Abstract
We have previously shown that the human developing pancreas, as a tissue under construction and remodeling, is composed of epithelial ducts and differentiated endocrine cells surrounded by mesenchyme. The physiologic importance of resident tissue leukocytes, expected to enter through the mesenchyme in remodeling functions, prompted us to investigate human developing pancreases for the presence of leukocyte lineages and for expression of cytokines and receptors involved in their recruitment and differentiation. Immunohistochemistry studies were performed on 69 human, paraffin-embedded pancreases at 6–12 weeks of development (WD). Cytokines and receptor transcripts were monitored by reverse transcription (RT)–PCR, by immunohistochemistry when antibodies were available or by in situ hybridization (ISH). We show that numerous cells expressing CD45RA, HLADR and CD68 antigens are cellular components of the mesenchyme of all the pancreases at 6–12 WD. So-called constitutive chemokines (SLC (CCL19), stromal-derived factor 1 (SDF1) (CXCL12)) and a macrophage-specific growth/survival factor, colony-stimulating factor 1 (CSF1), were detected in epithelial duct cells. Both epithelial and mesenchymal cells expressed chemokine receptors, suggesting a role in leukocyte recruitment and possibly in early pancreatic development. In conclusion, we demonstrated the presence of CD45RA resident leukocyte-derived lineages, mostly macrophages, in the early human pancreatic mesenchyme. These cells may have migrated in the tissue through the vascular system, attracted by constitutively expressed chemokines, and locally surviving through CSF1 signaling. The role of macrophages in epithelium/mesenchyme interaction-mediated pancreatic development remains to be demonstrated.
Introduction
The pancreas develops from the fusion of distinct, endoderm-derived, dorsal and ventral diverticula (Slack 1995). In man, by day 35 (5 weeks) of development, the ventral pancreatic bud begins to migrate backward; it comes into contact and fuses with the dorsal pancreatic bud during week 6 of development (WD) (O’Rahilly & Müller 1987, Moore et al. 1995). The dorsal bud gives rise to most of the head, body and tail of the future pancreas, whereas the ventral bud gives rise to the lower part of the head. We previously showed that, at 6–11 WD, the human embryonic pancreas is a branched organ, a complex structure containing epithelial ducts surrounded by mesenchyme (Polak et al. 2000). Cells of the exocrine and endocrine pancreas have a common precursor presumably located in the epithelial ducts. Early pancreas differentiation and growth are the result of epithelial/mesenchyme interactions. Signals from the mesenchyme that surrounds the embryonic pancreatic epithelium regulate the proliferation of immature pancreatic epithelial cells and their differentiation into endocrine or exocrine tissue both in murine models and in man (Miralles et al. 1998, Scharfmann 2000). Development of mouse pancreas has been extensively investigated, but human pancreas development is still poorly understood. We have also shown that the human endocrine cell population is detectable at 8 WD in or close to central ducts and expands at a brisk pace at 9–11 WD in a centrifuge manner toward the peripheral mesenchyme. This endocrine expansion results from a high proliferative activity probably associated with neogenesis (Polak et al. 2000). The epithelial precursor pool is greatly expanded, as the epithelial proliferation rate is very high at 6–11 WD. This also may contribute to the early endocrine cell expansion observed (Polak et al. 2000). Moreover, when grafted into immunodeficient mice, human pancreas at 6–16 WD keeps its ability to develop into a functional endocrine tissue (Usadel et al. 1980, Lafferty & Hao 1993, Beattie et al. 1997, Castaing et al. 2001). Peripheral mesenchyme and/or its secreted factors seem to have a stimulating effect on the proliferation of the neighboring epithelial cells, as we found peripheral ducts to proliferate 2–3 times more than the central ones (Polak et al. 2000).
The potential usefulness of embryonic and fetal human pancreas as a source of insulin cells for diabetes mellitus cell therapy implies the need to improve our knowledge of early human pancreatic development. Therefore, particular attention has to be given to the mesenchyme cellular components and epithelial/mesenchyme interactions, focusing on the presence of otherwise motile leukocytes.
The present study had three objectives: 1) to monitor the presence of CD45RA-cells in the mesenchyme of the human developing pancreas; 2) to characterize their phenotype and determine whether they are macrophages, leukocytes otherwise known to exert physiologic remodeling functions (Horton 1996), or auto- or allo-reactive, T-cell-stimulating dendritic cells; 3) to investigate, in addition to the receptors of the chemokines involved in tissue leukocyte recruitment, the putative locally induced differentiation by specific growth and survival factor expression.
Materials and Methods
Tissue collection, preparation and histologic monitoring
Human pancreases (n=69) and other organs (thyroid, liver, gastrointestinal tract, lung, kidney and adrenal glands) were dissected from embryonic and fetal tissue fragments obtained immediately after abortions performed at 6–11 WD, as is permitted by French law.
The late fetal (n=5), postnatal (n=2) and adult (n=2) pancreases were obtained from the Pathology Department at the Robert Debré Hospital, Paris, France.
Specimens for immunohistochemistry and in situ hybridization (ISH) were formalin-fixed and embedded in paraffin, additional specimens for immunohistochemistry were embedded in Tissue Teck OCT (Sakura, Finetechnical, CA, USA) and then frozen, and specimens for RNA extraction were immediately frozen.
Slides sampled at regular intervals from all the organs studied were stained with hematoxylin-eosin, and none showed abnormal inflammatory infiltrates or developmental anomalies.
Immunohistochemistry
Sectioning
All paraffin-embedded and OCT-embedded pancreases were cut from the dorsal to the ventral aspect into 4–6 μm coronal sections. Six to ten sections through each pancreas were selected for each antibody.
Selection of antibodies
We chose antibodies known to bind to cells of the leukocyte lineages and to chemokines. We looked for leukocyte-derived lineage cells (CD45RA), namely, potentially antigen-presenting cells, including mature dendritic cells (which express HLADR (Gosh et al. 1984), CD1a, CD83 or CD11c), immature dendritic cells (CD80), macrophages (CD68), T cells (CD3), B cells (CD20, CD79) and endothelial cells as well as progenitor hematopoietic cell (CD34) markers.
Expression of the constitutively expressed chemokines stromal-derived factor-1 (SDF1), (CXCL12) and 6 Ckine (SLC, CCL19) was examined (Pablos et al. 1999). To determine whether immune cells were located within or outside the blood vessel lumen, the antibody to CD34 was used in double staining with two other antibodies (HLADR or CD45RA).
Staining
Paraffin-embedded and frozen sections were processed as previously described (Polak et al. 2000).
Control experiments involved parallel staining by the secondary reagents (no primary antibodies), and parallel staining by only the chromogen (no primary antibodies or secondary reagents). They were performed on pancreases and control tissues. All the controls were negative. The antibodies, dilutions and controls are listed in Table 2.
Quantitative data
Leukocyte lineage cells were counted on representative sections per pancreas. The pancreatic surface of each CD45RA-stained section was quantified with a Leica DMRB microscope equipped with a color video camera connected to a Quantimet 500 IW computer (screen magnification × 10), as previously described (Polak et al. 2000). Results were expressed as CD45RA-positive cell number per mm2 of pancreatic surface.
TUNEL analysis
Apoptosis was studied by TUNEL (TdT-mediated Bio 11 dUTP Nick End Labeling coupled to digoxigenin), using the ApopTag kit (Q Biogene, Irvine, CA, USA). Cells were counted and results expressed as apoptotic cell number per mm2 of pancreatic surface.
Monitoring of chemokines, macrophage growth factors (CSF1 and GM–CSF) and their receptor transcripts
RT–PCR studies
Transcripts of the constitutively expressed chemokines SDF1, 6 Ckine, MIP3-beta (CCL20), and the inflammatory induced chemokine MIP3-alpha (CCL21), which are involved in chemotaxis and/or migration-maturation of lymphocytes and dendritic cells, were studied by reverse transcription and polymerase chain reaction (RT–PCR). CSF1 (colony-stimulating factor 1) and GM–CSF (granulocyte/macrophage colony-stimulating factor), which are known to be the most important growth factors involved in macrophage proliferation and differentiation, were also studied by RT–PCR. Transcripts of their receptors, CXCR4 (for SDF1), CCR7 (for 6 Ckine and MIP3-beta), CCR6 (for MIP3-alpha), CSF1R and GMCSFR were detected by RT–PCR and by ISH for CXCR4, CCR7, CSF1R and GMCSFR. To this end, total RNA was extracted from 34 human pancreases and from control tissues (Tables 1 and 3) by standard methods (TRIzol RNA extraction kit) and reverse-transcribed (Superscript II reverse transcriptase, Invitrogen Life Technologies). The cDNA of the chemokines, CSF1, GM–CSF and their receptors was amplified by PCR, using the primers listed in Table 3. Aliquots of each sample were amplified with cyclophilin primers as the cDNA loading control. The specificity was assessed by sequencing the first 300 bp of each PCR product. The results for mRNA expression are qualitative because the amount of cDNA from each pancreas was too small to allow quantitation, as pancreases were not pooled by age. The annealing temperature was 52 °C for CSF1, 56 °C for CSF1R and 54 °C for the other primers.
In situ hybridization (ISH)
Digoxigenin-UTP-labeled anti-sense CXCR4 riboprobes were synthesized by in vitro transcription of the CXCR4 cDNA cloned in pcDNA3 plasmid (a generous gift from F Arenzana-Seisdedos, Pasteur Institute, France). Linearized plasmids were transcribed in vitro with SP6 polymerase according to the manufacturer’s instructions (Boehringer Mannheim). The full-length antisense CXCR4 riboprobe was 600 bp.
CCR7 sense and antisense probes were synthesized from the 800 bp PCR product (Table 3), inserted in a pGEM Easy Vector and cloned in competent bacteria. CSF1, GMCSF, CSF1 and GMCSF receptor sense and antisense probes were synthesized from PCR products (Table 3), inserted in a pGEM Easy Vector and cloned in competent bacteria. Specificity was assessed by sequencing the first 300 bp of each cDNA insert.
ISH was performed on 6-μm-thick, paraffin-embedded sections mounted on Super Frost Plus slides (Menzel-Glaser, Braunschweig, Germany) and dried at 37 °C overnight. Slides were permeabilized with pepsin 0.35 in 0.04 mol/l HCl, washed, and then prehybridized with SSC 4X and 50% formamide for 1 h, at 50 °C for CXCR4, 42 °C for CCR7 and 45 C°for CSF1, GMCSF, CSF1 and GMCSF receptors. Then, the probe was heated for 5 min at 80 °C and incubated overnight, at 50 °C for CXCR4, 42 °C for CCR7 and 45 °C for CSF1, GMCSF, CSF1 and GMCSF receptors. After several washes, the slides were incubated overnight at 4 °C with the antidig-oxigenin antibody coupled with alkaline phosphatase 1/500 in Tris–HCl 100 mM and then stained blue with 4-nitro blue tetrazolium chloride (NBT)–X-phosphate/5-bromo-4-chloro-3-indolyl-phosphate (BCIP) chromogen. To localize both the expression of CXCR4 and CCR7 mRNA and the CD45RA-positive cells, immunofluorescence was performed after ISH. Control experiments involved parallel staining with CCR7, CSF1, GMCSF, CSF1 and GM–CSF receptor sense probes.
Results
Identification and characterization of leukocyte lineage cells within the extravascular space of the early human pancreas
Presence of leukocytic lineage cells in the early human pancreas
The human fetal pancreas at 6–12 WD is composed of epithelial duct cells in a loose connective tissue surrounded by mesenchyme (Fig. 1A, E and F). Endocrine cells become detectable at 8 WD as isolated cells, which subsequently aggregate into clusters (Polak et al. 2000). We found numerous leukocytes expressing CD45RA, HLADR or CD68 in the 29 human pancreases studied, of which the youngest were 6 WD (Fig. 1). We counted a mean per pancreas of 53 cells at 6 WD, 150 at 8 WD, 273 at 9 WD, 292 at 10 WD and 310 at 11 WD. When related to the pancreatic surface, the CD45RA-positive leukocyte number was 14 ± 0 at 7.5 WD, 21 ± 4 at 8–8.5 WD and 57 ± 37 at 9–11 WD. These results showed that leukocytes were abundant throughout the early pancreatic development. As the pancreas grows continuously between 6 and 11 WD, a marked increase in leukocyte density is noticed after 9 WD.
Of note, among the 29, only two pancreases, aged 6 and 9 WD, contained very few (fewer than 10) CD3-, CD20- or CD79-expressing cells (data not shown), suggesting that T and B cells may be absent from the human pancreas at the earliest developmental stages. At 6–8 WD, the leukocytic lineage, CD45RA-positive cells were located mainly in the mesenchyme surrounding the peripheral ducts (Fig. 1A, E and F). At 8.5–12 WD, these cells were more numerous in the surrounding mesenchyme and were found also in the loose connective tissue between the central epithelial ducts and the endocrine tissue (Fig. 1B and H).
At 18 WD, at birth and in adulthood, immune cells were found in the connective tissue throughout the pancreas (data not shown). At these later stages, the surrounding mesenchyme was a thin layer of connective tissue, and lymphoid follicular structure was visible within the pancreas (data not shown).
To determine whether these cells are extravascular, probably resident, or intravascular circulating leukocytes, we determined their location with regard to the blood vessels.
Blood vessels were present in 6-WD pancreases, and the leukocytic lineage cells were mainly extravascular
We first looked for vascular structures on hematoxylin-eosin-stained sections (Fig. 1C). From 6 WD, vessels were seen in the surrounding mesenchyme and the connective tissue of all the pancreases studied. They produced a positive stain for the endothelial cell marker CD34 (Fig. 1D), and some of them contained circulating cells (Fig. 1C).
Double immunostaining with antibody to CD34 and with immune antigen markers showed that most of the cells expressing immune antigens (CD45RA, HLADR and CD68) were outside the CD34-positive vessels (Fig. 1D). This suggested that these cells are not intravascular circulating leukocytes and that they may be resident cells.
The lack of CD34-positive hematopoietic progenitor cells suggested that the pancreatic mesenchyme was not an extralymphoid hematopoietic tissue in the early stages of development. Moreover, the mesenchyme surrounding the pancreas was not related to the retro-aorta-gonad-metanephros zone, which is rich in CD34-positive cells, previously reported to be a source of hematopoietic stem cells before 6 WD in man (Solvason & Kearney 1992, Nunez et al. 1996, Tavian et al. 1996, 1999, Marshall et al. 1999).
Coexpression of leukocyte markers in the developing human pancreas
The numerous CD45RA-positive cells found in all the developing pancreases were round when located in the peripheral mesenchyme but exhibited a dendritic cell-like morphology with cytoplasmic protrusions close to the ducts (Fig. 1B, insert ♦ and insert *, and G). Almost all the CD45RA-positive cells coexpressed HLADR and CD68 antigens (data not shown).
This indicates that the leukocyte cell population expressing CD45RA and HLADR could be antigen-presenting cells, which include dendritic cells, macrophages and B cells. Very few cells (n=3) expressing CD20 and CD79 antigens were detected in two early human pancreases (data not shown), therefore excluding B cells (which express CD20) from the leukocyte cell population found in the early developing pancreas. All the HLADR-positive cells expressed CD45RA (Fig. 1E–G). Double staining with CD45RA (or HLADR) and insulin showed no insulin-positive cell expressing CD45RA or HLADR at all the studied time points (Fig. 1H and data not shown), suggesting that at 6–12 WD human endocrine pancreatic cells may not constitutively express HLADR antigens.
This study provides the first evidence that the early human pancreas contains numerous CD45RA-positive leukocytes of various lineages. These leukocytes had a dendritic cell-like morphology when they were close to the epithelial duct cells, suggesting that they could be dendritic cells in a different maturation and/or functional status, depending on their location. Most of them coexpressed HLADR. Their predominant antigen-expression pattern was that of monocyte/macrophage lineage-derived cells.
We then sought to determine whether these leukocytes in the early developing human pancreas were dendritic cells, macrophages, or both. Most of the cells expressed HLADR and CD68, suggesting that they were macrophages (Tosi et al. 2000, Marinova et al. 2001, Odeberg & Soderberg-Naucler 2001). However, as the CD68 antigen has also been found on immature dendritic cells in vitro and ex vivo, we investigated the presence of various dendritic cell populations (Sapp et al. 1999). Almost all the CD45RA-positive cells were CD1a-negative, ruling out the presence of epithelial Langerhans cells in the developing pancreas (data not shown). Fewer than 20 cells were CD11c-positive cells. These results suggest that some of the CD45RA-positive leukocytes may have been of the interstitial dendritic cell type. For better characterization of these cells, we tested the expression of dendritic differentiation markers in pancreases at 9, 10 and 11 WD (data not shown). We found a few (fewer than 20 cells) CD80-positive cells (dendritic cells with the capacity to process the antigens but not to present them efficiently) and a few CD83-positive cells (mature dendritic cells capable of stimulating lymphocytes in vitro) (Steinman 1991, Lindhout et al. 1998).
These experiments suggested that macrophages contributed most of the CD45RA-positive, leukocytic lineage cells detected in the early pancreatic human mesenchyme.
We were interested to compare these features with those of other developing tissues and organs. We looked for the presence of CD45RA- and HLADR-expressing cells in a transitory hematopoietic organ (the liver), endoderm-derived organs (the thyroid, lung, pancreas and intestine), branched organs (the kidney, lung and pancreas) and other organs (the adrenal glands and heart) of five embryos and fetuses (at 6, 8, 8.5, 9 and 11 WD). CD45RA-expressing cells were found in 6–11 WD liver, thyroid, kidney, heart and adrenal glands (data not shown). The pancreas and the intestinal mesenchyme contained numerous CD45RA- and HLADR-positive cells (Fig. 1E–G). As shown in Fig. 1E and F, few CD45RA- (and HLADR-) positive cells were seen in the duodenum.
As one of the known constitutive functions of macrophages is clearing apoptotic cells, we examined the degree and location of apoptosis in the human developing pancreas.
Apoptosis analysis
Apoptotic cells were mainly detected in the surrounding mesenchyme, except for two found in epithelial ducts. Their density per mm2 of pancreatic surface was 0.9 at 7–7.5 WD, 0 ± 0 at 8–8.5 WD (n=2) and 1.16 ± 0.9 at 9–11WD (n=5). The coefficient of correlation between the leukocyte density data and the apoptosis data was 0.6, suggesting that the presence of pancreatic macrophages could serve another physiologic function.
In order to understand the mechanisms underlying the presence of extravascular leukocytes in the developing pancreas, we investigated the expression of chemokines and cytokines otherwise known to act as signals that respectively promote leukocyte migration within the organ and/or promote both macrophage migration and survival.
Analysis of the expression of chemokines and their receptors within the developing human pancreas
Detection of chemokines by both RT–PCR and ISH
Because most of the monocyte-macrophage lineage cells were detected outside the blood vessels, we looked for pancreatic expression of chemokines involved in tissue leukocyte recruitment. We were studying a developing organ and, consequently, elected to focus on chemokines constitutively expressed by organs and involved in the recruitment of lymphocytes and dendritic cells (Dieu et al. 1998, Dieu-Nosjean et al. 1999, Coulomb-L’Herminé et al. 1999, Sozzani et al. 2000), namely, SDF1, 6 Ckine and MIP3-beta, using RT–PCR to detect the corresponding mRNAs. We found that SDF1 mRNA was expressed from 6 WD to adulthood. Expression of MIP3-beta and 6 Ckine mRNA was also found at 6–10 WD (Fig. 2A and Table 4). As a control, we looked for MIP 3-alpha mRNA expression. Neither this proinflammatory chemokine (Kleef et al. 1999) nor its receptor CCR6 were detected at the stages studied (Table 4).
The expression of SDF1 and 6 Ckine was confirmed by immunohistochemistry studies. As early as 6 WD, some epithelial ducts expressed the SDF1 protein. At 9 WD, epithelial SDF1 staining was detected in all the ducts (Fig. 2B). At 18 WD and later, a few duct cells expressed SDF1. At 6–18 WD, SDF1 expression was found only in the epithelial cells (Fig. 2B and C). Expression of 6 CKine protein was evidenced in a few duct cells at 6WD, in most ducts after 8 WD, and in a few endothelial cells at 9 and 11 WD (Fig. 2D and insert).
This evidence that the early human epithelium expresses chemokines involved in the recruitment of leukocytic lineage cells prompted us to look for expression of the corresponding receptors.
Detection of chemokine receptors by ISH
SDF1, 6 Ckine and MIP3-beta are constitutively expressed chemokines that act by binding to their receptors, namely, CXCR4 for SDF1 and CCR7 for 6 Ckine and MIP3-beta. CXCR4 and CCR7 mRNA expression was detected by RT–PCR in the human pancreases (Fig. 2A and Table 4). ISH studies to localize the expression of these receptors within the various pancreatic cell types visualized CXCR4 and CCR7 mRNA in the human pancreas at 7.5 and 8.5 WD. The staining was found not only in numerous cells of the surrounding mesenchyme, but also in the endothelial and epithelial duct cells (Fig. 2E and F). Some of the cells expressing CXCR4 and CCR7 mRNA also expressed the CD45RA protein (Fig. 2H and I).
Analysis of the expression of CSF1 and GMCSF, the major growth factors involved in macrophage differentiation, and their receptors (CSF1R, GMCSFR) within the developing human pancreas
As macrophages were present in the human developing pancreas, we looked for the expression of CSF1, which is the major growth factor involved in resident macrophage proliferation, differentiation and survival in nonlymphoid tissues (Witmer-Pack et al. 1993, Pollard & Stanley 1996, Wiktor-Jedrzejczak & Gordon 1996). We found CSF1 transcripts in the liver (our control organ for macrophage markers) (Fig. 3A and B, and Table 4) and in epithelial (part of ducts and aggregated cells) and endothelial cells of human pancreas at 7, 8.5 and 9.5 WD (Fig. 3D and insert, and Table 4). We then studied the CSF1 receptor, which has been shown to be a macrophage marker (Guilbert & Stanley 1980, Sherr et al. 1985, Sasmono et al. 2003). CSF1R transcripts were present in the liver (Fig. 3A and C) and in pancreatic cells of the connective tissue (Fig. 3A and E, and Table 4), a finding that characterizes them as macrophages (Guilbert & Stanley 1980, Sherr et al. 1985, Sasmono et al. 2003). GM–CSF and receptor expression was not found by RT–PCR or by ISH (Table 4).
Discussion
This is the first study documenting the presence of mesoderm-derived CD45RA leukocyte subsets in the early developing human pancreas. Most of these cells were components of the surrounding mesenchyme. Their location outside the developed vessels suggested that they were not circulating cells and may have migrated into the pancreatic extravascular space.
Most of the CD45RA-positive leukocytes were macrophages (expressing CD68 antigen), and dendritic cells positive for CD83 (a molecule known to be expressed late in dendritic cell differentiation). Passenger leukocytes expressing CD45RA have been found in the pancreas at 13–20 weeks of gestation (MacKenzie et al. 1999). HLA-positive macrophages and dendritic cells were detected in postnatal human pancreases and islets (Lu et al. 1996 and data not shown). Macrophages express the HLADR molecule (Tosi et al. 2000, Marinova et al. 2001) otherwise known to display antigenic peptides (Wright-Browne et al. 1997, Lindhout et al. 1998, Starzl & Zinkernagel 1998). Dendritic cells were frequent in adult organs but rare in organs from donors under 7 years of age (Lu et al. 1996). The presence of macrophages and dendritic cells during the early development in the absence of T or B cells militates against antigen selection from endocrine lineages. Indeed, recent data suggest that islet-induced macrophage recruitment could delay streptozotocin-induced diabetes in mice (Scott et al. 2001). The differences seen in pancreatic HLADR expression during the organ’s development in children and in adults may influence the immunogenicity and/or survival of pancreatic and islet grafts in an age-dependent manner (Lu et al. 1996 and present study).
Pancreatic development results from epithelial/mesenchyme interactions. The importance of the mesenchyme in organ morphogenesis has been well documented (Cardoso 2001). Growth factors secreted by the mesenchyme have been shown to be involved in pancreatic development, but their cellular source has not been identified (Miralles et al. 1998, Miralles et al. 1999, Scharfmann 2000). Macrophages are involved in numerous nonimmune functions during organ development, tissue remodeling and homeostasis (Horton 1996). These multiple functions are mediated by macrophage synthesis and secretion of a broad repertoire of cytokines, growth factors and extracellular matrix remodeling molecules, and by their ability to recognize and phagocytize apoptotic cells. Among these growth factors, FGF2 (Rappolee & Werb 1992), transforming growth factor-β, insulin-like growth factor-I and activinβA have been shown to play a role in pancreatic development (Le Bras et al. 1998, Miralles et al. 1999, Scharfmann 2000, Dichmann et al. 2003). The mesenchymal resident macrophages we found at 6–12 WD and/or their secreted factors may be involved in early pancreatic development. At this stage, endocrine tissue differentiates and massively expands, and the epithelial ducts containing endocrine and exocrine precursors develop (Fukayama et al. 1986, Polak et al. 2000). Moreover, we previously showed that the proliferation rate of peripheral epithelial ducts in contact with the surrounding mesenchyme was two- to threefold that of the central epithelial ducts where endocrine differentiation occurs (Polak et al. 2000). Taken together, these results suggest that the peripheral, leukocyte-rich mesenchyme may regulate positively the epithelial duct cell proliferation and/or repress endocrine differentiation, as shown in mice in vitro (Scharfmann 2000). However, the central loose mesenchyme containing dendritic leukocytes may exert opposite effects. Both effects should involve a different mesenchyme cell component and thus differentially regulated secreted growth factors. The investigation of the spatial and temporal heterogeneity of the cellular composition and the developmental features we observed during the early human pancreas development should considerably improve our knowledge of the mechanisms involved in the epithelium/mesenchyme interactions.
The role of macrophages in pancreatic development/remodeling has been further substantiated in our work in macrophage-deficient mice (Fadok et al. 2001).
A reasonable hypothesis to explain the presence of macrophages relies on their known constitutive function of apoptotic cell clearance (Banaei-Bouchareb et al. 2004), as recently described in mice for the thymus (Jacobson et al. 1997, Scott et al. 2001) and shown during mammary gland development (Gouon-Evans et al. 2000, 2002). The fact that apoptosis was rarely detected in the early human pancreas and was not particularly correlated with macrophage density argues that macrophages could play a broader role during pancreatic development.
Macrophage differentiation or survival depends on local macrophage-specific growth factor secretion. We found CSF1 transcripts expressed by the human fetal endothelial and epithelial duct cells. Furthermore, CSF1 receptor transcripts were expressed by cells of the connective tissue, probably macrophages (Guilbert & Stanley 1980, Sherr et al. 1985, Sasmono et al. 2003). GM–CSF and GM–CSF receptor transcripts were not detected in all the studied pancreases. Thus, as in mice, resident macrophages in the developing human pancreas should depend chiefly on locally secreted CSF1 for their differentiation (Witmer-Pack et al. 1993, Pollard & Stanley 1996, Wiktor-Jedrzejczak & Gordon 1996).
We have shown that the early human pancreas expresses constitutive chemokines involved in recruitment and migration of macrophages and dendritic cells (Luster 1998). There is evidence that the chemokine SDF1 is produced within the bone marrow and mediates chemokinesis and chemotaxis of macrophages, B cells, T cells and CD34+ cells that express the CXCR4 receptor (Moore et al. 2001). In our study, the epithelial duct cells expressed SDF1. Our finding that these cells also expressed the SDF1 receptor CXCR4 strongly suggests a role for SDF1 in the local recruitment of CD45RA-positive cells. The expression of the chemokines (SDF1 and 6 CKine) by the pancreatic epithelium, and their receptors (CXCR4 and CCR7) by CD45RA-positive cells and pancreatic epithelial cells also indicates that the contribution of these molecules to organ development may extend beyond their well-documented role in recruitment of leukocyte-derived cells (Dieu et al. 1998, Luster 1998, Tachibana et al. 1998, Zou et al. 1998, Coulomb-L’Herminé et al. 1999, Dieu-Nosjean et al. 1999, Sozzani et al. 2000). Resident leukocytes are recruited within tissues by local expression of chemoattractants. A previous study found few B cells in the connective tissue of the fetal lung, contrasting with marked epithelial SDF1 expression, which is involved in the recruitment and differentiation of these cells (Coulomb-L’Herminé et al. 1999). Thus, in addition to recruiting hematopoietic cells, chemokines should be studied within the context of morphogenesis (Tachibana et al. 1998, Zou et al. 1998). Indeed, a previous study reveals a pathway that links reverse signaling to neural guidance during brain development, involving Ephrin-B, SDF1 and CXCR4 (Lu et al. 2001). Moreover, the migration of the pancreatic epithelial duct, CXCR4-positive cells in response to SDF1 has been very recently documented in mice (Kayali et al. 2003).
The absence of expression of the proinflammatory induced chemokine (MIP3 alpha) and its receptor (CCR6) by the developing human pancreas argues that developmental processes should occur through constitutively expressed molecules rather than inducible ones.
However, the scarcity of human material and the fact that experiments involving disturbance of the human embryonic pancreas development are extremely difficult to conduct in vivo and even ex vivo are major obstacles to further work aimed at elucidating the role played by chemokines and their receptors in pancreatic development in man. Consequently, we are now turning to animal models that are genetically engineered to lack various chemokines and their receptors.
Tissue collection and preparation
Immunohistochemistry and ISH | mRNA extraction | ||||
---|---|---|---|---|---|
Paraffin-embedded pancreas | Frozen pancreas | ||||
Age (WD) | Number | Age (WD) | Number | Age (WD) | Number |
32–35 days | 1 | 7 days | 2 | 6–6.5 days | 5 |
6–6.5 | 5 | 8.5 | 2 | 7–7.5 | 7 |
7–7.5 | 3 | 9 | 1 | 8–8.5 | 9 |
8–8.5 | 9 | Total | 5 | 9–9.5 | 5 |
9–9.5 | 5 | 10–10.5 | 2 | ||
10–10.5 | 4 | 11 | 1 | ||
11 | 2 | 14 | 1 | ||
12 | 1 | 17 | 1 | ||
Total | 30 | 19 | 1 | ||
28 | 1 | ||||
33 | 1 | ||||
Total | 34 |
Antibodies and control tissues used in immunohistochemical studies
Species | Supplier | Dilution | Control tissue | Secondary antibody | Dilution | Supplier | |
---|---|---|---|---|---|---|---|
*Works on frozen tissues. †Generous gift from F. Arenzana-Seisdedos, Institut Pasteur, France. | |||||||
Dako: Dako-Cytomation, Glostrup, Denmark; Immunotech: Immunotech, Fullerton, CA, USA; Biogenex: Biogenex, San Raman, CA, USA. | |||||||
Primary antibodies | |||||||
CD45 RA (leukocyte common antigen) | Mouse anti-human | Dako | 1/100 | Lymph node | Multispecies Biotinylated (KIT) | 1/40 | Biogenex |
CD34 | Mouse anti-human | Immunotech | PUR | Vascular tumor | Goat anti-mouse FITC | 1/200 | Immunotech |
CD3 | Rabbit anti-human | Dako | 1/100 | Lymph node, spleen | Multispecies Biotinylated (KIT) | 1/40 | Biogenex |
CD1a | Mouse anti-human | Immunotech | PUR | Skin | Multispecies Biotinylated (KIT) | 1/20 | Biogenex |
CD20 | Mouse anti-human | Dako | 1/25 | Lymph node | Multispecies Biotinylated (KIT) | 1/40 | Biogenex |
CD68 | Mouse anti-human | Dako | 1/50 | Lymph node | Multispecies Biotinylated (KIT) | 1/40 | Biogenex |
HLADR | Mouse anti-human | Dako | 1/50 | Lymph node | Multispecies Biotinylated (KIT) | 1/40 | Biogenex |
CD80* | Mouse anti-human | Immunotech | 1/50 | Tonsil | Goat anti-mouse FITC | 1/100 | Immunotech |
CD83* | Mouse anti-human | Immunotech | 1/50 | Tonsil | Goat anti-mouse FITC | 1/100 | Immunotech |
CD11c* | Mouse anti-human | Dako | 1/25 | Tonsil | Goat anti-mouse FITC | 1/100 | Immunotech |
SDF1† | Mouse anti-human | 1.5/100 | Salivary gland | Multispecies Biotinylated (KIT) | 1/40 | Biogenex | |
6CKine | Goat anti-human | R&D Systems | 12 μg/ml (1/10) | Tonsil | Donkey anti-goat Texas Red | 1/100 | Jackson Immuno Research Lab |
Insulin | Mouse anti-human | Sigma | 1/1000 | Pancreas | Multispecies Biotinylated (KIT) | 1/40 | Biogenex |
Primers and control tissues used in RT–PCR and ISH studies
Accession no. | Primer sequence | Product size (bp) | Control tissue | |
---|---|---|---|---|
CCL: CC chemokine ligand, CXCL: CXC chemokine ligand; L: left; R: right. | ||||
Name | ||||
L cyclophilin Multispecies | GGTCAACCCCACCGTGTTCT | |||
R cyclophilin Multispecies | Y00052 | TGCCATCCAGCCACTCAGTCT | 352 | |
L CSF1 | AGAAGACAGACCATCCATCTG | Fetal liver | ||
R CSF1 | M37435 | TCAGTCAAAGGAACGGAGTTA | 282 | |
L GMCSF | AGCCCCAGCACACAGCCCTGG | Plasmide | ||
R GMCSF | M 11734 | CTCCTGGACTGGCTCCCAGCAGTCAAAGGG | 365 | |
L CSF1R | TTCAAATGACTCCTTCTCTGAGC | |||
R CSF1R | X03663 | CGTGTAGACACAGTCAAAGATGC | 306 | Foetal liver |
L GMCSFR | AGAGAAATCGGATCTGCGAA | Liver | ||
R GMCSFR | X54934 | CCACATGGGTTCCTGAGTCT | 466 | |
L SDF1 (CXCL12) | CGTCAGCCTGAGCTACAGATGC | |||
R SDF1 | L36034, 36033 | TTCTCCAGGTACTCCTGAATCCAC | 188 | Liver |
L MIP3 beta (ELC, CCL19) | GTGCTGCTTCACCTACACTACCT | |||
R MIP3 beta | Z49270 | CTTTATGTCTCTGAGCTGTGCCT | 236 | Lung |
L 6CKINE (SLC, CCL21) | CTCAAGTACAGCCAAAGGAAGAT | |||
R 6CKINE | U88320 | CCTTTCCTTTCTTGCCAGTCT | 258 | Lung |
L MIP3 alpha (LARC, CCL20) | ATGTGCTGTACCAAGAGTTTGCT | |||
R MIP3 alpha | D86955 | GGAGACGCACAATATATTTCACC | 267 | Lung |
L CCR6 | TCTCCAGCTCAACTTTTGTCTTC | |||
R CCR6 | NM 004367 | CAAGCACCACAGCTATGATTACA | 254 | Lung |
L CCR7 | CTCCTTGTCATTTTCCAGGTATG | |||
R CCR7 | L31581 | AGGTAGGTATCGGTCATGGTCTT | 250 | Human-derived dendritic cells |
L CXCR4 | CTCCTGCTGACTATTCCCGACTT | |||
R CXCR4 | M99293 | AAGAAAGCCAGGATGAGGATGAC | 258 | Blood T cells |
ISH | ||||
L CCR7 | CCATGAGCTTCTGTTACCTTGTC | |||
R CCR7 | L31581 | GTCTCCCCACTATCTCTGGTCTT | 882 | Human-derived dendritic cells |
Macrophage growth factors, chemokines and their receptor mRNA expression in the developing human pancreas
Age (WD) | ||||||
---|---|---|---|---|---|---|
6 | 7 | 8–8.5 | 9–9.5 | 10–10.5 | 11 | |
ND: not done. | ||||||
Growth factor | ||||||
CSF1 | ND | +(n=5) | +(n=5) | +(n=2) | ND | ND |
GM–CSF | ND | +(n=5) | +(n=5) | +(n=2) | ND | ND |
Receptor | ||||||
CSF1R | ND | +(n=5) | +(n=5) | +(n=2) | ND | ND |
GMCSFR | ND | +(n=5) | +(n=5) | +(n=2) | ND | ND |
Chemokines | ||||||
SDF1(CXCL12) | +(n=5) | +(n=1) | +(n=3) | +(n=4) | +(n=1) | +(n=1) |
MIP3beta (CCL19) | +(n=3) | ND | +(n=2) | +(n=3) | +(n=1) | ND |
6CKine (CCL21) | +(n=2) | ND | +(n=2) | +(n=2) | +(n=1) | ND |
MIP3 alpha (CCL20) | −(n=2) | |||||
Receptors | ||||||
CXCR4 | +(n=2) | ND | +(n=2) | +(n=4) | +(n=1) | +(n=1) |
CCR7 | +(n=2) | ND | +(n=1) | +(n=2) | +(n=1) | ND |
CCR6 | ND | ND | ND | −(n=2) | ND | ND |
This work was supported in part by two nonprofit organizations, Aide aux Jeunes Diabétiques (AJD) and Fondation pour la Recherche Médicale to L B and M P. We thank Geneviè ve Milon (Institut Pasteur) for fruitful discussion and advice throughout this study; Marie Claire Castellotti and Aurore Carré for technical assistance with the ISH experiments; F Arenzana-Seisdedos (Institut Pasteur) for donating the anti-human SDF1 antibody and the human CXCR4 cDNA cloned in pCDNAIII plasmid; and M C Dieu-Nosjean for donating the human CD34-derived dendritic cells expressing CCR7. The authors declare that there is no conflict of interest that would prejudice the impartiality of this scientific work.
References
Banaei-Bouchareb L, Gouon-Evans V, Samara-Boustani D, Castellotti MC, Czernichow P, Pollard JW & Polak M 2004 Insulin cell mass is altered in Csf1op/Csf1op macrophage-deficient mice. Journal of Leukocyte Biology 762 359–367.
Beattie GM, Otonkoski T, Lopez AD & Hayek A 1997 Functional beta-cell mass after transplantation of human foetal pancreatic cells. Differentiation or proliferation? Diabetes 46 244–248.
Cardoso WV 2001 Molecular regulation of lung development. Annual Review of Physiology 63 471–494.
Castaing M, Peault B, Basmaciogullari A, Casal I, Czernichow P & Scharfmann R 2001 Blood glucose normalization upon transplantation of human embryonic pancreas into beta-cell-deficient SCID mice. Diabetologia 4411 2066–2076.
Coulomb-L’Herminé A, Amara A, Schiff C, Durand-Gasselin I, Foussat A, Delaunay T, Chaouat G, Capron F, Ledee N, Galanaud P, et al.1999 Stromal cell-derived factor 1 SDF-1 and antenatal human B cell lymphopoiesis expression of SDF-1 by mesothelial cells and biliary ductal plate epithelial cells. PNAS 96 8585–8590.
Dichmann DS, Miller CP, Jensen J, Scott Heller R & Serup P 2003 Expression and misexpression of members of the FGF and TGFbeta families of growth factors in the developing mouse pancreas. Developmental Dynamics 226 663–674.
Dieu MC, Vandervliet B, Vicari A, Bridon JM, Oldham E, Ait-Yahia S, Briere F, Zlotnik A, Lebecque S & Caux C 1998 Selective recruitment of immature and mature dendritic cells by distinct chemokines expressed in different anatomic sites. Journal of Experimental Medicine 188 373–386.
Dieu-Nosjean M, Vicari A, Lebecque S & Caux C 1999 Regulation of dendritic cell trafficking, a process that involves the participation of selective chemokines. Journal of Leukocyte Biology 66 252–262.
Fadok VA, Bratton DL & Henson PM 2001 Phagocyte receptors for apoptotic cells: recognition, uptake, and consequences. Journal of Clinical Investigation 108 957–962.
Fukayama M, Ogawa M, Hayashi Y & Koike M 1986 Development of human pancreas. Immunohistochemical study of foetal pancreatic secretory proteins. Differentiation 31 127–133.
Ghosh AK, Gatter KC & Mason DY 1984 Immunocytochemical characterization of monoclonal anti-HLA-DR antibodies. In Disease Markers, pp 223–233. Ed CM Steel. Chichester: Wiley.
Gouon-Evans VE, Rothenberg M & W Pollard J 2000 Postnatal mammary gland development requires macrophages and eosinophils. Development 127 2269–2282.
Gouon-Evans V, Lin EY & Pollard JW 2002 Requirement of macrophages and eosinophils and their cytokines/chemokines for mammary gland development. Breast Cancer Research 4 155–164.
Guilbert LJ & Stanley ER 1980 Specific interaction of murine colony-stimulating factor with mononuclear phagocytic cells. Journal of Cell Biology 85 153–159.
Horton MA 1996 Macrophages and related cells. In Blood Cell Biochemistry, 5th edn, vol 5. Ed MA Horton. New York: Plenum Press.
Jacobson MD, Weil M & Raff MC 1997 Programmed cell death in animal development. Cell 88 347–354.
Kayali AG, Van Gunst K, Campbell IL, Stotland A, Kritzik M, Liu G, Flodstrom-Tullberg M, Zhang YQ & Sarvetnick N 2003 The stromal cell-derived factor-1/CXCR4 ligand–receptor axis is critical for progenitor survival and migration in the pancreas. Journal of Cell Biology 163 859–869.
Kleef J, Kusama T, Rossi LD, Ishiwata T, Maruyama H, Freiss H, Buchler MW, Zlotnik A & Korc M 1999 Detection and localization of MIP-3α/LARC/Exodus, a macrophage proinflammatory chemokine, and its CCR6 receptor in human pancreatic cancer. International Journal of Cancer 81 650–657.
Krakowski M, Abdelmalik R, Mocnik L, Krahl T & Sarvetnick N 2002 Granulocyte macrophage-colony stimulating factor (GM-CSF) recruits immune cells to the pancreas and delays STZ-induced diabetes. Journal of Pathology 196 103–112.
Lafferty KJ & Hao L 1993 Fetal pancreas transplantation for treatment of IDDM patients. Diabetes Care 16 383–386.
Le Bras S, Miralles F, Basmaciogullary A, Czernichow P & Scharfmann R 1998 Fibroblast growth factor 2 promotes pancreas epithelial cell proliferation via functional fibroblast growth factor receptors during embryonic life. Diabetes 47 1236–1242.
Lindhout E, Figdor GC & Adema GJ 1998 Dendritic cells: migratory cells that are attractive. Cell Adhesion and Communication 6 117–123.
Lu W, Pipeleers DG, Kloppel G & Bouwens L 1996 Comparative immunocytochemical study of MHC class II expression in human donor pancreas and isolated islets. Virchows Archiv 429 205–211.
Lu Q, Sun EE, Klein RS & Flanagan JG 2001 Ephrin-B reverse signalling is mediated by a novel PDZ-RGS protein and selectively inhibits G protein-coupled chemoattraction. Cell 105 69–79.
Luster AD 1998 Chemokines – chemotactic cytokines that mediate inflammation. New England Journal of Medicine 338 436–445.
MacKenzie DA, Sollinger HW & Hullett DA 1999 Analysis of passenger cell composition of human foetal pancreas: implications for transplantation. Transplantation Proceedings 31 651.
Marinova T, Altankova I, Dimitrova D & Pomakov Y 2001 Presence of HLA-DR immunopositive cells in human foetal thymus. Archives of Physiology and Biochemistry 109 74–79.
Marshall CJ, Moore RL, Thorogood P, Brickell PM, Kinnon C & Thrasher AJ 1999 Detailed characterization of the human aorta-gonad-mesonephros region reveals morphological polarity resembling a hematopoietic stromal layer. Developmental Dynamics 215 139–147.
Miralles F, Czernichow P & Scharfmann R 1998 Follistatin regulates the relative proportions of endocrine versus exocrine tissue during pancreatic development. Development 125 1017–1024.
Miralles F, Czernichow P, Ozaki K, Itoh N & Scharfmann R 1999 Signaling through fibroblast growth factor receptor 2b plays a key role in the development of the exocrine pancreas. PNAS 96 6267–6272.
Moore KL, Persaud TVN & Shiota K 1995 In Color Atlas of Clinical Embryology, pp 153–163. Philadelphia, PA, USA: WB Saunders.
Moore MA, Hattori K, Heissig B, Shieh JH, Dias S, Crystal RG & Rafii S 2001 Mobilization of endothelial and hematopoietic stem and progenitor cells by adenovector-mediated elevation of serum levels of SDF-1, VEGF, and angiopoietin-1. Annals of the New York Academy of Sciences 938 36–45.
Nunez C, Nishimoto N, Gartland GL, Billips LG, Burrows PD, Kubagawa H & Cooper MD 1996 B cells are generated throughout life in humans. Journal of Immunology 156 866–872.
Odeberg J & Soderberg-Naucler C 2001 Reduced expression of HLA class II molecules and interleukin-10- and transforming growth factor beta1-independent suppression of T-cell proliferation in human cytomegalovirus-infected macrophage cultures. Journal of Virology 75 5174–5181.
O’Rahilly R & Müller F 1987 In Developmental Stages in Human Embryos. Carnegie Institution of Washington, Publication no. 637. Meriden, CT, USA: Meriden-Stinehour Press.
Pablos JL, Amara A, Bouloc A, Santiago B, Caruz A, Galindo M, Delaunay T, Virelizier JL & Arenzana-Seisdedos F 1999 Stromal-cell derived factor is expressed by dendritic cells and endothelium in human skin. American Journal of Pathology 155 1577–1586.
Polak M, Bouchareb-Banaei L, Scharfmann R & Czernichow P 2000 Early pattern of differentiation in the human pancreas. Diabetes 49 225–232.
Pollard JW & Stanley ER 1996 Pleiotropic roles for CSF-1 in development defined by the mouse mutation, osteopetrotic. Advances in Developmental Biochemistry 4 153–193.
Rappolee DA & Werb Z 1992 Macrophage-derived growth factors. Current Topics in Microbiology and Immunology 181 87–140.
Sapp M, Engel Mayer J, Larsson M, Granelli-Piperno A, Steinman R & Bhardwaj N 1999 Dendritic cells generated from blood monocytes of HIV-1 patients are not infected and act as competent antigen presenting cells eliciting potent T-cell responses. Immunology Letters 66 121–128.
Sasmono RT, Oceandy D, Pollard JW, Tong W, Pavli P, Wainwright BJ, Ostrowski MC, Himes SR & Hume DA 2003 A macrophage colony-stimulating factor receptor-green fluorescent protein transgene is expressed throughout the mononuclear phagocyte system of the mouse. Blood 101 1155–1163.
Scharfmann R 2000 Control of early development of the pancreas in rodents and humans: implications of signals from the mesenchyme. Diabetologia 43 1083–1092.
Scott RS, McMahon EJ, Pop SM, Carrichio R, Reap EA, Cohen PJ, Earp HS & Matsushima GK 2001 Phagocytosis and clearance of apoptotic cells is mediated by MER. Nature 411 207–211.
Sherr CJ, Rettenmier CW, Sacca R, Roussel MF, Look AT & Stanley ER 1985 The c-fms proto-oncogene product is related to the receptor for the mononuclear phagocyte growth factor, CSF-1. Cell 41 665–676.
Slack JMW 1995 Developmental biology of the pancreas. Development 121 1569–1580.
Solvason N & Kearney JF 1992 The human foetal omentum A site of B cell generation. Journal of Experimental Medicine 175 397–404.
Sozzani S, Allavena P, Vecchi A & Mantovani A 2000 Chemokines and dendritic cell traffic. Journal of Clinical Immunology 20 151–160.
Starzl TE & Zinkernagel RM 1998 Antigen localization and migration in immunity and tolerance. New England Journal of Medicine 339 1905–1913.
Steinman R 1991 The dendritic cell system and its role in immunogenicity. Annual Review of Immunology 9 271–296.
Tachibana K, Hirota S, Lizasa H, Yoshida H, Kawabata K, Kataoka Y, Kitamura Y, Matsushima K, Yoshida N, Nishikawa S-I et al.1998 The chemokine receptor CXCR4 is essential for vascularization of the gastrointestinal tract. Nature 393 591–594.
Tavian M, Coulombel L, Luton D, Clemente HS, Dieterlen-Lievre F & Peault B 1996 Aorta-associated CD34+ hematopoietic cells in the early human embryo. Blood 87 67–72.
Tavian M, Hallais MF & Peault B 1999 Emergence of intraembryonic hematopoietic precursors in the pre-liver human embryo. Development 126 793–803.
Tosi G, De Lerma Barbaro A, D’Agostino A, Valle MT, Megiovanni AM, Manca F, Caputo A, Barbanti-Brodano G & Accolla RS 2000 HIV-1 Tat mutants in the cysteine-rich region downregulate HLA class II expression in T lymphocytic and macrophage cell lines. European Journal of Immunology 30 19–28.
Usadel KH, Schwedes U, Bastert G, Steinau U, Klempa I, Fassbinder W & Schöffling K 1980 Transplantation of human foetal pancreas. Experience in thymus-aplastic mice and rats and in a diabetic patient. Diabetes 29 74–79.
Wiktor-Jedrzejczak W & Gordon S 1996 Cytokine regulation of the macrophage M phi system studied using the colony-stimulating factor-1-deficient op/op mouse. Physiological Reviews 76 927–947.
Witmer-Pack MD, Hughes DA, Schuler G, Lawson L, McWilliam A, Inaba K, Steinman RM & Gordon S 1993 Identification of macrophages and dendritic cells in the osteopetrotic op/op mouse. Journal of Cell Science 104 1021–1029.
Wright-Browne V, McClain KL, Talpaz M, Ordonez N & Estrov Z 1997 Physiology and pathophysiology of dendritic cells. Human Pathology 28 563–579.
Zou Y, Kottmann AH, Kuroda M, Taniuchi I & Littman DR 1998 Function of the chemokine receptor CXCR4 in haematopoiesis and in cerebellar development. Nature 393 595–599.