Dysfunction of fibroblasts of extrarenal origin underlies renal fibrosis and renal anemia in mice

In chronic kidney disease, fibroblast dysfunction causes renal fibrosis and renal anemia. Renal fibrosis is mediated by the accumulation of myofibroblasts, whereas renal anemia is mediated by the reduced production of fibroblast-derived erythropoietin, a hormone that stimulates erythropoiesis. Despite their importance in chronic kidney disease, the origin and regulatory mechanism of fibroblasts remain unclear. Here, we have demonstrated that the majority of erythropoietin-producing fibroblasts in the healthy kidney originate from myelin protein zero–Cre (P0-Cre) lineage-labeled extrarenal cells, which enter the embryonic kidney at E13.5. In the diseased kidney, P0-Cre lineage-labeled fibroblasts, but not fibroblasts derived from injured tubular epithelial cells through epithelial-mesenchymal transition, transdifferentiated into myofibroblasts and predominantly contributed to fibrosis, with concomitant loss of erythropoietin production. We further demonstrated that attenuated erythropoietin production in transdifferentiated myofibroblasts was restored by the administration of neuroprotective agents, such as dexamethasone and neurotrophins. Moreover, the in vivo administration of tamoxifen, a selective estrogen receptor modulator, restored attenuated erythropoietin production as well as fibrosis in a mouse model of kidney fibrosis. These findings reveal the pathophysiological roles of P0-Cre lineage-labeled fibroblasts in the kidney and clarify the link between renal fibrosis and renal anemia.

Erythropoietin (EPO) is a hormone that is essential for the production of red blood cells (1). EPO is mainly produced in the kidney after birth, and its production is severely reduced in patients with chronic kidney disease (CKD) (2) with renal anemia. The cloning of the EPO gene and the production of recombinant human EPO revolutionized the management of patients with CKD, providing an opportunity for safe long-term anemia correction, which improves cognitive function, quality of life, exercise capacity, and cardiac function (3). Nowadays, an increasing number of patients with CKD receive erythropoiesis-stimulating agents (ESAs), and annual US prescription sales for ESAs reached 10 billion dollars in 2006. In addition to the hematopoietic function of EPO, increasing evidence suggests that EPO plays a role as a tissue-protective factor in various nonhematopoietic tissues (4), raising the possibility that it will be prescribed for various types of tissue injuries in the near future.

Studies using in situ hybridization and the transgenic mice approach indicate that EPO is mainly produced by the interstitial fibroblasts in the deep cortex and the outer medulla in the kidney (5, 6). Although the essential role of EPO has been well recognized, the identity and developmental origin of EPO-producing cells in the kidneys remain unknown. This has hindered both the generation of EPO-producing cells from stem cells as well as the clarification of the molecular mechanisms underlying the defective production of EPO in patients with CKD — EPO-producing cells may lose their ability to produce EPO after injury, or, alternatively, EPO-producing cells may be lost during CKD.

Fibroblasts are interstitial mesenchymes that structurally support epithelia by producing ECM. Recently, the tissue-resident fibroblasts have attracted considerable attention because of their prominent roles in fibrosis, endocrine, organogenesis, and tumorigenesis (7). In chronic kidney injury, sustained inflammation accompanies the proliferation of interstitial fibroblasts and myofibroblasts (8), leading to renal fibrosis, which is the final common pathway for all CKD and eventually leads to renal failure (9). The origin of scar-producing myofibroblasts is of prime importance in understanding the mechanisms of renal fibrosis, which remain controversial. The contribution of bone marrow–derived fibrocytes (10) and the epithelial-mesenchymal transition (EMT) of injured tubular epithelial cells (11) has been proposed in some studies, but other groups have reported conflicting results (12–16).

Myelin protein zero (P0) is expressed in migrating neural crest cells in the early embryonic stages as well as in Schwann cells, which also originate from neural crest, in later stages (17). In P0-Cre mice, the activation of the P0 promoter induces the Cre-mediated recombination of reporter alleles, permanently tagging the cells arising from P0-Cre–expressing precursors with the expression of indicator genes.

Here, we demonstrate the occurrence of P0-Cre lineage–labeled fibroblasts in the kidney interstitium and their 2 different characteristics — they produce EPO, which possesses hematopoietic and tissue-protective functions in healthy kidney, and they transdifferentiate into scar-producing myofibroblasts and lose their EPO-producing activity after kidney injury, leading to renal fibrosis and renal anemia. We further demonstrate that the attenuated EPO production in scar-producing myofibroblasts can be restored through the administration of neuroprotective reagents, such as low-dose dexamethasone and neurotrophins. A similar effect was also obtained by the administration of a renoprotective reagent HGF. Moreover, the in vivo administration of tamoxifen, a selective estrogen receptor modulator (SERM), restored attenuated EPO production as well as fibrosis in a mouse model of kidney fibrosis.

These findings provide insights that we believe to be novel into the mechanisms of renal fibrosis and renal anemia in CKD and provide clues that assist in the development of therapeutic approaches to treat these disorders.

Most fibroblasts in the kidney arise from P0-Cre–expressing precursors. P0-Cre mice were mated to 4 indicator mouse lines: floxed EGFP (18), R26R (19), R26ECFP (20), and R26tdRFP mice (21). Immunostaining of the kidneys of P0-Cre/R26ECFP mice using anti-GFP antibody revealed numerous enhanced cyan fluorescent protein–positive (ECFP⁺) cells in the cortex and outer medulla (Figure 1, A–C). The kidneys of R26ECFP mice without the P0-Cre allele contained no ECFP⁺ cells, confirming the specificity of staining (data not shown). LacZ staining in P0-Cre/R26R kidney also revealed the same results (Figure 1, D–F). These P0-Cre fate-mapped cells are located in the interstitium, the space between the tubules and nutritive peritubular capillaries, exhibiting a unique stellar shape with numerous projections.

Figure 1

Most of the fibroblasts in the kidney arise from P0-Cre–expressing precursors. (A) Schematic drawing of the kidney, showing the localization of P0-Cre fate-mapped cells in the cortex and outer medulla. (B and C) ECFP⁺ cells were detected in the interstitium of the kidneys of P0-Cre/R26ECFP mice. A higher-magnification view is shown in C. (D–F) LacZ staining of P0-Cre/R26R kidney. (D and E) Note the similar distribution of LacZ⁺ cells in P0-Cre/R26R kidneys with ECFP⁺ cells in P0-Cre/R26ECFP mice. A higher-magnification view is shown in E. (F) No LacZ⁺ cells were observed in the kidneys of R26R mice without the P0-Cre allele. (G–I) Double immunostaining of ECFP⁺ cells in the P0-Cre/R26ECFP kidney. (G) Most, if not all, of the ECFP⁺ cells were negative for PECAM. (H and I) ECFP⁺ cells (H) in the kidney were positive for PDGFR-β, the fibroblast marker, whereas ECFP⁺ cells (I) in the cortex were also positive for CD73/5′NT, the marker for cortical fibroblasts. (J) Graph illustrating the proportion of PDGFR-β⁺ interstitial cells in the cortex and outer medulla coexpressing ECFP (% ECFP/PDGFR-β cells) and the proportion of ECFP⁺ interstitial cells coexpressing PDGFR-β (% PDGFR-β/ECFP cells). (K) In neonatal kidneys of P0-Cre/floxed-EGFP mice most of the EGFP⁺ cells were also positive for p75, a neural crest marker. Scale bars: 10 μm (B–I); 100 μm (K).

The interstitium of the healthy adult kidney is composed of heterogeneous cells, including macrophages, dendritic cells, and peritubular capillaries (9). Most, if not all, P0-Cre fate-mapped cells in the kidney did not express the macrophage marker CD11b (Supplemental Figure 1A; supplemental material available online with this article; doi: 10.1172/JCI57301DS1), F4/80 (Supplemental Figure 1B), the dendritic cell marker major histocompatibility complex class II (Supplemental Figure 1C and Supplemental Figure 7), or the common leukocyte antigen CD45 (Supplemental Figure 7). Although a previous report indicated that endothelial cells in certain tissues are of neural crest origin (22), ECFP⁺ cells in the kidney of P0-Cre/R26ECFP mice did not express PECAM/CD31, a marker for endothelial cells (Figure 1G and Supplemental Figure 7). Very few ECFP⁺ cells expressed tyrosine hydroxylase, a marker for sympathetic nerve fibers, whereas most ECFP⁺ cells did not (Supplemental Figure 2). Instead, ECFP⁺ cells expressed PDGFR-β, a marker for interstitial fibroblasts in the kidney (refs. 9, 23, and Figure 1H). ECFP⁺ cells in the cortex expressed CD73/ecto-5′-nucleotidase (CD73/5′NT), a marker for cortical fibroblasts (ref. 9 and Figure 1I). More than 98% of ECFP⁺ interstitial cells expressed PDGFR-β, whereas more than 99% of PDGFR-β⁺ cells in the interstitium of the cortex and outer medulla expressed ECFP (Figure 1J). Furthermore, EGFP⁺ cells in the neonatal kidneys of P0-Cre/floxed-EGFP mice expressed p75, a marker for the neural crest (ref. 24 and Figure 1K). These results indicate that most, if not all, of the kidney fibroblasts in the cortex and outer medulla are fate-mapped with P0-Cre.

P0-Cre lineage-labeled cells enter the embryonic kidney during nephrogenesis. To investigate the route and stage of migration of P0-Cre lineage-labeled cells into the kidney, we analyzed P0-Cre/R26ECFP embryos serially from E11.5, when nephrogenesis is initiated, to the neonatal stage. ECFP⁺ cells were first detected from E13.5 onward in metanephric kidneys. These cells were predominantly located along the outer capsule and the ureter of the kidney (Figure 2, A–C), which is possibly indicative of their routes of migration. The number of ECFP⁺ cells in the metanephric kidney dramatically increased in later stages of development, and the ECFP⁺ cells began to express PDGFR-β (Figure 2, D–F).

Figure 2

P0-Cre lineage-labeled cells in the developing kidney. (A–C) ECFP⁺ cells were first observed in the kidneys of P0-Cre/R26ECFP mice at E13.5. ECFP⁺ cells were predominantly located along the outer capsule and the ureter (arrows) of the kidney. (D–F) The number of ECFP⁺ cells in the kidneys of P0-Cre/R26ECFP mice increased at E16.5, and (F) the cells began to express PDGFR-β. The boxed region in E is shown at higher magnification in F. (G and H) ECFP⁺ cells populating the cortex appeared to surround the territory of Six2⁺ renal progenitor cells. (I) ECFP⁺ cells in the kidney were negative for class III β-tubulin. K, kidney. Scale bars: 100 μm (A, B, D, E, and I); 10 μm (C and F–H).

These ECFP⁺ cells populating the cortex appeared to avoid the territory of the Six2⁺ renal progenitor cells, which give rise to all nephron segments in the kidney (ref. 25 and Figure 2, G and H). Furthermore, these ECFP⁺ cells did not express class III β-tubulin, a neuron-specific marker (Figure 2I). Importantly, immunostaining of the P0-Cre embryo failed to identify any Cre⁺ cells in the kidneys throughout development (Supplemental Figure 3, C–E), whereas it detected the expression of Cre in the migrating neural crest and the dorsal root ganglion (Supplemental Figure 3, A and B). Immunostaining of P0 protein also failed to detect any P0 protein in the kidneys throughout development (Supplemental Figure 4, A–C). We further demonstrated that the expression of P0 and Cre mRNA was almost undetectable in embryonic kidneys (Supplemental Figure 8), whereas the number of P0-Cre–labeled cells increased substantially during development. Thus, it is highly unlikely that the ectopic expression of Cre in the kidney accounts for ECFP⁺ cells in the kidney.

P0-Cre lineage-labeled fibroblasts produce EPO. Renal EPO-producing cells are thought to be peritubular interstitial cells that exhibit a stellar shape with projections and express neuronal markers (6). Given these features of EPO-producing cells, we examined whether EPO-producing cells arise from P0-Cre–expressing precursors. We used Epo-GFP mice, a BAC transgenic mouse line that expresses GFP as a reporter under the control of a 180-kb mouse Epo gene locus, and identified the EPO-producing population in the kidney (6).

GFP expression in the kidney of Epo-GFP mice was detected in the interstitium and was colocalized with the fibroblast markers CD73/5′NT and PDGFR-β (Figure 3, A–C). Analysis of the kidneys of P0-Cre/R26R/Epo-GFP mice revealed that most of the GFP⁺ cells were positive for LacZ (Figure 3, D and E), supporting the idea that EPO-producing cells originate from the P0-Cre–expressing precursors. We further analyzed the kidneys of P0-Cre/R26tdRFP/Epo-GFP mice and found that more than 75% of GFP⁺ cells were also positive for tandem-dimer red fluorescent protein (tdRFP) (Figure 3F).

Figure 3

P0-Cre lineage-labeled fibroblasts produce EPO. (A–C) Stellate-shaped GFP⁺ cells with projections were detected (A) in the interstitium of the kidneys of Epo-GFP mice and were positive for (B) CD73 and (C) PDGFR-β. (D and E) Analysis of the kidney of P0-Cre/R26R/Epo-GFP mice revealed that GFP⁺ cells were also positive for LacZ. (F) Analysis of the kidney of P0-Cre/R26tdRFP/Epo-GFP mice revealed that most GFP⁺ cells were also positive for tdRFP. Scale bars: 10 μm. (G) Sorting EGFP⁺ cells (P5) out of EGFP^– cells (P4) from the kidneys of adult P0-Cre/floxed-EGFP mice. RT-PCR analysis of these populations revealed high expression of Epo and p75 in EGFP⁺ cells but not in EGFP^– cells. Cells were stained either with or without anti-GFP antibody. (H) Sorting p75⁺ cells (P5) out of p75^– cells (P4) from the kidneys of P0-Cre/R26ECFP mice at 2 weeks of age. Cells were stained either with or without anti-p75 antibody. Quantitative PCR analysis demonstrated that the expression of p75, Epo, and ECFP was higher in p75⁺ cells. Expression of p75 was normalized to that of Gapdh and expressed relative to that in p75^– cells.

EPO expression in P0-Cre fate-mapped cells was further examined using EGFP⁺ and EGFP^– cells FACS-purified from kidneys of adult P0-Cre/floxed-EGFP mice. PCR analysis of these populations revealed the expression of EPO and p75 in EGFP⁺ cells but not in the EGFP^– population (Figure 3G). Sorting p75⁺ cells in the kidneys of P0-Cre/R26ECFP mice at 2 weeks of age also revealed higher expression of p75, EPO, and ECFP in p75⁺ cells than in p75^– cells (Figure 3H).

P0-Cre lineage-labeled fibroblasts transdifferentiate into myofibroblasts and contribute to fibrosis. To investigate the response of P0-Cre lineage-labeled fibroblasts to injury, we performed unilateral ureteral obstruction (UUO), a well-established model for renal fibrosis (26), using P0-Cre/R26ECFP mice. Five days after UUO, the number of ECFP⁺ cells had markedly increased in the interstitium of the operated kidney, in parallel with the progression of interstitial fibrosis (Figure 4, A–D). FACS analysis also revealed a substantial increase in the number of EGFP⁺ cells in the operated kidney in P0-Cre/floxed-EGFP mice (Figure 4, E and F). It is unlikely that the increase in ECFP⁺ or EGFP⁺ cells is due to the ectopic expression of Cre or P0 protein in the kidney, because neither Cre protein nor P0 protein was detectable in the operated kidney (Supplemental Figure 3F and Supplemental Figure 4D). We further demonstrated that the expression of P0 and Cre mRNA was almost undetectable in fibrotic kidneys (Supplemental Figure 8), whereas the number of P0-Cre–labeled cells increased substantially during kidney fibrosis. Immunostaining of the operated kidney revealed the presence of ECFP/Ki67 double-positive cells, indicating that the increase in the number of ECFP⁺ cells in the operated kidney is at least partially attributable to the proliferation of ECFP⁺ cells (Figure 4G).

Figure 4

P0-Cre lineage-labeled fibroblasts transdifferentiate into myofibroblasts and contribute to fibrosis. (A and B) Five days after obstruction of the ureter, (B) interstitial fibrosis was prominent in the operated kidney (A) compared with that in the control kidney. (C and D) The number of ECFP⁺ cells was markedly increased in the operated kidneys of P0-Cre/R26ECFP mice compared with that in the control kidney. (E and F) FACS analysis also confirmed the increase in the number of EGFP⁺ cells in the operated kidneys of P0-Cre/floxed-EGFP mice. Cells were stained with anti-GFP antibody. (G) ECFP/Ki67 double-positive cells were abundant in the operated kidneys of P0-Cre/R26ECFP mice. (H and I) Most ECFP⁺ cells were also positive for (H) PDGFR-β and (I) α-SMA. (J) Graph illustrating the proportion of PDGFR-β⁺ or α-SMA⁺ interstitial cells coexpressing ECFP (% labeled/ECFP cells) and the proportion of interstitial ECFP⁺ cells coexpressing either PDGFR-β or α-SMA (% ECFP/labeled cells). (K and L) The number of ECFP⁺ cells was also markedly increased in the kidneys of P0-Cre/R26ECFP mice after (K) folic acid nephrotoxicity and (L) severe ischemic reperfusion injury. Most ECFP⁺ cells were also positive for α-SMA. Scale bars: 10 μm.

Another possibility is that macrophages and dendritic cells infiltrated the operated kidneys in response to inflammation. This would account for the increase in the number of ECFP⁺ cells in the operated kidneys. However, ECFP⁺ cells did not overlap with the inflammatory cells, which infiltrated the operated kidneys in response to inflammation (Supplemental Figure 1, D–F). Instead, these ECFP⁺ cells were almost exclusively positive for PDGFR-β (Figure 4H), a marker for fibroblasts, and α-SMA (Figure 4I), a marker for myofibroblasts. Quantitative analysis demonstrated that 97% of PDGFR-β⁺ interstitial cells as well as 94% of α-SMA⁺ myofibroblasts were positive for ECFP, whereas 97% and 93% of ECFP⁺ cells were positive for PDGFR-β and α-SMA, respectively (Figure 4J). Similar results were obtained in other models of renal fibrosis — folic acid nephropathy (Figure 4K) and severe ischemic reperfusion injury (Figure 4L). These results indicate that P0-Cre lineage-labeled fibroblasts are the main source of scar-producing myofibroblasts, contributing to fibrosis after kidney injury.

EPO-producing fibroblasts also transform into myofibroblasts at the cost of EPO production. Regardless of the cause of the damage, diseased kidneys invariably suffer from fibrosis with reduced EPO production. The mechanism underlying this defective production of EPO in the diseased kidney remains unknown (27). In situ hybridization detected Epo mRNA in the interstitium of the control kidney (Figure 5A, left) but not in the operated kidney at day 14 of UUO (Figure 5A, middle). Suppression of Epo mRNA in the operated kidney was evident as early as 12 hours after ligation of the ureter (Figure 5B). Interestingly, the attenuated production of Epo mRNA in the fibrotic kidney was restored by the induction of anemia (Figure 5C and Supplemental Figure 10). Similarly, GFP⁺ cells were almost undetectable in the fibrotic kidneys of Epo-GFP mice, whereas many GFP⁺ cells emerged in the fibrotic kidneys after the induction of anemia (Figure 5D). These GFP⁺ cells in the fibrotic kidneys were positive for α-SMA (Figure 5D), whereas GFP⁺ cells in the healthy kidneys were negative for α-SMA (Supplemental Figure 5). These results indicate that transdifferentiated myofibroblasts still possess the capacity to produce EPO after the induction of anemia.

Figure 5

EPO-producing fibroblasts also transform into myofibroblasts, hindering EPO production. (A) In situ hybridization demonstrated that Epo mRNA detected in the control kidney was almost undetectable in the operated kidneys at day 14 of UUO. Hybridization with a sense probe did not elicit any signals in the control kidney. AS, antisense; S, sense. (B) Expression of Epo mRNA in the operated kidney decreased significantly, as early as 12 hours after ligation of the ureter (n = 5/group, except for the group sacrificed at day 5 [d5], in which n = 4). NT, nontreated. (C) The suppression of Epo mRNA in the operated kidneys was reversed after the induction of anemia. Each dot represents renal expression of Epo in an individual mouse (n = 5 in UUO group; n = 9 in UUO plus anemia group). (D) Undetectable GFP⁺ cells in the operated kidneys of Epo-GFP mice became detectable after the induction of anemia and colocalized with α-SMA. (E) The expression of Epo in cultured myofibroblasts was augmented by the administration of low-dose dexamethasone, neurotrophins, and HGF. Vehicle used was 0.05% ethanol for the control of dexamethasone. Data are representative of results of 5 independent experiments. (F) Expression of Epo mRNA in UUO kidneys was significantly restored, whereas the expression of Col1a1 and fibronectin (Fn1) mRNA was decreased by tamoxifen treatment (n = 5). Vehicle used was 100 μl of 10% ethanol in sunflower oil. Expression of Epo, Col1a1, and Fn1 was normalized to that of Gapdh and expressed relative to that in (B and E) the nontreated group or (F) vehicle-treated group. Scale bars: 10 μm.

Taken together, the results indicate that the cause of renal anemia is not the loss of EPO-producing cells but is at least partially due to the transdifferentiation of EPO-producing cells into scar-producing myofibroblasts that express a smaller amount of EPO (Figure 6).

Figure 6

Hypothetical model for 2 different characteristics of P0-Cre lineage-labeled fibroblasts in the kidney, with a possible therapeutic implication for treatment of renal anemia. P0-Cre lineage-labeled fibroblasts in the kidney produce EPO, which possesses hematopoietic and tissue-protective functions in healthy kidneys. In contrast, they transdifferentiate into scar-producing myofibroblasts and lose EPO-producing activity after kidney injury, leading to renal fibrosis and renal anemia. We further demonstrated that the administration of various neuroprotective and renoprotective reagents restores the EPO-producing ability in fibrotic kidneys.

These findings led us to seek the possible stimuli to restore the EPO-producing ability in myofibroblasts. Because EPO-producing cells in the kidney possess both renal and neural characteristics, we administered renoprotective and neuroprotective agents to cultured primary myofibroblasts obtained from fibrotic kidneys to determine whether or not the EPO-producing ability could be restored. Among the agents that we tested, we found that the administration of neuroprotective agents, such as low-dose dexamethasone, neurotrophins (neurotrophin-3 [NT-3], NT-4, and brain-derived neurotrophic factor [BDNF]), and the renoprotective agent, HGF, increased EPO production in myofibroblasts (Figure 5E and Supplemental Figure 11). We sought other possible reagents to alleviate the loss of EPO in fibrotic kidney in vivo and found that the administration of tamoxifen, a SERM, restored the attenuated expression of EPO in fibrotic kidney (Figure 5F). SERMs activate the estrogen receptors and function as estrogen agonists in some tissues, such as bone and kidneys, and as estrogen antagonists in other tissues, such as breast. The administration of tamoxifen also restored fibrosis significantly (Figure 5F). These results indicate that the attenuated production of EPO in scar-producing myofibroblasts can be restored (Figure 6).

In the present study, we demonstrated that EPO-producing cells in healthy kidney and scar-producing myofibroblasts during fibrosis originate from the same P0-Cre lineage-labeled extrarenal cells, which enter the embryonic kidney at E13.5 to become renal fibroblasts and transit from one another depending on the condition of the kidney. A greater understanding of the mechanisms that control the balance between these 2 functions will lead to better treatment for patients with CKD. More importantly, the reversibility of EPO production in the fibrotic kidney (Figure 5, C–F) raised the possibility of a therapeutic approach toward renal anemia (Figure 6).

Neurotrophins are reported to promote neuronal growth and survival in many neuronal populations through the activation of Trk receptor tyrosine kinases (28–30). Furthermore, NT-3 rescues neuronal precursors from apoptosis and promotes neuronal differentiation in embryonic kidney explants (31). Given the origin of EPO-producing cells from P0-Cre–expressing precursors, it is plausible that neurotrophins restore EPO-producing ability in transdifferentiated myofibroblasts. Recently, pharmaceutical strategies to stimulate neurotrophin signaling established proof of principle of these as effective therapeutics for Alzheimer disease (30). Among these strategies, small molecule activators of neurotrophin are appealing and should be tested in the model of renal anemia in the future.

Glucocorticoid exhibits both protective and destructive effects in the nervous system. In excess, glucocorticoid produces neuronal death, whereas low-dose glucocorticoid is neuroprotective. Recently, it has been demonstrated that glucocorticoid provides a neuroprotective effect through the activation of neurotrophin signaling (32), indicating that the beneficial effect of low-dose dexamethasone in the restoration of the EPO-producing ability might be exerted through neurotrophin functions as well.

On the other hand, HGF has been reported to promote the regeneration of damaged cells after kidney injury (33). The regenerative capacity of HGF may be effective in the restoration of EPO-producing ability in myofibroblasts.

Recently, Bernhardt et al. demonstrated that the pharmacological manipulation of the HIF system increases EPO production in patients with end-stage renal disease. They suggested that the disturbed oxygen sensing of EPO-producing cells causes renal anemia (34). However, the activation of HIF has multiple downstream effects, and intermittent HIF activation over a prolonged period of time may lead to profound changes in cellular metabolism, growth, and differentiation (35).

The results of the present study prompt the proposal of another possible therapeutic approach to treat renal anemia by protecting damaged EPO-producing cells. The administration of neurotrophins, low-dose dexamethasone, and HGF as well as the reagents activating these pathways may be a promising therapeutic approach either alone or in combination with other approaches. In addition, SERMs, including tamoxifen, are promising candidate drugs to treat renal anemia as well as kidney fibrosis. Recently, raloxifene, another SERM, has been reported to be renoprotective in postmenopausal women with osteoporosis (36). Clinical trials of SERMs in patients with CKD with renal anemia and further clarification of the mechanism of action of SERMs are needed.

Although we demonstrated that almost all cortical fibroblasts in the kidney arise from P0-Cre–expressing precursors (Figure 1J), less than 20% of cortical fibroblasts produce EPO, even at the maximum induction of EPO under an anemic or hypoxic condition (27). Whether or not a certain subpopulation of P0-Cre lineage-labeled fibroblasts is designated to possess EPO-producing ability, or all P0-Cre lineage-labeled fibroblasts are born with the ability but some environmental factors determine which cells produce EPO, remains to be clarified. Although the lack of specific cell surface markers of EPO-producing cells has hindered further analysis, we demonstrated that p75 is an effective surface marker in the purification of EPO-producing cells in neonatal kidney (Figure 3H). In the adult kidney, however, p75 was not an effective marker because the expression of p75 protein in the kidney was downregulated (data not shown), whereas its mRNA was still detected by reverse-transcription PCR (RT-PCR) in EPO-producing cells (Figure 3G). A genome-wide analysis of EPO-producing cells purified from neonatal kidneys may enable the identification of suitable surface markers for the purification of the cells in adult kidneys.

We also demonstrated that more than 94% of myofibroblasts arise from P0-Cre lineage-labeled fibroblasts, providing what we believe to be new insights into the long-lasting controversy regarding the source of scar-producing myofibroblasts in the kidney. Recently, EMT has become widely accepted as a mechanism by which injured tubular cells transform into scar-producing cells and contribute to fibrosis during kidney injury (11). However, a number of conflicting data about the existence of EMT in vivo have been published (8, 37–39). Our data support the idea that myofibroblasts emerge from the proliferation and transdifferentiation of fibroblasts in the kidney (8), but it is not likely that they emerge from EMT of injured tubules. Recently, Humphreys et al. proposed that FoxD1⁺ mesenchymal cells in the developing kidneys differentiate into myofibroblasts in fibrosis (37), although the labeling efficiency of myofibroblasts by FoxD1-inducible Cre transgenic mice was 20%, due to the toxicity of tamoxifen in embryogenesis. P0-Cre fate-mapped cells populate around Six2⁺ progenitor cells in the developing kidney (Figure 2, G and H), and some of them express FoxD1 (Supplemental Figure 9). From this result, we speculate that the P0-Cre–labeled population and the FoxD1-Cre–labeled population are overlapping.

Recent reports indicate that neural crest cells contribute to bone marrow cells (22) and mesenchymal stem cells (40), in addition to the classical neural crest derivatives, such as peripheral nervous system, melanocytes, and adrenal medulla (24). The previous chick-quail xenotransplantation experiment (41, 42) as well as the dye injection into the dorsal neural tube (43) indicate that neural crest–derived cells contribute to the developing kidneys (44). This observation is also supported by the results in this manuscript, showing that most fibroblasts in the kidney arose from P0-Cre–expressing precursors (Figure 1, H–J) and that neonatal kidney fibroblasts expressed neural crest marker p75 (Figure 1K). On the contrary, a previous lineage-tracing study using Wnt1-Cre mice (45), another tool to trace the fate of neural crest derivatives, does not reveal a pattern overlapping with that of P0-Cre mice in the kidney. One possible explanation for the discordant results might be a difference in the efficiency and timing of labeling neural crest–derived cells in these Cre drivers. For instance, in Wnt1-Cre mice, Cre is predominantly expressed in the dorsal neural tube from which neural crest cells delaminate, whereas in P0-Cre mice, Cre expression is sustained in migrating neural crest cells. Although there is a possibility that the P0-Cre driver is ectopically expressed in other precursor tissues, P0-Cre–driven lineage labeling has to occur outside of the embryonic kidney, since we did not detect Cre expression within the developing kidney (Supplemental Figures 3 and 8).

Although further study is needed before one can definitely conclude that P0-Cre lineage-labeled cells are of neural crest origin, the hypothesis that EPO-producing cells derive from neural crest cells might open a new avenue of research toward the in vitro generation of EPO-producing cells from patient-derived induced pluripotent stem (iPS) cells. This is promising because an efficient method for converting iPS cells to neural crest cells has already been established in a recent study (46).

View Supplemental data

We thank M. Yanagisawa, M. Nakafuku, R. Kageyama, T. Ohtsuka, Y. Kaziro, T. Nakamura, Y. Nabeshima, S. Nagata, K. Ohta, and A. Economides for their valuable comments and discussion. We also greatly appreciate F. Costantini for the R26ECFP mice, J. Miyazaki for the CAG-CAT^loxP/loxP-EGFP mice, T. Yamamoto for his kind help with in situ hybridization, T. Kurosaki and H. Kawamoto for the valuable help in the transfer of the R26tdRFP mice, and A.P. McMahon and J.W. Mugford for the FoxD1 antibodies and protocol for immunostaining. We sincerely thank C. Gerle for the critical reading of this manuscript. We are very grateful to A. Hosotani for her excellent technical assistance. This study was supported by the Funding Program for Next-Generation World-leading Researchers; Grants-in-Aid from the Ministry of Education, Culture, Science, Sports and Technology of Japan (Shin-gakujyutsu 21200080); a grant-in-aid for Research on Biological Markers for New Drug Development; Health and Labor Sciences Research Grants from the Ministry of Health, Labor and Welfare of Japan (08062855); grants from Astellas Foundation for Research on Metabolic Disorders; Novartis Foundation for the Promotion of Science; Senri Lifescience Foundation; Mochida Memorial Foundation; Takeda Science Foundation; and Japan Foundation for Applied Enzymology.

Address correspondence to: Motoko Yanagita, Career-Path Promotion Unit for Young Life Scientists, E-wing, Kyoto University Graduate School of Medicine, Yoshida-konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan. Phone: 81.75.753.9310; Fax: 81.75.753.9311; E-mail: motoy@kuhp.kyoto-u.ac.jp.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2011;121(10):3981–3990. doi:10.1172/JCI57301.

1. Fisher JW. A quest for erythropoietin over nine decades. Annu Rev Pharmacol Toxicol. 1998;38:1–20.
   View this article via: PubMed CrossRef Google Scholar
2. Erslev AJ, Besarab A. Erythropoietin in the pathogenesis and treatment of the anemia of chronic renal failure. Kidney Int. 1997;51(3):622–630.
   View this article via: PubMed CrossRef Google Scholar
3. Soni RK, Weisbord SD, Unruh ML. Health-related quality of life outcomes in chronic kidney disease. Curr Opin Nephrol Hypertens. 2010;19(2):153–159.
   View this article via: PubMed CrossRef Google Scholar
4. Noguchi CT, Wang L, Rogers HM, Teng R, Jia Y. Survival and proliferative roles of erythropoietin beyond the erythroid lineage. Expert Rev Mol Med. 2008;10:e36.
   View this article via: PubMed CrossRef Google Scholar
5. Bachmann S, Le Hir M, Eckardt KU. Co-localization of erythropoietin mRNA and ecto-5′-nucleotidase immunoreactivity in peritubular cells of rat renal cortex indicates that fibroblasts produce erythropoietin. J Histochem Cytochem. 1993;41(3):335–341.
   View this article via: PubMed CrossRef Google Scholar
6. Obara N, Suzuki N, Kim K, Nagasawa T, Imagawa S, Yamamoto M. Repression via the GATA box is essential for tissue-specific erythropoietin gene expression. Blood. 2008;111(10):5223–5232.
   View this article via: PubMed CrossRef Google Scholar
7. Bhowmick NA, Neilson EG, Moses HL. Stromal fibroblasts in cancer initiation and progression. Nature. 2004;432(7015):332–337.
   View this article via: PubMed CrossRef Google Scholar
8. Kriz W, Kaissling B, Le Hir M. Epithelial-mesenchymal transition (EMT) in kidney fibrosis: fact or fantasy? J Clin Invest. 2011;121(2):468–474.
   View this article via: JCI PubMed CrossRef Google Scholar
9. Kaissling B, Le Hir M. The renal cortical interstitium: morphological and functional aspects. Histochem Cell Biol. 2008;130(2):247–262.
   View this article via: PubMed CrossRef Google Scholar
10. Wada T, Sakai N, Matsushima K, Kaneko S. Fibrocytes: a new insight into kidney fibrosis. Kidney Int. 2007;72(3):269–273.
    View this article via: PubMed CrossRef Google Scholar
11. Iwano M, Plieth D, Danoff TM, Xue C, Okada H, Neilson EG. Evidence that fibroblasts derive from epithelium during tissue fibrosis. J Clin Invest. 2002;110(3):341–350.
    View this article via: JCI PubMed Google Scholar
12. Duffield JS, Bonventre JV. Kidney tubular epithelium is restored without replacement with bone marrow-derived cells during repair after ischemic injury. Kidney Int. 2005;68(5):1956–1961.
    View this article via: PubMed CrossRef Google Scholar
13. Lin F, Moran A, Igarashi P. Intrarenal cells, not bone marrow-derived cells, are the major source for regeneration in postischemic kidney. J Clin Invest. 2005;115(7):1756–1764.
    View this article via: JCI PubMed CrossRef Google Scholar
14. Roufosse C, et al. Bone marrow-derived cells do not contribute significantly to collagen I synthesis in a murine model of renal fibrosis. J Am Soc Nephrol. 2006;17(3):775–782.
    View this article via: PubMed CrossRef Google Scholar
15. Lin SL, Kisseleva T, Brenner DA, Duffield JS. Pericytes and perivascular fibroblasts are the primary source of collagen-producing cells in obstructive fibrosis of the kidney. Am J Pathol. 2008;173(6):1617–1627.
    View this article via: PubMed Google Scholar
16. Humphreys BD, et al. Intrinsic epithelial cells repair the kidney after injury. Cell Stem Cell. 2008;2(3):284–291.
    View this article via: PubMed CrossRef Google Scholar
17. Yamauchi Y, et al. A novel transgenic technique that allows specific marking of the neural crest cell lineage in mice. Dev Biol. 1999;212(1):191–203.
    View this article via: PubMed CrossRef Google Scholar
18. Kawamoto S, et al. A novel reporter mouse strain that expresses enhanced green fluorescent protein upon Cre-mediated recombination. FEBS Lett. 2000;470(3):263–268.
    View this article via: PubMed CrossRef Google Scholar
19. Soriano P. Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat Genet. 1999;21(1):70–71.
    View this article via: PubMed CrossRef Google Scholar
20. Srinivas S, et al. Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev Biol. 2001;1:4.
    View this article via: PubMed CrossRef Google Scholar
21. Luche H, Weber O, Nageswara Rao T, Blum C, Fehling HJ. Faithful activation of an extra-bright red fluorescent protein in “knock-in” Cre-reporter mice ideally suited for lineage tracing studies. Eur J Immunol. 2007;37(1):43–53.
    View this article via: PubMed CrossRef Google Scholar
22. Nagoshi N, et al. Ontogeny and multipotency of neural crest-derived stem cells in mouse bone marrow, dorsal root ganglia, and whisker pad. Cell Stem Cell. 2008;2(4):392–403.
    View this article via: PubMed CrossRef Google Scholar
23. Andrae J, Gallini R, Betsholtz C. Role of platelet-derived growth factors in physiology and medicine. Genes Dev. 2008;22(10):1276–1312.
    View this article via: PubMed CrossRef Google Scholar
24. Nagoshi N, Shibata S, Nakamura M, Matsuzaki Y, Toyama Y, Okano H. Neural crest-derived stem cells display a wide variety of characteristics. J Cell Biochem. 2009;107(6):1046–1052.
    View this article via: PubMed CrossRef Google Scholar
25. Kobayashi A, et al. Six2 defines and regulates a multipotent self-renewing nephron progenitor population throughout mammalian kidney development. Cell Stem Cell. 2008;3(2):169–181.
    View this article via: PubMed CrossRef Google Scholar
26. Grande MT, Lopez-Novoa JM. Fibroblast activation and myofibroblast generation in obstructive nephropathy. Nat Rev Nephrol. 2009;5(6):319–328.
    View this article via: CrossRef PubMed Google Scholar
27. Nangaku M, Eckardt KU. Pathogenesis of renal anemia. Semin Nephrol. 2006;26(4):261–268.
    View this article via: PubMed CrossRef Google Scholar
28. Chao MV. Neurotrophins and their receptors: a convergence point for many signalling pathways. Nat Rev Neurosci. 2003;4(4):299–309.
    View this article via: PubMed CrossRef Google Scholar
29. Patapoutian A, Reichardt LF. Trk receptors: mediators of neurotrophin action. Curr Opin Neurobiol. 2001;11(3):272–280.
    View this article via: PubMed CrossRef Google Scholar
30. Webster NJ, Pirrung MC. Small molecule activators of the Trk receptors for neuroprotection. BMC Neurosci. 2008;9 suppl 2:S1.
    View this article via: PubMed Google Scholar
31. Karavanov A, Sainio K, Palgi J, Saarma M, Saxen L, Sariola H. Neurotrophin 3 rescues neuronal precursors from apoptosis and promotes neuronal differentiation in the embryonic metanephric kidney. Proc Natl Acad Sci U S A. 1995;92(24):11279–11283.
    View this article via: PubMed CrossRef Google Scholar
32. Jeanneteau F, Garabedian MJ, Chao MV. Activation of Trk neurotrophin receptors by glucocorticoids provides a neuroprotective effect. Proc Natl Acad Sci U S A. 2008;105(12):4862–4867.
    View this article via: PubMed CrossRef Google Scholar
33. Liu Y. Hepatocyte growth factor in kidney fibrosis: therapeutic potential and mechanisms of action. Am J Physiol Renal Physiol. 2004;287(1):F7–F16.
    View this article via: PubMed CrossRef Google Scholar
34. Bernhardt WM, et al. Inhibition of prolyl hydroxylases increases erythropoietin production in ESRD. J Am Soc Nephrol. 2010;21(12):2151–2156.
    View this article via: PubMed CrossRef Google Scholar
35. Haase VH. Hypoxic regulation of erythropoiesis and iron metabolism. Am J Physiol Renal Physiol. 2010;299(1):F1–F13.
    View this article via: PubMed CrossRef Google Scholar
36. Melamed ML, et al. Raloxifene, a selective estrogen receptor modulator, is renoprotective: a post-hoc analysis. Kidney Int. 2011;79(2):241–249.
    View this article via: PubMed CrossRef Google Scholar
37. Humphreys BD, et al. Fate tracing reveals the pericyte and not epithelial origin of myofibroblasts in kidney fibrosis. Am J Pathol. 2010;176(1):85–97.
    View this article via: PubMed CrossRef Google Scholar
38. Li L, Zepeda-Orozco D, Black R, Lin F. Autophagy is a component of epithelial cell fate in obstructive uropathy. Am J Pathol. 2010;176(4):1767–1778.
    View this article via: PubMed CrossRef Google Scholar
39. Koesters R, et al. Tubular overexpression of transforming growth factor-beta1 induces autophagy and fibrosis but not mesenchymal transition of renal epithelial cells. Am J Pathol. 2010;177(2):632–643.
    View this article via: PubMed CrossRef Google Scholar
40. Morikawa S, et al. Development of mesenchymal stem cells partially originate from the neural crest. Biochem Biophys Res Commun. 2009;379(4):1114–1119.
    View this article via: PubMed CrossRef Google Scholar
41. Le Douarin NM, Teillet MA. Experimental analysis of the migration and differentiation of neuroblasts of the autonomic nervous system and of neurectodermal mesenchymal derivatives, using a biological cell marking technique. Dev Biol. 1974;41(1):162–184.
    View this article via: PubMed CrossRef Google Scholar
42. Le Douarin NM. The avian embryo as a model to study the development of the neural crest: a long and still ongoing story. Mech Dev. 2004;121(9):1089–1102.
    View this article via: PubMed CrossRef Google Scholar
43. Bronner-Fraser M, Fraser SE. Cell lineage analysis reveals multipotency of some avian neural crest cells. Nature. 1988;335(6186):161–164.
    View this article via: PubMed CrossRef Google Scholar
44. Sainio K, Nonclercq D, Saarma M, Palgi J, Saxen L, Sariola H. Neuronal characteristics in embryonic renal stroma. Int J Dev Biol. 1994;38(1):77–84.
    View this article via: PubMed Google Scholar
45. Itaranta P, Viiri K, Kaartinen V, Vainio S. Lumbo-sacral neural crest derivatives fate mapped with the aid of Wnt-1 promoter integrate but are not essential to kidney development. Differentiation. 2009;77(2):199–208.
    View this article via: PubMed CrossRef Google Scholar
46. Chambers SM, Fasano CA, Papapetrou EP, Tomishima M, Sadelain M, Studer L. Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nat Biotechnol. 2009;27(3):275–280.
    View this article via: PubMed CrossRef Google Scholar
47. Menon MP, Karur V, Bogacheva O, Bogachev O, Cuetara B, Wojchowski DM. Signals for stress erythropoiesis are integrated via an erythropoietin receptor-phosphotyrosine-343-Stat5 axis. J Clin Invest. 2006;116(3):683–694.
    View this article via: JCI PubMed CrossRef Google Scholar
48. Yanagita M, et al. Uterine sensitization-associated gene-1 (USAG-1), a novel BMP antagonist expressed in the kidney, accelerates tubular injury. J Clin Invest. 2006;116(1):70–79.
    View this article via: JCI PubMed CrossRef Google Scholar
49. Tanaka M, et al. Expression of BMP-7 and USAG-1 (a BMP antagonist) in kidney development and injury. Kidney Int. 2008;73(2):181–191.
    View this article via: PubMed CrossRef Google Scholar
50. Tanaka M, et al. Loss of the BMP antagonist USAG-1 ameliorates disease in a mouse model of the progressive hereditary kidney disease Alport syndrome. J Clin Invest. 2010;120(3):768–777.
    View this article via: JCI PubMed CrossRef Google Scholar
51. Mugford JW, Sipila P, McMahon JA, McMahon AP. Osr1 expression demarcates a multi-potent population of intermediate mesoderm that undergoes progressive restriction to an Osr1-dependent nephron progenitor compartment within the mammalian kidney. Dev Biol. 2008;324(1):88–98.
    View this article via: PubMed CrossRef Google Scholar
52. Vielhauer V, Anders HJ, Perez de Lema G, Luckow B, Schlondorff D, Mack M. Phenotyping renal leukocyte subsets by four-color flow cytometry: characterization of chemokine receptor expression. Nephron Exp Nephrol. 2003;93(2):e63.
    View this article via: PubMed CrossRef Google Scholar