Translation of ESTs can provide partial aa sequences of the proteome. Because the chemokines are comparatively small and contain signature aa motifs, many novel family members have been identified through searches of EST databases. The availability of genomic databases will no doubt lead to the discovery of additional chemokines, particularly those that display very restricted expression patterns.

Because of the large number of chemokines and rapid progress within the field, many chemokines have been reported by different research groups and, as a result, given multiple names. This situation has caused considerable confusion leading to an effort to create a unified nomenclature. The remainder of this article will focus only on newly described chemokines (i.e., identified by the new nomenclature). More comprehensive reviews of chemokines in general can be found in references 1-4.

Fig. 1. CCL23 plays a role in bone formation. Osteoblasts (cells that replace bone) secrete high levels of CCL23 which is a potent chemoattractant for precursors of osteoclasts (cells that resorb bone).


CCL23 is a 99 aa CC chemokine originally reported under the names CKß8 and MPIF-1 (myeloid progenitor inhibitory factor 1). It was isolated from human systems and no mouse equivalent has been described. The most closely related chemokines include human MIP-1δ, mouse C10, and mouse MIP-1γ. All of these chemokines share the unusual features of having a relatively long N-terminal segment and possessing six cysteine residues rather than the customary four. CCL23 is expressed in a variety of tissues and acts as a chemoattractant for monocytes, dendritic cells, and resting T cells but not activated T cells. It also inhibits colony formation by myeloid progenitors. Desensitization studies measuring calcium fluxes indicate that CCL23 reacts with CCR1 and possibly other receptors.5-8

CCL23 also plays a role in bone formation. Two cell types are responsible for bone remodeling; osteoclasts resorb bone and osteoblasts replace bone. CCL23 is a potent chemoattractant for osteoclast precursors, but not for fully differentiated osteoclasts or for osteoblasts.9 This chemokine is highly expressed in bone tissue with the highest levels of mRNA present in osteoblasts.

Recent reports indicate that multiple forms of CCL23 arise from alternative splicing and post-translational processing. A variant designated CKß8-1 is encoded by an alternatively spliced transcript that leads to the insertion of 17 aa within the codon for aa 25 of Ckß8. These two forms appear to be quite similar in their expression patterns and functional activities.

CCL23 expressed in a baculovirus system is proteolytically truncated in the absence of protease inhibitors. Surprisingly, one of the truncated forms lacking the first 24 aa is two orders of magnitude more potent than the full-length form in chemotaxis and calcium mobilization assays.10 These results suggest that CCL23 and possibly other chemokines having an extended N-terminal segment may be secreted in a partially active pro form and further activated by proteolytic cleavage.


This CC chemokine is found in human and mouse and was discovered independently by several laboratories. It was identified through searches of EST databases for chemokine homologies and reported initially as CTACK (cutaneous T cell-attracting chemokine), ALP (named for its N-terminal peptide sequence), or ILC (IL-11 Ra-locus chemokine because the gene is adjacent to the IL-11 Ra gene).13 It was also isolated from a mouse embryonic stem cell library and reported as ESkine.14

The CCL27 genes are found on syntenic regions of human chromosome 9p13 and mouse chromosome 4 near the genes for two other CC chemokines, 6Ckine and MIP-3ß. The mature human and mouse proteins share 84 percent aa identity, but have no more than 20 percent identity compared to other human CC chemokines.11, 13 CCL27 is most closely related to a CC chemokine from the molluscum contagiosum virus.

CCL27 is reported to exist as two splicing variants. One form encodes a classical secreted chemokine with a cleavable signal peptide. This form is expressed strongly and selectively in the skin.11, 14 Cultured keratinocyes demonstrate constitutive mRNA expression that is upregulated 8 to 30 fold upon treatment with a combination of the proinflammatory cytokines TNF-α and IL-1ß. Expression is not observed, however, in other skin-derived cells such as melanocytes, dermal fibroblasts, and γδ T cells. This chemokine appears to function by attracting T cells to the skin. Cutaneous lymphocyte-associated antigen (CLA) is expressed on a subset of circulating memory T cells and participates in the cutaneous localization of these cells via its interaction with the vascular ligand E-selectin.15, 16 In chemotaxis assays, CCL27 attracts a subset of the CD4+ CLA memory T cells (2-24 percent in different experiments) but fails to attract CLA-CD4+ or CD8 T cells. CCL27 interacts with an orphan receptor, G-protein-coupled receptor 2, which has been tentatively designated CCR10.

A second form of this protein is generated by use of an alternative exon at the 5' end of the mRNA. In this alternative form, designated PESKY, the signal peptide is replaced by a 40 aa segment that appears to function as a nuclear localization signal. Transfection of several cell lines with PESKY fused to green fluorescent protein has confirmed the nuclear localization of this protein. This transcript shows an expression pattern that is very distinct from the secreted form of CCL27 with high levels of expression in brain and testis and weak expression in liver and kidney. The function of this nuclear form is unknown.


CCL28 is a novel CC chemokine that was recently identified from a human EST database; the mouse homologue was cloned from a kidney cDNA library.19 The mature protein contains 105 aa, six cysteine residues, and an extended C-terminal segment. The human and mouse proteins share 63 percent aa identity. The most closely related chemokine is CCL27, which shares 28 percent aa identity with CCL28.

CCL28 elicited calcium mobilization in CCR10 transfectants, and CCL28 desensitized the CCL27-induced calcium flux in these transfectants.19 These results indicate that CCL28 and CCL27 both react with the CCR10 receptor. Recombinant human CCL28 attracted resting CD4 T cells, but activated T cells were not responsive.

Expression of mRNA was observed in a variety of organs but was less abundant in lymphoid or hematopoietic cells. Highest levels of mRNA expression were observed in colon and inflamed bowel tissue. Immunohistochemical data indicate that CCL28 is produced by epithelial cells. CCR10 is strongly expressed in normal gut samples suggesting a role for CCL28 in immunological surveillance within the gut.


SDF-1α and SDF-1ß are CXC chemokines encoded by alternatively spliced mRNAs. The mature a and ß forms differ only in that the ß form has four additional aa at its C-terminus. These proteins are highly conserved between species. Most functional studies have been performed with SDF-1α and suggest a variety of roles for this molecule. It is necessary for normal development of B cells and brain. It is a potent chemoattractant for CD34 bone marrow progenitor cells and dendritic cells. It also appears to play a role in trafficking and adhesion of lymphocytes and megakaryocytes. These chemokines have been known for some time and have been reviewed elsewhere.20

An additional alternatively-spliced product from rat, designated SDF-1?, was reported recently.21 The SDF-1γ mRNA is similar to the SDF-1ß message but with an additional exon inserted near the C-terminal end of the coding region. The four aa of SDF-1ß that are normally appended to the C-terminus of SDF-1α are replaced in SDF-1γ by a 30 aa segment containing 17 positively charged residues. The SDF-1ßand ? transcripts display different patterns of expression in a number of tissues. They are also reciprocally expressed in developing rat brain. SDF-1ß is expressed in embryonic and neonatal brain, whereas SDF-1? is expressed in adult brain. The function of this variant is still unknown.

A unique chemotactic activity was recently reported for SDF-1 in which subpopulations of T cells were attracted by SDF-1α concentrations of 100 ng/mL but repelled by concentrations of 1 µg/mL. 22 The higher concentration that elicited repulsion is comparable to the concentration of SDF-1 that occurs in the bone marrow. Inhibitor studies reveal that migration in both directions requires the CXCR4 receptor, G-proteins, and phosphatidylinositol 3-kinase. However, tyrosine kinase inhibitors block chemoattraction and have no affect on chemorepulsion, whereas a cAMP agonist inhibits chemorepulsion but does not affect chemoattraction.


Two groups using very different approaches identified this mouse ELR+ CXC chemokine. In an attempt to isolate the mouse homologue of MIP-3a, mouse cDNA libraries were screened with human MIP-3α.23 Positive clones were confined to libraries derived from lung tissue, and thus the chemokine was designated lungkine. WECHE (for Weird CHEmokine) was identified as a differentially expressed transcript in a comparison of two embryonic cell lines having different abilities to maintain hematopoietic stem cells in a pluripotent state.24

The CXCL15 cDNA predicts a 166 aa protein with a 25 aa signal peptide.23,24 The 141 aa mature protein has an unusually long C-terminal tail. Within the 70 aa chemokine domain, it shares about 40 percent aa identity with the mouse ELR CXC chemokines KC, LIX, and MIP-2. Expression is very restricted with high levels of mRNA in adult lung epithelial cells and much weaker expression in heart and fetal lung. A two-fold up-regulation of mRNA can be observed in inflamed lung tissue. The protein is released into the airway and functions as a neutrophil chemoattractant, suggesting that CXCL15 is involved in the homing of neutrophils in inflamed lungs.23

This chemokine may also regulate hematopoiesis. Immunohistochemistry of mouse embryonic stages reveals expression in the dorsal aorta and yolk sac, sites where hematopoietic stem cells are maintained.24 Expression is also observed in fetal liver, bladder, and umbilical vein. The protein functions as a chemoattractant for hematopoietic progenitor cells. It also inhibits proliferation of hematopoietic progenitors and blocks development of the erythroid lineage.


CXCL14 is an ELR- CXC chemokine that was initially named BRAK named because it was identified from human breast and kidney derived ESTs.25 The mouse homologue, which shares 97 percent aa identity with the human form, was also identified from EST databases. The most closely related chemokines, MIP-2α and MIP-27β share about 30 percent aa identity with CXCL14.

Northern blot studies demonstrate expression of CXCL14 mRNA in a variety of non-lymphoid tissues although there is little consistency between the observed expression in the human and mouse systems.25-27 In several cases, expression in normal tissue is down-regulated in the corresponding tumor tissue. Although absent in unstimulated peripheral blood mononuclear cells, CXCL14 is expressed at high levels in infiltrating inflammatory lymphocytes in nearly all cancers examined.27 Expression is also observed in B cells and monocytes following lipopolysaccharide activation.27 Direct binding to and chemoattractant properties have been demonstrated for B cells and monocytes, but not T cells.26 The loss of expression in tumors and the presence of CXCL14 in infiltrating lymphocytes suggest that this protein may play a role in host-tumor interactions.


  1. Sallusto, F. et al. (2000) Annu. Rev. Immunol. 18:593.
  2. Rossi, D. & A. Zlotnik (2000) Annu. Rev. Immunol. 18:217.
  3. Murdoch, C. & A. Finn (2000) Blood 95:3032.
  4. Zlotnik, A. & O. Yoshie (2000) Immunity 12:121.
  5. Forssmann, U. et al. (1997) FEBS Lett. 408:211.
  6. Patel, V. et al. (1997) J. Exp. Med. 185:1163.
  7. Youn, B. et al. (1998) Blood 91:3118.
  8. Nardelli, B. et al. (1999) J. Immunol. 162:435.
  9. Votta, B. et al. (2000) J. Cell Physiol. 183:196.
  10. Macphee, C. et al. (1998) J. Immunol. 161:6273.
  11. Morales, J. et al. (1999) Proc. Natl. Acad. Sci. USA 96:14470.
  12. Hromas, R. et al. (1999) Biochem. Biophys. Res. Comm. 258:737.
  13. Ishikawa-Mochizuki, I. et al. (1999) FEBS Lett. 460:544.
  14. Baird, J. et al. (1999) J. Biol. Chem. 274:33496.
  15. Berg, E. et al. (1991) J. Exp. Med. 174:1461.
  16. Picker, L. et al. (1991) Nature 349:796.
  17. Homey, B. et al. (2000) J. Immunol. 164:3465.
  18. Jarmin, D. et al. (2000) J. Immunol. 164:3460.
  19. Wang, W. et al. (2000) J. Biol. Chem. 275:22313.
  20. Zlotnik, A. et al. (1999) Crit. Rev. Immunol. 19:1.
  21. Gleichmann, M. et al. (2000) Eur. J. Neurosci. 12:1857.
  22. Poznansky, M. et al. (2000) Nature Med. 5:543.
  23. Rossi, D. et al. (1999) J. Immunol. 162:5490.
  24. Ohneda, O. et al. (2000) Immunity 12:141.
  25. Hromas, R. et al. (1999) Biochem. Biophys. Res. Comm. 255:703.
  26. Sleeman, M. et al. (2000) Int. Immunol. 12:677.
  27. Frederick, M. et al. (2000) Amer. J. Pathol. 156:1937.