The renal cortical interstitium: morphological and functional aspects
The renal interstitial compartment, situated between basement membranes of epithelia and vessels, contains two contiguous cellular networks. One network is formed by interstitial fibroblasts, the second one by dendritic cells. Both are in intimate contact with each other. Fibroblasts are interconnected by junctions and connected to basement membranes of vessels and tubules by focal adhesions. Fibroblasts constitute the “skeleton” of the kidney. In the renal cortex, fibroblasts produce erythropoietin and are distinguished from other interstitial cells by their prominent F-actin cytoskeleton, abundance of rough endoplasmic reticulum, and by ecto-5′-nucleotidase expression in their plasma membrane. The resident dendritic cells belong to the mononuclear phagocyte system and fulfil a sentinel function. They are characterized by their expression of MHC class II and CD11c. The central situation of fibroblasts suggests that signals from tubules, vessels, and inflammatory cells converge in fibroblasts and elicit an integrated response. Following tubular damage and inflammatory signals fibroblasts proliferate, change to the myofibroblast phenotype and increase their collagen production, potentially resulting in renal fibrosis. The acquisition of a profibrotic phenotype by fibroblasts in renal diseases is generally considered a main causal event in the progression of chronic renal failure. However, it might also be seen as a repair process.
KeywordsMyofiboblasts Dendritic cells Unilateral ureter obstruction Thiazide diuretics Morphology
The fibroblasts in the interstitium provide the “skeleton” of the tissue and maintain the three-dimensional architecture of the tissue. The interstitium is necessarily involved in all intrarenal exchange processes since the reabsorption and secretion of fluid and solutes implicates a transit across the interstitial compartment. Cortical fibroblasts play a role in the production of regulatory substances, such as extracellular adenosine (Le Hir and Kaissling 1993) and erythropoietin (Bachmann et al. 1993; Maxwell et al. 1993). In the healthy organ dendritic cells, belonging to the mononuclear phagocyte system, reside in the interstitium in line with their sentinel function. In disease, various inflammatory cells infiltrate the interstitium. In the diseased kidney, the cellular content as well as the content of fibers and matrix in the interstitium increases, eventually yielding renal fibrosis.
The cortical interstitium, and in particular the fibroblasts, are often overlooked in conventional tissue preparations of healthy kidneys. However, under inflammatory conditions the interstitial space becomes massively expanded and fibroblasts acquire the myofibroblast phenotype: they express alpha smooth muscle actin (αSMA) and have an increased capacity for production of collagen fibrils and matrix.
In this review based on studies in rats and mice we describe the morphology of interstitial cells of the healthy renal cortex and in various models of diseases. The observations suggest that fibroblasts are the pivotal cells in the renal interstitium.
Problems in studying interstitial cells in the kidney
Following older electron microscopic investigations (Bulger and Nagle 1973; Lemley and Kriz 1991; Takahashi-Iwanaga 1991) only a few reports dealt with the structure of interstitial cells in healthy kidney. The reasons are mostly technical. The study of the cells in the cortical interstitium is difficult since after immersion fixation the interstitial space and the capillaries are usually collapsed whereas tubules are swollen. The light microscopic distinction between fibroblasts, endothelial cells, and dendritic cells is awkward in such tissue preparations. Only perfusion fixation and the use of thin light microscopic (<3 μm) sections and of electron microscopy provide a reasonably safe distinction of fibroblasts and other interstitial cells.
Identification of cortical fibroblasts by immunostainings
The study of renal fibroblasts is further hampered by the lack of marker proteins expressed in all fibroblasts in all functional stages and exclusively in fibroblasts.
Nevertheless, a couple of proteins have been proposed to identify fibroblasts in healthy and diseased kidneys, the most propagated one being the “fibroblast specific protein 1 (FSP1) (Strutz et al. 1995) which corresponds to the small cytoplasmic calcium-binding protein S100A4 and plays a role in cell motility (Garrett et al. 2006, 2008). In the healthy kidney only very few cells in the interstitial space are recognized by antibodies against S100A4/FSP1. This number is by far too small to comprehend the entire population of fibroblasts in the healthy kidney. In the diseased kidney, the incidence of S100A4/FSP1-positive cells increases in parallel with the amount of infiltrating leukocytes (Inoue et al. 2005). In fact, in the healthy as well as in the diseased kidney S100A4/FSP1 co-localizes with leukocyte markers (Inoue et al. 2005; Le Hir et al. 2005; Rossini et al. 2005). S100A4/FSP1 also labels smooth muscle cells of arterioles (Le Hir et al. 2005) and metastasing cells (Davies et al. 1993; Gibbs et al. 1994; Mazzucchelli 2002; Garrett et al. 2006). Recently the stromal cadherin 9 (Thedieck et al. 2007) has been proposed as surface marker for fibroblasts in the human kidney. Cadherin 9 is expressed in a very small population of interstitial cells in healthy kidneys. In the diseased kidney the incidence of labeled cells is higher. Cadherin 9 partially co-localizes with cells labeled by FSP1/S100A4 and with mononuclear cells (Thedieck et al. 2007).
Alpha SMA is expressed in fibroblasts of the adult kidney only under pathological conditions. In other words αSMA allows the identification myofibroblasts but not of the fibroblasts of the healthy kidney.
In contrast with cortical interstitial fibroblasts, fibroblasts of the renal capsule, of the periarterial connective tissue sheath, and of the medulla do not or only weakly express 5′NT.
Developmental aspects of cortical peritubular fibroblasts
The sharing by renal fibroblasts of many proteins ascribed typically to either mesenchymal cells or to epithelial cells might be rooted in the developmental origin of renal fibroblasts. Unlike in other epithelial organs, where epithelial and stromal cells arise from different blastema, the metanephric mesenchyme gives rise to both, epithelia and stroma. Whereas the signaling cascades promoting the transformation of metanephric mesenchymal cells to epithelial cells are extensively studied, little is known about the differentiation of the mesenchymal cells to interstitial fibroblasts (Ekblom and Weller 1991; Alcorn et al. 1999). During embryogenesis part of the metanephric mesenchyme is induced to develop into nephron epithelia. These prospective epithelial cells are surrounded by a population of supporting mesenchymal cells, which express the winged helix transcription factor BF-2. BF-2 seems to determine the fate of this population of mesenchymal cells to maintain the mesenchymal phenotype and to differentiate later to cortical fibroblasts (Hatini et al. 1996). The expression of αSMA apparently depends on signaling by Wnt-4, a member of the Wnt family of secreted signaling molecules. Wnt-4 probably performs this function in the developing kidney by activating the Bmp-4 gene encoding a known differentiation factor for αSMA (Itaranta et al. 2006).
In newborn rats the stromal cells in the kidney display αSMA, vimentin (Marxer-Meier et al. 1998), and nestin, which is associated with vimentin (Sakairi et al. 2007). Yet, they are all negative for 5′NT, although at that time point the enzyme is well detectable in its epithelial locations. In the latter the “mesenchymal” cytoskelettal proteins (αSMA and vimentin) are downregulated. The cortical peritubular fibroblasts (but not the medullary ones) loose αSMA- and vimentin as well as nestin expression gradually during the first two postnatal weeks, simultaneously with the gradual gain of 5′NT. During a short period within the second postnatal week αSMA, vimentin, and 5′NT are co-expressed by the same cells. Three weeks after birth the peritubular cortical fibroblasts in healthy rat kidneys display the adult fibroblast phenotype, i.e. 5′NT in their plasma membrane, but neither vimentin, nor αSMA in the cytoplasm (Marxer-Meier et al. 1998). The fibroblasts in the juxtamedullary layer of the cortex, containing the most mature, oldest generation of nephrons, are the first to gain the “mature” stage. Interestingly, the fibroblasts in the periarterial interstitial tissue maintain during adult life a moderate αSMA- and vimentin expression associated with an extremely weak or undetectable 5′NT expression (Kaissling et al. 1996; Marxer-Meier et al. 1998).These observations strengthen the notion of the dependence of the fibroblast phenotype on the microenvironment.
Cellular constituents of the healthy cortical interstitium
Morphology of fibroblasts
In healthy kidneys the pericaryon of the fibroblasts reveals a stellate profile with clear outlines (Figs.3, 4 and 5). Characteristic for cortical fibroblasts in healthy kidneys is the thin organelle-free cytoplasmic rim around the nucleus and the thin layer of actin filaments immediately under the plasma membrane, well apparent in transmission electron microscopy (Figs. 3, 5) and by staining for F-actin. The cellular processes are connected to the basement membranes by pedicle-like (rats) (Fig. 3) or spine-like (mouse) (Fig. 5) extensions. In these attachments the abundant F-actin filaments are suggestive of stress fibers. In addition to the contacts to basement membranes of capillaries and epithelia the fibroblast processes are also interconnected with each other by intermediate junctions (Fig. 3) (Lemley and Kriz 1991; Kaissling et al. 1996). By bracing the space between vessels, renal corpuscles, and loops of convoluted tubules the fibroblasts form a continuous network through the entire cortex thereby maintaining the tissue architecture.
Protein synthesis apparatus
The synthesis of collagen fibers (types I, III and VI), of unbanded microfilaments and of matrix is part of the scaffolding function by fibroblasts (Lemley and Kriz 1991). This function is mirrored by the prominent apparatus for protein synthesis. Indeed, the extensive rough endoplasmatic reticulum (rER) distinguishes fibroblasts from all other interstitial cells (Figs. 3, 5). The abundant meandering and anastomosing cisterns of the rough ER are particularly densely studded with ribosomes and show large profiles, which are filled by a flocculent, rather electron dense material. Free ribosomes are abundant. The rER cisterns are almost exclusively located in the large cytoplasmic processes of fibroblasts that might be quite distant from the pericaryon. The Golgi fields are usually located together with the rER in the periphery of the cells, where in electron microscopy the formation of collagen fibrils is often well detectable in the direct vicinity of the plasma membrane. This clearly designates the cell as fibroblast (Fig. 5). In “quiescent” fibroblasts the nucleus shows often angular profiles and reveals large amounts of heterochromatin condensations (Fig. 5), whereas in “activated” stages with increased matrix production the nucleus becomes enlarged, rounded and displays much more euchromatin (see below).
Other cell organelles are situated among the rER cisterns. Mitochondria are numerous. Fibroblasts can accumulate lipid droplets. These are not common in cortical fibroblasts of healthy kidneys; yet, they constitute a constant and characteristic feature of fibroblaststs in the medulla, in particular in the inner medulla. They appear also in cortical fibroblasts of anemic kidneys (Kaissling et al. 1993). Lysosomal bodies, which point to a phagocytotic capacity (Lullmann-Rauch 1987; Sundelin and Bohman 1990) are rarely observerd under control conditions.
Morphology of dendritic cells
The broad ramified and flat “veil-like” processes, that often appear perforated like laces are largely devoid of cell organelles and they lack the prominent stuffing with F-actin filaments, characterizing fibroblasts. In addition they have long filiform processes, of approximately the same diameter as the filiform fibroblast processes. In transmission electron microscopy a distinction of fibroblast and DC processes is, however, possible without any marker. Due to the lack of F-actin the processes of dendritic cells appear always much less electron dense than the fibroblast processes (Fig. 7). The lack of the prominent F-actin skeleton is in line with the lack of junctional connections between DCs and tubules or vessels, in contrast to fibroblasts. However, frequently the plasma membranes of dendritic cells and of fibroblasts, or dendritic cells and lymphocytes form spots of membrane adhesions,so-called “kisses” (Fig. 7). Double immuno-staining using 5′NT for fibroblasts and MHC class II for dendritic cells allows an appreciation of the narrow intermingling of both cell types (Fig. 6), suggesting the possibility of extensive cross-talk between them.
Macrophages, lymphocytes and granulocytes
Macrophages and lymphocytes are rarely found in the healthy renal interstitium but they invade the interstitial spaces under inflammatory conditions (Eddy 2005). A large proportion of the invading mononuclear cells display, besides the established marker proteins (e.g. CD 45, CD3, CD4, CD 8; ED1, ED2, CD44), the cytosolic protein S100A4/FSP1 (Le Hir et al. 2005). Neutrophil granulocytes are found occasionally, basophil and eosinophil granulocates and plasma cells are rare in the healthy cortical renal interstitium.
Morphology of myofibroblasts
Any chronic inflammation of the interstitium goes along with the appearance of interstitial cells with a specific phenotype, the myofibroblasts. The myofibroblasts are a landmark of renal fibrosis which is the final common pathway for all chronic kidney diseases and is eventually associated with the loss of kidney function.
Functionally myofibroblasts resemble, on one hand, fibroblasts because of their matrix production and, on other hand, smooth muscle cells because of their contractility due to the presence of αSMA.
The hallmark of myofibroblasts is the presence of the αSMA (Desmouliere et al. 2005; Hinz 2007; Hinz et al. 2007) which builds bundles of microfilaments with dense bodies (Fig. 9). These bundles may run across the cytoplasm and are not necessarily associated with the plasma membrane, as the F-actin filaments in quiescent fibroblasts, and they are concentrated in cellular processes. The frequency of intercellular attachments by intermediate junctions and of hemidesmosome-like attachments to basement membranes of tubules and vessels is much higher than in quiescent fibroblasts.
Furthermore, the intermediate filament protein vimentin is expressed in myofibroblasts, but not in quiescent fibroblasts of adults. Desmin, another intermediate filament protein, has been described to be present in myofibroblasts (Maxwell et al. 1997); however, we and others (Essawy et al. 1997) never detected desmin in renal myofibroblasts.
Role of fibroblasts in regulation of regional renal hemodynamics and in renal erythropoietin production
The presence of 5′NT in cortical fibroblasts indicates that adenosine can be produced in the cortical interstitium. 5′NT is responsible for conversion of extracellular 5′-AMP to adenosine (Eltzschig et al. 2003). The bulk of extracellular 5′-AMP in kidney most probably originates from tubular cells and any disturbances in the energy balance of tubules might increase 5′-AMP levels in the interstitial space (Le Hir and Kaissling 1993). Thus, hydrolysis of extracellular 5′-AMP by fibroblast 5′NT would elicit local adenosine-mediated responses, like adjustment of blood flow. Under inflammatory conditions also infiltrating activated polymorphonuclear neutrophils may contribute to the pool of interstitial 5′-AMP and adenosine might modulate inflammation (Linden 2006; Yang et al. 2006).
Extracellular adenosine has been widely implicated in adaptive responses to hypoxia (Eltzschig et al. 2003; Grenz et al. 2007). The role of the fibroblasts in the adaptive response to systemic hypoxia is further highlighted by the discovery that renal erythropoietin is synthesized exclusively by 5′NT-positive renal fibroblasts (Bachmann et al. 1993; Maxwell et al. 1993; Fisher et al. 1996). The kidneys adjust their erythropoietin production to changes in oxygen supply (Maxwell 2003). Regulation of erythropoietin production and transcription of erythropoietin are mediated by hypoxia-inducible factor (HIF Hif-2α) (Maxwell 2003) which has been located as well to 5′NT-positve fibroblasts (Rosenberger et al. 2003, 2005).
In kidneys of normoxic rats and mice erythropoietin-mRNA is detected only in the 5′NT-positive peritubular fibroblasts in the deep cortex. At this level 5′NT expression by peritubular fibroblasts is highest and it decreases towards the renal surface. Both, hypoxia and anemia, involve a progressive recruitment of additional interstitial fibroblasts for erythropoietin production in a pattern that spreads from the deep cortex toward the capsule (Eckardt et al. 1993). In parallel the 5′NT activity by the peritubular fibroblasts increases towards the capsule (Kaissling et al. 1993). This points again to an obvious link between 5′NT expression (or adenosine) and erythropoietin synthesis, which also has been confirmed in a fibroblast cell line (Plotkin and Goligorsky 2006). The details of this link are, however, still inexplicit. Interestingly, Ito-cells (or fat-storing cells) in the liver, which express 5′NT (Schmid et al. 1994) and which resemble morphologically in all respects to the renal cortical fibroblasts, also produce erythropoietin (Maxwell et al. 1994).
Role of fibroblasts in chronic renal failure
Fibrosis is the morphological hallmark of chronic renal failure. It is characterized by an exaggerated accumulation of fibrous collagens, by an increased incidence of interstitial cells in the renal interstitium and by the emergence of myofibroblasts.
Origin of fibroblasts/myofibroblsts in renal fibrosis
Until the late 1990s it was generally assumed that myofibroblasts developed by proliferation and differentiation from resident fibroblasts (Bulger and Nagle 1973). Yet, this concept has been challenged in more recent studies. The difficulty to detect interstitial fibroblasts in the kidney, discussed above, resulted in the erroneous belief that fibroblasts are sparse in the healthy kidney. Therefore, sources other than the resident fibroblasts population were proposed for the numerous myofibroblasts found in various models of kidney disease. Three origins have been considered.
Epithelial-mesenchymal transition (EMT)
The hypothesis that in renal diseases, fibroblasts are generated via epithelial-mesenchymal transition (EMT) grounds on two observations. First, in many instances tubular cells lose some of their epithelial features and acquire a “mesenchymal” appearance. Second, in such instances some tubular cells express proteins, which are considered specific for fibroblasts/myofibroblasts, namely fibroblasts-specific protein-1 (FSP1) (Zeisberg et al. 2001; Iwano et al. 2002), heat shock protein HSP47 (Iwano et al. 2002) and αSMA (Liu 2004). However the interpretation of the latter findings requires some caution. FSP1 is not a reliable fibroblast marker (Le Hir et al. 2005). HSP47 is supposed fibroblast-specific because it is a chaperone for collagens. However it is not a chaperone for fibrous collagens only. Accordingly, accumulation of HSP47 was observed in tubular cells, which produced excessive amounts of collagen type IV in a model of partial nephrectomy (Kimura et al. 2005). The occasional expression of αSMA in some tubular cells in models or renal disease is intriguing. However, it remains to be proven that αSMA-positive tubular cells acquire the most characteristic function of fibroblasts/myofibroblasts, namely the production of collagen fibers.
In some experimental models, like 5/6 nephrectomy (Kimura et al. 2005), DCT lesion with thiazide diuretics (Le Hir et al. 2005) and UUO (Picard et al. 2008), accumulation of αSMA-positive interstitial cells was observed whereas the tubules remained αSMA-negative. The same observation was made in biopsies of patients with nephritic syndrome (Kuusniemi et al. 2005). At least in those instances myofibroblasts did not appear to originate via EMT.
Bone marrow (BM) origin
There is evidence from animal models and from studies with BM transplant recipients that circulating precursors from BM origin give rise to fibroblasts/myofibroblasts in various organs in adults (Quan et al. 2004; Desmouliere et al. 2005; Hinz et al. 2007). Studies using transgenic rodents have shown a BM origin of a limited fraction of renal myofibroblasts after ischemia/reperfusion injury (Broekema et al. 2007) and in the model of adriamycin-induced chronic renal failure (Li et al. 2007).
Proliferation and differentiation of resident fibroblasts
It is widely accepted that under conditions leading to fibrosis in various organs fibroblasts transform into myofibroblasts, and both fibroblasts and myofibroblasts proliferate (Desmouliere et al. 2005; Hinz 2007; Hinz et al. 2007). Apparently, that straightforward view of the cellular basis of fibrosis has not much attracted the interest of investigators in the field of renal pathophysiology. Thus, very few studies have analyzed the response of resident fibroblasts to renal diseases.
Maxwell et al. (1997) previously observed an expression of αSMA in 5′NT-positive fibroblasts in UUO and after mechanical lesions (needlestick) in mice kidneys, which was associated with, first, a decrease in 5′NT, and second, a reduced capacity of the fibroblasts to respond with erythropoietin production under hypoxia. Proliferation of interstitial cells has also been described in the model of UUO (Duymelinck et al. 2000) and in another model of renal fibrosis induced by aristolochic acid (Pozdzik et al. 2008).
Mechanical stress seems to play a central role in the transformation of fibroblasts into myofibroblasts after uranyl acetate-induced tubular lesion (Fujigaki et al. 2005). The dilatation or shrinkage of tubules likely exerts a tension on the fibroblasts that are directly attached to damaged tubules. Mechanical forces could be transmitted to further fibroblasts by cell-cell contacts.
Fibrosis and nephron loss: which comes first?
The often massive extent of fibrosis in chronic renal failure unavoidably suggests that fibrosis might be responsible for the destruction of nephrons. Accordingly, in the case of glomerular diseases, the most common cause of chonic renal failure, it has been proposed that proteinuria elicits fibrosis via the release of profibrotic factors by tubular cells. Tubular degeneration would then be a consequence of fibrosis (Remuzzi and Bertani 1998; Meyer 2003; Eddy 2005). That view is widely accepted but not unchallenged. Indeed, morphological observations suggest that tubular degeneration may be a direct consequence of the encroachment of the glomerular lesion on the tubule (Kriz et al. 2001, 2003; Bertani et al. 2002; Le Hir and Besse-Eschmann 2003; Kriz and LeHir 2005).
In the vast majority of chronic renal diseases there is a mix of tubular, glomerular and vascular alterations mostly accompanied by interstitial inflammation and fibrosis. Obviously, in such situations, it is difficult to decide whether the activation and transformation of fibroblasts is the cause or the consequence of tubular lesions. Therefore it is particularly interesting that in a few models fibrosis appears to be induced by a primary tubular insult. As described above the lesion induced by thiazide diuretics in the DCT in the rat induces an inflammatory reaction and a transformation of fibroblasts into myofibroblasts, which are strictly associated with the injured DCT (Loffing et al. 1996; Le Hir et al. 2005). In the kidney only DCT cells express the transporter NCC, which is the binding site for thiazide diuretics. In this model the fibrotic process came to a halt when under continued metolazone application the NCC-expressing cells were eliminated by apoptosis. A similar specific association of fibrotic processes with injured tubules was found in other models, in which there is little doubt that the experimental protocol affects primarily tubular cells. Local interstitial fibrosis was observed mainly around distal tubules after treatment with lithium (Zhu et al. 1996). Substances, which are toxic to specific segments of the proximal tubules, induced fibrosis specifically around the affected segments (Fujigaki et al. 2005; Pozdzik et al. 2008).
In those models of primary tubular lesion the association in localization and in time between tubular damages and interstitial responses suggests that fibrosis might represent a repair process. Interestingly, inhibition of transformation of fibroblasts into myofibroblasts after induction of acute renal failure with uranyl acetate resulted in aggravation of renal dysfunction and inhibition of renal repair (Fujigaki et al. 2005).
The cortical fibroblasts contribute essentially to renal functions, not only as skeleton of the organ but also by producing regulatory substances such as, e.g. adenosine and erythropoietin. In response to tubular lesions fibroblasts proliferate and transform into myofibroblasts, thus initiating fibrosis. The role of that process in renal diseases remains to be elucidated. On one hand, fibrosis might contribute to the deterioration of renal function. On other hand, in the same way as in other organs, fibrosis in the kidney might contribute to repair by containing inflammation and by preserving the organ architecture. Experimental models in which fibrosis is induced by lesions to specific tubular cell types appear particularly promising for the study of the cross-talk between tubular cells and interstitial cells in disease.
- Broekema M, Harmsen MC, van Luyn MJ, Koerts JA, Petersen AH, van Kooten TG, van Goor H, Navis G, Popa ER (2007) Bone marrow-derived myofibroblasts contribute to the renal interstitial myofibroblast population and produce procollagen I after ischemia/reperfusion in rats. J Am Soc Nephrol 18:165–175PubMedCrossRefGoogle Scholar
- Eltzschig HK, Ibla JC, Furuta GT, Leonard MO, Jacobson KA, Enjyoji K, Robson SC, Colgan SP (2003) Coordinated adenine nucleotide phosphohydrolysis and nucleoside signaling in posthypoxic endothelium: role of ectonucleotidases and adenosine A2B receptors. J Exp Med 198:783–796PubMedCrossRefGoogle Scholar
- Fujigaki Y, Muranaka Y, Sun D, Goto T, Zhou H, Sakakima M, Fukasawa H, Yonemura K, Yamamoto T, Hishida A (2005) Transient myofibroblast differentiation of interstitial fibroblastic cells relevant to tubular dilatation in uranyl acetate-induced acute renal failure in rats. Virchows Arch 446:164–176PubMedCrossRefGoogle Scholar
- Hashizume T, Imayama S, Hori Y (1992) Scanning electron microscopic study on dendritic cells and fibroblasts in connective tissue. J Electron Microsc (Tokyo) 41:434–437Google Scholar
- Hinz B, Phan SH, Thannickal VJ, Galli A, Bochaton-Piallat ML, Gabbiani G (2007) The myofibroblast. One function, multiple origins. Am J PatholGoogle Scholar
- Kaissling B, Hegyi I, Loffing J, Le Hir M (1996) Morphology of interstitial cells in the healthy kidney. Anat Embryol (Berl) 193:303–318Google Scholar
- Kruger T, Benke D, Eitner F, Lang A, Wirtz M, Hamilton-Williams EE, Engel D, Giese B, Muller-Newen G, Floege J, Kurts C (2004) Identification and functional characterization of dendritic cells in the healthy murine kidney and in experimental glomerulonephritis. J Am Soc Nephrol 15:613–621PubMedCrossRefGoogle Scholar
- Mc Dermott R, Ziylan U, Spehner D, Bausinger H, Lipsker D, Mommaas M, Cazenave JP, Raposo G, Goud B, de la Salle H, Salamero J, Hanau D (2002) Birbeck granules are subdomains of endosomal recycling compartment in human epidermal Langerhans cells, which form where Langerin accumulates. Mol Biol Cell 13:317–335PubMedCrossRefGoogle Scholar
- Picard N, Baum O, Vogetseder A, Kaissling B, Le Hir M (2008) Origin of renal myofibroblasts in the model of unilateral ureter obstruction in the rat. Histochem Cell BiolGoogle Scholar
- Yang D, Zhang Y, Nguyen HG, Koupenova M, Chauhan AK, Makitalo M, Jones MR, St Hilaire C, Seldin DC, Toselli P, Lamperti E, Schreiber BM, Gavras H, Wagner DD, Ravid K (2006) The A2B adenosine receptor protects against inflammation and excessive vascular adhesion. J Clin Invest 116:1913–1923PubMedCrossRefGoogle Scholar