M. Samson, INSERM U620, Université de Rennes 1, 2 avenue du Prof. Léon Bernard, 35043 Rennes cedex, France Fax: (+33) 02 23 23 47 94 Tel: (+33) 02 23 23 48 06 E-mail: email@example.com
Prokineticins are a novel family of secreted peptides with diverse regulatory roles, one of which is their capacity to modulate immunity in humans and in other species. Prokineticins are small peptides of 8 kDa that mediate their biological activities by signaling through two homologous G-protein-coupled receptors (prokineticin receptor 1 and prokineticin receptor 2). This family of peptides is characterized by a completely conserved N-terminal hexapeptide crucial for their bioactivities and a unique structural motif comprising five disulfide bonds. Prokineticins and their receptors are highly expressed in bone marrow, in peripheral circulating leukocytes, in inflamed tissues and in resident organ immune cells. Their structure, size, signaling and biological activities are reminiscent of the chemokine superfamily. In this review, emphasis is placed on the properties of prokineticins as cytokines and their role in the immune system.
Prokineticin 1 (PROK1) and prokineticin 2 (PROK2), otherwise known as endocrine gland vascular endothelial factor (EG-VEGF) and Bombina variegata 8 (Bv8), respectively, are a novel family of peptides that are highly conserved across species. Indeed, prokineticin-like peptides are present in invertebrates (crayfish, shrimp), vertebrates (black mamba snake, frogs, trout, fugu) and mammals (rodents, bull, humans) [1,2]. One of the first biological activities reported was the ability of the amphibian orthologue of PROK2 (Bv8) to elicit hyperalgesia in rats and to induce gastrointestinal motility of guinea-pig ileum . Subsequently, the human recombinant form of prokineticins was shown to have similar gastrointestinal motility activities in guinea-pigs [3,4]. Both studies confirm that prokineticins are structurally conserved across species and show highly conserved bioactivities. Not long after, it was demonstrated that PROK1 and PROK2 also had potent angiogenic properties independent of vascular endothelial growth factor . Since then, numerous other biological activities have been associated with prokineticins, such as angiogenesis, neurogenesis, ingestive behaviours and hormone release, gastrointestinal motility, circadian rhythms, pain sensation, and blood cell function and development [2,6–11]. Furthermore, the most elegant demonstration of the implication of prokineticins in a human disease has been brought by the identification that different point mutations in the genes encoding PROK2 or the Gprotein-coupled prokineticin receptor-2 (PROKR2) can lead to Kallmann syndrome [6,12,13].
This review aimed to retrace all the cytokine properties of prokineticins. This novel family of peptides share many common features with the chemokine superfamily, such as their small size (8 kDa), receptors [G-protein coupled receptors (GPCRs)], signaling mechanisms, and chemotactic and immunomodulatory activities. Therefore, in the present study we first described the gene, protein structure, signaling and biochemical properties of prokineticins, which showed similarity to some cytokines, notably chemokines and defensins. Second, we overviewed the cellular source of both prokineticins and their receptors throughout immune cells. Finally, their numerous activities as cytokines were overviewed.
Comparison of gene and protein structure of prokineticins and prokineticin receptors with chemokines
Ligands PROK1 and PROK2
The gene that encodes PROK1 (86 amino acids) is located on chromosome 1p21. It is composed of three exons, with no known alternative splicing product . The PROK2 gene maps to chromosome 3p21.1 and it is composed of four exons, which gives rise to two mature proteins: PROK2 (81 amino acids) (exons 1, 2 and 4) and a splice variant with a 21-amino acid insert called PROK2L (102 amino acids) (exons 1, 2, 3 and 4). PROK1 and PROK2 share approximately 44% amino acid identity . In contrast to PROK1 and PROK2, which are located on different chromosomes, the genes encoding chemokines are often on the same chromosome and their loci are physically very close (i.e. CXC chemokines are almost all located on chromosome 4q21) . Furthermore, the genetic loci of both prokineticins are not physically near any cytokine family genes and it thus appears unlikely that both prokineticin and cytokine genes could be involved in translocation phenomenon pathologies.
After signal peptide processing, the secreted peptide contains distinctive structural motifs that are highly conserved across species. One such motif is the N-terminal AVIT sequence and another comprises 10 conserved cysteine residues [1,9]. This N-terminal AVIT sequence is essential for the correct binding of receptors . However, it has not yet been identified if prokineticins can undergo in vivo proteolytic cleavage by extracellular proteases. As observed with chemokines, the N-terminal domain is also essential for receptor binding because cleavage of peptides from the N terminus by extracellular proteases regulates chemokine bioavailability by either increasing or decreasing the biological activity of the chemokine . Furthermore, prokineticins are highly basic. Indeed, PROK1 binds with high affinity to heparin–sepharose, and PROK2 and PROK2L are predicted to be highly basic with a pI of 8.85 and a pI of 10.68, respectively. Thus, prokineticin activity may also be regulated through the binding of extracellular components, such as sulphate proteoglycans . Interestingly, the members of the chemokine family are also highly basic, and their bioactivity is tightly regulated through interactions with heparin sulfates of the extracellular matrix . Chemokines contain four to six cysteine residues, and this is a marked structural difference from prokineticins, which contain 10 cysteine residues. The 10 cysteine residues form five disulphide bonds that confer a compact structure on the molecule, with N-and C-termini present on the surface. One side of the roughly ellipsoid protein has a positive net charge, whereas the opposite side is hydrophobic .
We performed a phylogenic study in order to compare the degree of similarity among prokineticins, chemokines and defensins, (which, similarly to prokineticins, also contain a high number of cysteine residues) [19–22]. Interestingly, the results of the pylogenic study revealed a higher similarity of amino acid sequence between defensins and prokineticins than with chemokines (Fig. 1A).
Receptors PROKR1 and PROKR2
There are two prokineticin receptors, named prokineticin receptor 1 (PROKR1) and prokineticin receptor 2 (PROKR2); they are two closely related seven-transmembrane GPCRs that belong to the family of the neuropeptide Y receptor. PROKR1 is located on chromosome 2p13.1 and PROKR2 is located on chromosome 20p12.3. No genes encoding cytokine receptors are located near prokineticin receptors. Interestingly, similarly to their ligands, both receptors are on different chromosomes, and this contrasts with chemokine receptors that are often located on the same chromosome, such as CCR1, CCR2, CCR3, CCR4, CCR5 that are located very closely on chromosome 3p21.3-24 .
Furthermore, all chemokine receptors are also G protein seven-transmembrane-coupled receptors, which is another common element shared between chemokines and prokineticins. However, they diverge in sequence with PROKRs, as can be seen in the phylogenetic tree (Fig. 1B), where the closest related chemokine receptor to PROKRs is XCR1.
Expression and regulation of prokineticins and prokineticin receptors in immune cells
Several studies have reported a differential expression of PROK1 and PROK2 within immune cells. Indeed, LeCoulter et al. showed, in 2004, that within human peripheral leukocytes only PROK2 was detected, and the highest expression was observed in bone marrow, neutrophils, dendritic cells and monocytes . In contrast, another study reported a significant expression of PROK1 in CD14+ cells, T cells and B cells; however, PROK2 expression was not evaluated . We measured the expression of PROK1 and PROK2 in fresh human monocytes obtained from 10 different healthy donors and our results consistently showed very high expression of PROK2 but undetectable expression of PROK1 (data not shown). Furthermore, we recently showed that the resident macrophage population of human liver (called Kupffer cells) was the specific source of PROK2 in liver, whereas PROK1 was weakly expressed .
Prokineticins have been associated differentially with inflammatory and immune tissues. Transcript analysis showed higher expression of PROK1 in rheumatoid arthritis synoviocytes and in Crohn’s disease compared with normal tissue . In situ localization in appendicitis and tonsillitis samples revealed high expression of PROK2 in infiltrating neutrophils . Interestingly, PROK1 transcripts were detected in tumor-infiltrating T lymphocytes in ovarian carcinoma .
In other species, immunohistochemical staining revealed that PROK1 was expressed in macrophages of bovine corpus luteum regression and follicular atresia [27,28]. In BALB/c mice, peritoneal macrophages were shown to express only PROK2 . However, in C57BL6 mice, PROK1 was shown to bind monocyte/macrophage cells in the spleen . Recently, Shojaei et al.  demonstrated in BALB/c nude mice implanted with several tumor cell lines that the bone marrow mononuclear cells subset enriched in PROK2 were CD11b+ GR1+ cells, consisting mainly of macrophage and neutrophil lineage cells. These results might reflect intraspecies variations.
Taken together, from the literature and from our own results it seems that PROK2 is preferentially expressed by cells from the monocyte–granulocyte lineage and PROK1 seems to be less specifically expressed by immune cells.
PROKR2 has been shown to be expressed more highly than PROKR1, in particular in CD8 cells, monocytes and neutrophils . In contrast, Dorsch et al. reported a low expression of PROKR2 in monocytes, but a substantial expression of both receptors in B cells . Recently, we studied the expression of PROKR1 and PROKR2 in fresh human monocytes from 10 different donors at the protein level using flow cytometry and we observed that both receptors were expressed on the monocyte surface (J. Monnier, V. Quillien, C. Piquet-Pellorce, C. Leberre, L. Preisser, H. Gascan & M. Samson, unpublished data). In addition, we showed that in liver hepatic cells both PROKR1 and PROKR2 mRNA were only expressed at high levels by Kupffer cells . In mice peritoneal macrophages a predominance of PROKR1 transcripts was found, and macrophage cells extracted from PROKR1 knockout mice were unable to respond to PROK2 .
In humans, it has not yet been determined which receptor is essential for the immunoactivities linked to PROK1 or PROK2. Future directions for research could come from the PROKR knockout mice that have been recently generated. Indeed, while PROKR1−/− mice have been obtained at the expected Mendelian rate without critical abnormalities, > 50% of PROKR2−/− mice died at an early neonatal stage and the surviving mice presented abnormalities similar to Kallmann syndrome . Challenging the PROKR1−/− and PROKR2−/− mice with various pathogens could be a very informative model for studying the role of prokineticin receptors in the immune system.
Very little is known about the regulation of prokineticin and prokineticin receptors within immune cells. However, very recently PROK2 was shown to be positively regulated in CD11b+ Gr1+ myeloid cells (consisting mainly of neutrophils and cells of the macrophage lineage), specifically by granulocyte colony-stimulating factor and not by any other cytokine tested [interleukin (IL)-4, IL-10, IL-13, monocyte chemotactic protein-1, macrophage inflammatory protein (MIP)-1α, MIP-1β, MIP-2, interferon-γ, keratinocyte-derived chemokine, fibroblast growth factor, vascular endothelial factor, PROK2, granulocyte–macrophage colony-stimulating growth factor (GM-CSF), granulocyte colony-stimulating factor, tumor necrosis factor-α (TNF-α), stromal cell-derived factor 1α and transforming growth factor-β] . Furthermore, we evaluated if the expression levels of PROK2 and PROK1 could be regulated according to the differentiation state of monocytic cells. To achieve this we compared the expression of PROK1 and PROK2 in monocytes and monocyte-derived macrophages, in undifferentiated and differentiated THP1 cells, and in undifferentiated and differentiated U937 cells. Our results showed that PROK1 was undetectable in all samples measured; however, PROK2 was highly expressed in all cells in the undifferentiated state and showed a large decrease in cells in the differentiated state (data not shown).
For the regulation of PROKR1 and PROKR2 in myeloid cells there is still much to be learned. However, we recently showed, by flow cytometry, that receptors PROKR1 and PROKR2 were present on the surface of monocytes, but that their expression was almost absent on the same monocytes derived into macrophages by GM-CSF or into dendritic cells by GM-CSF + IL-4 (J. Monnier, V. Quillien, C. Piquet-Pellorce, C. Leberre, L. Preisser, H. Gascan & M. Samson, unpublished data). Altogether, these data suggest that the differentiation status of monocytic cells may have an impact on the regulation of expression of prokineticin and prokineticin receptors.
Functions of prokineticins and prokineticin receptors in the immune system
Prokineticin receptor signaling in immune cells
The affinity of prokineticins for their receptors are in a similar range, with PROK2 showing a moderately higher affinity for both receptors: the Kd (nm) values for PROK1 and PROK2 binding to PROKR1 are 12.3 ± 4.2 and 1.4 ± 0.5, respectively, and the Kd (nm) values for PROK1 and PROK2 binding to PROKR2 are 1.8 ± 0.1 and 2.0 ± 0.7, respectively [9,32–34].
Based on the literature it seems that intracellular calcium mobilization and Gq protein activation is one of the major signaling mechanisms of prokineticin receptor activation [33–37]. However, other studies have shown that prokineticin receptors can also couple to Gi and Gs [26,32,33]. Only a few reports have studied the prokineticin signaling mechanisms in immune cells. It has been shown that human monocytes exposed to PROK2 induced extracellular signal-regulated kinase phosphorylation that was abolished by pertussis toxin, suggesting involvement of the Gi protein signaling pathway . Interestingly, in mouse macrophages, it seems that pertussis toxin was unable to block the actions of PROK2, but rather inhibition of the Gq protein pathway blocked the secretion of cytokines mediated by PROK2 . Recently, we tested the effect of inhibitors of Gi protein (pertussis toxin) and calcium [using the intracellular calcium chelator 1, 2-bis (2-aminophenoxy) ethane-N, N, N’, N’-tetraacetic acid (BAPTA)] on human monocytes for their ability to block PROK1-mediated CXCL8 secretion. Our results show that PROK1-mediated CXCL8 monocyte production was sensitive to pertussis toxin and BAPTA (J. Monnier, V. Quillien, C. Piquet-Pellorce, C. Leberre, L. Preisser, H. Gascan & M. Samson, unpublished data). Taken altogether, the results suggest that multiple pathways are involved in prokineticin signaling in monocytes, and that there might be some species-to-species variation.
Furthermore, another molecular mechanism that could influence signaling is GPCR dimerization. Indeed, it is now well described for chemokine receptors such as CCR2 to CCR5 or CXCR4 to CCR2 that dimerization can modulate signaling by negative or positive binding cooperativity . Thus, it would be very informative to determine if prokineticin receptors can homodimerize or heterodimerize, and how this would influence signaling.
Hematopoietic activity of prokineticins
Identification that a prokineticin-like peptide (Astakine) induced a strong hematopoietic response in vivo, as well as in vitro growth and differentiation of hematopoietic cells in invertebrates (shrimp and in crayfish), suggests a primitive role for prokineticins as hematopoietic cytokines (Table 1) . In vertebrates, the systemic expression of PROK1 or PROK2 by injection of adenovirus into nude mice resulted in a potent hematopoietic response. Indeed, the total leukocyte, neutrophil and monocyte count was increased, and the mouse spleen was enlarged as a result of the large number of immune cells produced (Table 1) .
Table 1. Summary of the effects induced by prokineticins on blood cells.
1. Action of PKs on blood cells
↑ Monocyte + neutrophil count ↑ CFU-G and CFU-M (Lecouter et al. )
↑ Monocyte + neutrophil count ↑ CFU-G and CFU-M (Lecouter et al. )
↑ Hematopoiesis in vivo and in vitro (Soderhall et al. )
Monocytes (Lecouter et al. )
Monocytes (Lecouter et al. ) Macrophages (Martucci et al. )
↑ Monocyte survival ↑ Macrophage differentiation (Dorsch et al. ; Lecouter et al. )
↑ CXCL8, CCL4, CXCL1, TNF-α, IL-1β (Monnier et al. ; Kisliouk et al. )
↑ CCL18, CCL20 (Monnier et al. )
↑ IL-1, ↓ IL-10 (Martucci et al. )
With LPS and IFN-γ
↑ IL-12 (Martucci et al. [29)]
7 days with PK1, then 24 h with LPS
↑ TNF-α, IL-12, ↓ IL-10 (Dorsch et al. )
In addition to their hematopoietic properties, PROK1 and PROK2 were also shown to be potent chemoattractants for human monocytes (Table 1). Indeed, migration of monocytes to prokineticins was observed at concentrations as low as 10−8m , and mouse macrophages migrated in response to even lower concentrations of PROK2 (10−12m) . This is another property that prokineticins share with chemokines, which are first and foremost characterized by their ability to induce the migration of immune cells at very low concentrations.
PROK1 has been shown to induce the differentiation of both human and mouse bone marrow cells into the monocyte/macrophage lineage (Table 1). In vitro, human and mouse hematopoietic stem cells treated with either PROK1 or PROK2 showed an increase in the number of granulocytic and monocytic colony-forming units . This was further observed in another study where human and mouse CD34+ cells showed a decrease in expression of CD34 and increase in expression of CD14 after treatment with PROK1. As well as inducing monocyte survival for 7 days, PROK1 differentiated monocytes into macrophage-like cells, as observed by the morphological changes induced in monocytes and the down-regulation of surface expression of B7-1, CD14, CXCR4 and CCR5 .
Cytokine and chemokine induction
By observing the cytokine signature induced in monocyte/macrophages by prokineticins, several studies have demonstrated that prokineticins function as pro-inflammatory mediators (Table 1) [24,27,29]. Indeed, monocytes treated for 7 days with PROK1 are primed to release TNF-α and IL-12, and to decrease IL-10 after treatment with lipopolysaccharide (LPS) . In mouse macrophages, PROK2, in conjunction with LPS, induced IL-1 and decreased IL-10 production, and co-stimulation of PROK2 with LPS plus interferon-γ induced IL-12 . Interestingly, those two studies only demonstrated an indirect response to PROK1 or PROK2 in differentiated monocytes.
However, other reports using fresh monocytes were able to show a direct response for cytokine production by prokineticins. In bovine monocytes, incubation with either PROK1 or PROK2 for 48 h increased the levels of integrin β2, elevated the number of adherent cells, and stimulated TNF-α mRNA expression, implying that prokineticins participate in the activation of bovine monocytes . We observed, in human monocytes treated with 1 μg·mL−1 of PROK1 for 8 h just after plating, that IL-1β and TNF-α transcripts were strongly induced (data not shown). Furthermore, we observed that monocytes treated with PROK1 for 24 h secreted the chemokines CXCL1, CXCL8 and CCL4. In addition, costimulation with LPS and PROK1 showed synergy for CCL18 and CCL20 production. Finally, we observed that CXCL1 and CXCL8 secretion after PROK1 induction is only observed in monocytes and not in monocyte-derived macrophages or dendritic cells, probably because of the decrease of receptors in macrophages and dendritic cells described previously (J. Monnier, V. Quillien, C. Piquet-Pellorce, C. Leberre, L. Preisser, H. Gascan & M. Samson, unpublished data). Taken together it seems that, in vitro, the differentiation state of monocytes has an impact on the ability of prokineticins alone to induce chemokines and cytokines.
In conclusion, this review attempted to present the many similar traits between prokineticins and cytokines. Indeed, prokineticins show greater similarity to the chemokine family than to members of the interleukin family. They share with chemokines many similar aspects: they are small secreted peptides, they are highly basic and bind heparane sulfates, they both contain cysteine residues, their N terminus is essential for proper signaling, they signal through GPCRs, they show multiple receptors with ligand cross-reactivity and they are potent chemoattractants. Furthermore, prokineticins also induce survival, differentiation and activation of the granulocytic and monocytic lineages, and they can be considered as pleiotropic chemokine-like cytokines. However, there is still much to be understood on how prokineticins modulate the innate and adaptive immune systems.
Justin Monnier was supported by a PhD fellowship from the Region Bretagne. Michel Samson was supported by the Institut National de la Santé et de la Recherche Médicale (INSERM).