Dickkopf-3 in aberrant endothelial secretome triggers renal fibroblast activation and endothelial–mesenchymal transition

Dickkopf-3 in aberrant endothelial secretome triggers renal fibroblast activation and... Abstract Background Our laboratory has previously demonstrated that Sirt1endo−/− mice show endothelial dysfunction and exaggerated renal fibrosis, whereas mice with silenced endothelial transforming growth factor beta (TGF-β) signaling are resistant to fibrogenic signals. Considering the fact that the only difference between these mutant mice is confined to the vascular endothelium, this indicates that secreted substances contribute to these contrasting responses. Methods We performed an unbiased proteomic analysis of the secretome of renal microvascular endothelial cells (RMVECs) isolated from these two mutants. We cultured renal fibroblasts and RMVECs and used microfluidic devices for coculturing. Results Dickkopf-3 (DKK3), a putative ligand of the Wnt/β-catenin pathway, was present exclusively in the fibrogenic secretome. In cultured fibroblasts, DKK3 potently induced myofibroblast activation. In addition, DKK3 antagonized effects of DKK1, a known inhibitor of the Wnt pathway, in conversion of fibroblasts to myofibroblasts. In RMVECs, DKK3 induced endothelial–mesenchymal transition and impaired their angiogenic competence. The inhibition of endothelial outgrowth, enhanced myofibroblast formation and endothelial–mesenchymal transition were confirmed in coculture. In reporter DKK3-eGFP × Col3.6-GFPcyan mice, DKK3 was marginally expressed under basal conditions. Adriamycin-induced nephropathy resulted in upregulation of DKK3 expression in tubular and, to a lesser degree, endothelial compartments. Sulindac sulfide was found to exhibit superior Wnt pathway-suppressive action and decreased DKK3 signals and the extent of renal fibrosis. Conclusions In conclusion, this unbiased proteomic screen of the profibrogenic endothelial secretome revealed DKK3 acting as an agonist of the Wnt pathway, enhancing formation of myofibroblasts and endothelial–mesenchymal transition and impairing angiogenesis. A potent inhibitor of the Wnt pathway, sulindac sulfide, suppressed nephropathy-induced DKK3 expression and renal fibrosis. endothelial–fibroblast cocultures, endothelial secretome, microfluidic device, renal fibrosis, Wnt pathway INTRODUCTION It has been recognized that endothelial cells normally produce diverse paracrine-tropic and angiocrine factors necessary for differentiation and regeneration of tissues as dissimilar as pancreatic acini, neurons, hematopoietic precursors, hepatocytes and alveolar epithelia [1]. For instance, senescent endothelial cells [2, 3] or endothelial cells undergoing mesenchymal transition [4] not only disrupt the function of the endothelial lining of the vessels but also desert their instructive angiocrine functions and affect the neighboring cells by their secretome. The term ‘secretome’ comprises all proteins secreted from the plasma membrane and released with exosomes and other cell organellar products harvested from the extracellular compartment [5]. The recent analysis of the secretome of cultured human umbilical vein endothelial cells disclosed 123 secreted proteins [6]. We have recently generated two strains of mutant mice with endothelial cells either lacking exon 4 encoding for the deacetylase activity of sirtuin 1 (Sirt1endo−/−) or deficient in transforming growth factor beta (TGF-β) receptor II (TGFβRIIendo+/−); the former mutant is prone to develop fibrosis and the latter is protected from fibrosis [7, 8]. Since endothelial Sirt1 deficiency is a consistent companion of a range of chronic cardiovascular, metabolic and renal diseases [9–11], clearly the Sirt1endo−/− mouse model has broad applicability. On the other hand, silencing of TGF-β signaling, as in our TGFβRIIendo+/− model, is one of the goals of pharmaceutical research for preventing fibrosis [8, 12]. Sirt1-deficient endothelial cells have all the signs of dysfunction: reduced endothelium-dependent vasorelaxation, microvascular rarefaction, impaired migration, sprouting, matrilytic activity, loss of glycocalyx, premature senescence and exaggerated fibrotic response [7–11]. In contrast, the TGF-βRIIendo+/− model has no distinctive phenotype under basal conditions but develops much reduced fibrosis after induction of unilateral ureteral obstruction or the chronic phase of folic acid nephropathy [8]. These contrasting strains not only emphasize the fact that messages from the endothelium are instructive for fibroblasts but also represent a useful tool to juxtapose endothelial secretomes as the only differentially expressed variables potentially explaining the distinct end effects. Hence one of the goals of the present study was to perform mining of the profibrogenic endothelial secretome. The detection of Dickkopf-3 (DKK3) uniquely present in this aberrant secretome prompted the undertaking of studies in cultured renal microvascular endothelial cells (RMVECs) and fibroblasts, in their cocultures and in whole animals to elucidate the functional significance of this finding. MATERIALS AND METHODS Mouse models used in the study All animal experiments were performed in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and was approved by the Institutional Animal Care and Use Committee. Sirt1endo−/− mice. The endothelial Sirt1-deleted mouse model was established by cross-breeding B6 129-Sirt1tm1Ygu/J, which harbors a deletion of exon 4 encoding for the deacetylase catalytic domain, with Tie2-Cre transgenic mice expressing cre-recombinase in vascular endothelial cells (both from Jackson Laboratory, Bar Harbor, ME, USA) [7]. The resulting Sirt1endo+/− mice were mated with Sirt1Flox/Flox mice to produce endothelial-deleted Sirt1endo−/− mutant mice. TGFβRIIendo+/− mice. These mice were generated by Xavier et al. [8] as detailed previously. DKK3-eGFP × Col3.6-GFPcyan mice. These mice on a CD1 background were generated by Dr. David W. Rowe [13, 14]. Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice. These were established by cross-breeding Sirt−/− mice with DKK3-eGFP × Col3.6-GFPcyan mice. α-Smooth muscle actin–green fluorescent protein mice. α-Smooth muscle actin (SMA)–green fluorescent protein (GFP) transgenic mice on a C57BL/6 J background were generously supplied by Dr. Ivo Kalajzic (University of Connecticut Health Center, Farmington, CT, USA). Mice were originally developed by Dr. Jen-Yue Tsai (National Eye Institute, NIH) and carried a regulatory sequence of α-SMA gene spanning 1074 bp of the 5′-flanking region, the transcription start site, 48 bp of exon 1, the 2.5-kbp intron 1 and the 15-bp exon 2 of mouse α-SMA. GFP is specifically expressed in both vascular and nonvascular smooth muscle cells [15, 16]. Mice were housed in the animal care facility of New York Medical College (25°C, 50% humidity and 12-h dark/light cycle) with free access to food and water. Microfluidic device studies of spatially separated RMVEC fibroblast cocultures We performed all microfluidic device experiments in accordance with Kim et al.’s protocol [17]. Briefly, a fibrinogen solution was prepared by dissolving 2.5 mg/mL bovine fibrinogen in phosphate-buffered saline (Corning, Corning, NY, USA). RMVECs, preloaded with a CellTracker-red (Fisher Scientific, Pittsburgh, PA, USA), and renal fibroblasts, all obtained from α-SMA-GFP mice, were individually suspended in the fibrinogen solution. The cell solutions were mixed with thrombin (0.5 U/mL) and then promptly pipetted into the central and peripheral channels leaving the spaces between endothelial cells and fibroblasts for the respective ingrowth. After clotting for 3–5 min at room temperature, the inlet reservoirs of the cell culture medium channels were loaded with endothelial cell culture medium and then a vacuum was gently applied at the outlet reservoirs to fill the hydrophobic channels. The microfluidic devices were incubated at 37°C in a humidified 5% carbon dioxide atmosphere. The cell culture medium was refilled with fresh endothelial cell culture medium every 48 h. Statistical analysis Values are given as the mean ± standard error of the mean unless stated otherwise. Data were analyzed using independent t-test or analysis of variance (ANOVA) with post hoc analysis for multiple group comparisons using the Bonferroni method. A P-value <0.05 was considered statistically significant. All statistical analyses were performed with NCSS 10 (NCSS, Kaysville, UT, USA). RESULTS Mass spectrometry of the RMVEC secretome CD31 magnetically isolated high-purity populations of RMVEC from control wild-type, Sirt1endo−/− or TGF-βRIIendo+/− mice were expanded and exposed to a vehicle or TGF-β and conditioned medium was collected 48 h later. The latter was analyzed using unbiased, nontargeted tandem mass spectrometry as detailed in the Supplementary Methods section. The tryptic digests were analyzed using the Thermo Scientific Q Exactive Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) and the tandem mass spectra were extracted. Using this technique we were able to detect 332 nonredundant proteins that belong to diverse categories, as reported previously [18, 19]. We argue that the differential signatures of proteins secreted exclusively by Sirt1endo−/− versus TGF-βRIIendo+/− could theoretically contain protein culprits for pro- and antifibrogenic phenotypes of the respective mice. Among those differentially expressed secreted proteins, DKK3 was present exclusively in the fibrogenic secretome (Figure 1). Whether DKK3 secreted by dysfunctional endothelial cells may interfere with DKK1 in regulating Wnt signaling in surrounding cells is entirely unknown. For these reasons, we decided to explore the effects of DKK3 on the Wnt pathway in renal fibroblasts and RMVECs. FIGURE 1: View largeDownload slide DKK3 is elevated in the secretome of Sirt1endo−/− mice. RMVEC from control wild-type and Sirt1endo−/− mice were expanded, exposed to TGF-β (5 ng/mL) and conditioned media were collected and subjected to mass spectrometry analyses. Among the differentially regulated proteins, DKK3 was highly enriched in the secretome of RMVECs obtained from Sirt1endo−/− mice upon stimulation with TGF-β (***P < 0.05). FIGURE 1: View largeDownload slide DKK3 is elevated in the secretome of Sirt1endo−/− mice. RMVEC from control wild-type and Sirt1endo−/− mice were expanded, exposed to TGF-β (5 ng/mL) and conditioned media were collected and subjected to mass spectrometry analyses. Among the differentially regulated proteins, DKK3 was highly enriched in the secretome of RMVECs obtained from Sirt1endo−/− mice upon stimulation with TGF-β (***P < 0.05). Paracrine effects of DKK3 in renal fibroblasts Fibroblasts were isolated from α-SMA-GFP mouse kidneys using positive (enrichment of PDGFR+ cells) and negative (depletion of CD31+cells) selection with magnetic beads according to the previously established protocol as detailed in the Methods section. Under resting conditions, very few fibroblasts expressed α-SMA-GFP. Application of DKK3 was combined with the known agonist of the Wnt pathway, Wnt1, or a known antagonist of the Wnt pathway, DKK1, each of these components with or without the application of TGF-β to fibroblasts in culture and monitoring at the time of appearance of GFP fluorescence. We expected that this protocol might reveal modulatory effects of DKK3 on the time and extent of fibroblast-to-myofibroblast conversion. Data indicated that DKK3 (10 µg/mL) alone induced a myofibroblastic phenotype (Figure 2); however, it did not further alter TGF-β-induced responses. Application of Wnt1 (100 ng/mL) alone similarly induced GFP expression in α-SMA-GFP renal fibroblasts, whereas DKK1 coapplication reduced fibroblast activation. When DKK3 was applied in combination with Wnt1 and DKK1, it resulted in the antagonistic signal from DKK1 being weakly overridden by DKK3. The presence of TGF-β has not perceptibly altered the directionality of these responses but induced a strong counter-regulation by DKK3 of DKK1-induced inhibition of myofibroblasts formation. Hence data generated in renal fibroblasts is consistent with the notion that DKK3 acts as an agonist of the Wnt pathway and is capable of counteracting antagonistic effects of DKK1, especially in the presence of TGF-β. FIGURE 2: View largeDownload slide DKK3 induces fibroblast-to-myofibroblast transition. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured renal fibroblasts isolated from α-SMA-GFP reporter mice 24 h after exposure to the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Data indicate that DKK3 (10 µg/mL) alone induced the myofibroblastic phenotype; however, it did not further enhance responses to TGF-β. *, indicates the control group to be statistically significantly different from all other groups; #, indicates the Wnt/DKK3 group to be statistically significantly different from all other groups except the Wnt group (P < 0.05). (C) Quantification of α-SMA-GFP fluorescence intensity. DKK3 did not further enhance responses to TGF-β. *, indicates the control group to be statistically significantly different from all other groups; #, indicates the TGFβ/Wnt/DKK1 group to be statistical significantly different from all other groups as well (P < 0.05). FIGURE 2: View largeDownload slide DKK3 induces fibroblast-to-myofibroblast transition. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured renal fibroblasts isolated from α-SMA-GFP reporter mice 24 h after exposure to the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Data indicate that DKK3 (10 µg/mL) alone induced the myofibroblastic phenotype; however, it did not further enhance responses to TGF-β. *, indicates the control group to be statistically significantly different from all other groups; #, indicates the Wnt/DKK3 group to be statistically significantly different from all other groups except the Wnt group (P < 0.05). (C) Quantification of α-SMA-GFP fluorescence intensity. DKK3 did not further enhance responses to TGF-β. *, indicates the control group to be statistically significantly different from all other groups; #, indicates the TGFβ/Wnt/DKK1 group to be statistical significantly different from all other groups as well (P < 0.05). Next, we examined in depth (Figure 3) the dose–response relations between DKK1 and DKK3. Data showed a complex relationship: DKK3 is able to convert fibroblasts to myofibroblasts only at low levels of DKK1. The sharp increase in sensitivity of fibroblasts to DKK3 at the lower spectrum of DKK1 concentrations would argue that at distant sites from DKK3-secreting dysfunctional endothelial cells, DKK1 effects predominate and Wnt signaling is partially counteracted. At the sites nearest to the dysfunctional endothelial cells, such as pericytes, the gradients of DKK3 and DKK1 favor the former and therefore these cells could be most influenced by the endothelium-secreted DKK3. On the other hand, based on studies from Federico et al. [20], interstitial fibroblasts may be more influenced by the epithelium-secreted DKK3. In addition, Wnt secretion by the damaged epithelium has been shown by Zhou et al. [21], thus placing interstitial fibroblasts under the cooperative influence of both DKK3 and Wnt to convert them to myofibroblasts. This intricate gradient field of profibrogenic messengers focusing on pericytes and interstitial fibroblasts, still insufficiently investigated, may play a substantial role in conversion of these cell types into myofibroblasts. FIGURE 3: View largeDownload slide Dose response of DKK1 and DKK3 on renal fibroblasts. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured renal fibroblasts isolated from α-SMA-GFP reporter mice 48 h after exposure to the indicated proteins and an additional 48 h after exposure to TGF-β(n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Data indicate that DKK3 in a dose-responsive manner is able to convert fibroblasts to myofibroblasts only at the low levels of DKK1. (C) Quantification of α-SMA-GFP fluorescence intensity. TGF-β did not further enhance responses to DKK3. FIGURE 3: View largeDownload slide Dose response of DKK1 and DKK3 on renal fibroblasts. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured renal fibroblasts isolated from α-SMA-GFP reporter mice 48 h after exposure to the indicated proteins and an additional 48 h after exposure to TGF-β(n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Data indicate that DKK3 in a dose-responsive manner is able to convert fibroblasts to myofibroblasts only at the low levels of DKK1. (C) Quantification of α-SMA-GFP fluorescence intensity. TGF-β did not further enhance responses to DKK3. Autocrine effects of the secretome and how does DKK3 affect the angiogenesis of endothelial cells Since DKK3 released to the extracellular milieu acquires the functions of a cytokine [22], it is plausible that it may exert not only paracrine effects on mesenchymal cells but also autocrine effects. DKK3 autocrine actions as possible angiogenic modulators were studied using RMVECs obtained from α-SMA-GFP mouse kidneys, again employed as a reporter system. Application of DKK3 was performed in the presence or absence of Wnt1 (to induce the ‘on’ state of frizzled receptors), DKK1 (to confirm its well-described inhibition of Wnt actions) and TGF-β. Results demonstrated (Figure 4) that 48 h after these treatments, control cells gradually underwent endothelial–mesenchymal transition (as judged by the acquisition of α-SMA-GFP signal). Moreover, Wnt1 and DKK3 further promoted endothelial–mesenchymal transition, but their effects were nonadditive. As expected, the addition of DKK1 to Wnt1 resulted in a significant reduction of α-SMA-GFP fluorescence. Remarkably, the addition of DKK3 to the combined treatment with Wnt1 and DKK1 reduced the inhibitory effect of DKK1 on endothelial–mesenchymal transition. When the above-mentioned treatments were applied in the presence of TGF-β, acquisition of α-SMA-GFP fluorescence by RMVECs was also inhibited by DKK1. However, under those conditions, DKK3 afforded even higher antagonistic activity against DKK1. FIGURE 4: View largeDownload slide Autocrine effects of DKK3. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured RMVECs isolated from α-SMA-GFP reporter mice 48 h after exposure to the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Control cells gradually underwent endothelial–mesenchymal transition (as judged by the acquisition of α-SMA-GFP signal). Moreover, Wnt1 and DKK3 further promoted endothelial–mesenchymal transition, but their effects were nonadditive. As expected, the addition of DKK1 to Wnt1 resulted in a significant reduction in α-SMA-GFP fluorescence. *, labeled groups are statistically significantly different from the nonlabeled groups (P < 0.05). (C) Quantification of α-SMA-GFP fluorescence intensity. *, indicates the control group and TGFβ/Wnt/DKK1/DKK3 group to be significantly different from all other groups; #, indicates the TGFβ/Wnt/DKK1 group to be significantly different from all other groups (P < 0.05). FIGURE 4: View largeDownload slide Autocrine effects of DKK3. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured RMVECs isolated from α-SMA-GFP reporter mice 48 h after exposure to the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Control cells gradually underwent endothelial–mesenchymal transition (as judged by the acquisition of α-SMA-GFP signal). Moreover, Wnt1 and DKK3 further promoted endothelial–mesenchymal transition, but their effects were nonadditive. As expected, the addition of DKK1 to Wnt1 resulted in a significant reduction in α-SMA-GFP fluorescence. *, labeled groups are statistically significantly different from the nonlabeled groups (P < 0.05). (C) Quantification of α-SMA-GFP fluorescence intensity. *, indicates the control group and TGFβ/Wnt/DKK1/DKK3 group to be significantly different from all other groups; #, indicates the TGFβ/Wnt/DKK1 group to be significantly different from all other groups (P < 0.05). Having demonstrated the capability of DKK3 to induce endothelial–mesenchymal transition, we explored whether DKK3 influences the angiogenesis of RMVECs. Therefore we plated RMVECs on matrigel, exposed them to DKK3 and examined the ability to form a capillary network. We quantified angiogenic parameters using NIH-developed Angiotool software. The total length, percentage area occupied by capillary-like structures and the number of bifurcations were all significantly reduced by DKK3 (Figure 5). FIGURE 5: View largeDownload slide Angiogenesis assay of cultured RMVECs on matrigel. (A) Representative images of cultured RMVECs on matrigel with and without DKK3 (n = 5; scale bar 150 µm). (B–D) Quantification of the vessels percentage area, total number of junctions and total vessels length. RMVECs treated with DKK3 experienced a profound decrease in vessels percentage area, total number of branches and total vessels length after 96 h of treatment (*P < 0.05). FIGURE 5: View largeDownload slide Angiogenesis assay of cultured RMVECs on matrigel. (A) Representative images of cultured RMVECs on matrigel with and without DKK3 (n = 5; scale bar 150 µm). (B–D) Quantification of the vessels percentage area, total number of junctions and total vessels length. RMVECs treated with DKK3 experienced a profound decrease in vessels percentage area, total number of branches and total vessels length after 96 h of treatment (*P < 0.05). Microfluidic device cocultures of renal endothelial and fibroblastic cells Since the above-mentioned two-dimensional (2D) cultures of endothelial cells and fibroblasts lack the complexity of their biological three-dimensional topography and interactions, we established a coculture system using microfluidic devices according to the previously published protocols [17]. The microfluidic devices were fabricated using Sylgard 184 (Dow Corning, Corning, NY, USA) by lithography and replica modeling as described previously [17] and illustrated in Figure 6A. RMVECs and renal fibroblasts were isolated from kidneys of α-SMA-GFP mice. The RMVECs were prelabeled with red CellTracker and introduced into nanochannels in the fibrin gel (fibrinogen and thrombin) alongside fibroblasts. DKK3 was added to the endothelial culture medium and the outgrowth of RMVECs toward fibroblasts as well as the proportion of RMVECs undergoing endothelial–mesenchymal transition and the proportion of fibroblasts becoming α-SMA-GFP positive were monitored by intravital microscopy and quantified. Analysis of angiogenic parameters showed a reduction in the total length, percentage area occupied by capillary-like structures and number of bifurcations with the addition of DKK3 (Figure 6B–I). Furthermore, we documented that with the addition of DKK3 to the culture medium, the extent of endothelial–mesenchymal transition and fibroblast–myofibroblast transition was significantly increased (Figure 7). FIGURE 6: View largeDownload slide Angiogenesis assay of microfluidic cocultures of renal endothelial and fibroblastic cells. (A) Illustration of a microfluidic device with five nanochannels (one central, two outer channels and two intersectional channels). (B) Representative bright-field images of cocultured RMVECs and renal fibroblasts with and without DKK3 at Day 4 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. Under control conditions, RMVECs began forming a capillary network inside the central channel with early outgrowth toward the intersectional channel, whereas the addition of DKK3 decreased the vessels percentage, the total number of branches and the total vessels length (n = 6; scale bar 150 µm). (C–E) Quantification of the vessels percentage area, total number of junctions and total vessels length on Day 4 (*P < 0.05). (F) Representative bright-field images of cocultured RMVECs and renal fibroblasts with and without DKK3 on Day 9 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. By Day 9, RMVECs formed a capillary network in the intersectional channel under control conditions. A capillary network or branching could not be recognized with the addition of DKK3 (n = 6; scale bar 150 µm). (G–I) Quantification of the vessels percentage area, total number of junctions and total vessels length on Day 9 (*P < 0.05). FIGURE 6: View largeDownload slide Angiogenesis assay of microfluidic cocultures of renal endothelial and fibroblastic cells. (A) Illustration of a microfluidic device with five nanochannels (one central, two outer channels and two intersectional channels). (B) Representative bright-field images of cocultured RMVECs and renal fibroblasts with and without DKK3 at Day 4 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. Under control conditions, RMVECs began forming a capillary network inside the central channel with early outgrowth toward the intersectional channel, whereas the addition of DKK3 decreased the vessels percentage, the total number of branches and the total vessels length (n = 6; scale bar 150 µm). (C–E) Quantification of the vessels percentage area, total number of junctions and total vessels length on Day 4 (*P < 0.05). (F) Representative bright-field images of cocultured RMVECs and renal fibroblasts with and without DKK3 on Day 9 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. By Day 9, RMVECs formed a capillary network in the intersectional channel under control conditions. A capillary network or branching could not be recognized with the addition of DKK3 (n = 6; scale bar 150 µm). (G–I) Quantification of the vessels percentage area, total number of junctions and total vessels length on Day 9 (*P < 0.05). FIGURE 7: View largeDownload slide DKK3 induced endothelial–mesenchymal and fibroblast-to-myofibroblast transition in microfluidic device cocultures of renal endothelial and fibroblastic cells. (A–C) Representative images of cocultured RMVECs and renal fibroblasts with and without DKK3 on Days 1, 3 and 8 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. RMVECs prelabeled with a CellTracker are shown in red. Green = α-SMA-GFP (n = 6; scale bar 150 µm). (D–F) Quantification of the α-SMA-GFP fluorescence intensity on Days 1, 3, and 8 (*P < 0.05). The addition of DKK3 to the culture medium caused more intense endothelial–mesenchymal and fibroblast-to-myofibroblast transition. FIGURE 7: View largeDownload slide DKK3 induced endothelial–mesenchymal and fibroblast-to-myofibroblast transition in microfluidic device cocultures of renal endothelial and fibroblastic cells. (A–C) Representative images of cocultured RMVECs and renal fibroblasts with and without DKK3 on Days 1, 3 and 8 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. RMVECs prelabeled with a CellTracker are shown in red. Green = α-SMA-GFP (n = 6; scale bar 150 µm). (D–F) Quantification of the α-SMA-GFP fluorescence intensity on Days 1, 3, and 8 (*P < 0.05). The addition of DKK3 to the culture medium caused more intense endothelial–mesenchymal and fibroblast-to-myofibroblast transition. In another approach, we cocultured RMVECs and renal fibroblasts for 12 days in endothelial cell culture medium. We then added DKK3 (10 µg/mL) to the medium and cultured for 2 additional days. As shown in Figure 8, the RMVECs formed a dense capillary network inside the central channel with bifurcations reaching out to the intersectional channel after 12 days. After the addition of DKK3, the capillary network collapsed as judged by measuring the total length, percentage area occupied by capillary-like structures and number of bifurcations. Moreover, endothelial–mesenchymal transition was enhanced after the addition of DKK3 as judged by the intensity of α-SMA-GFP fluorescence. Collectively these findings confirm our results of the 2D cultures establishing DKK3 as an inducer of endothelial–mesenchymal/fibroblast–myofibroblast transition and as a suppressor of angiogenesis. FIGURE 8: View largeDownload slide DKK3 disrupts capillary network formation. (A) Representative images of cocultured RMVECs before and after exposure to DKK3. The central channel harboring RMVECs is shown on the right side of the image. The intersectional channel is shown on the left side of the image. RMVECs prelabeled with a CellTracker are shown in red. Green = α-SMA-GFP. Black dots indicate beads conjugated with CD31 antibody. RMVECs were cocultured with renal fibroblasts for 12 days (left panel). RMVECs formed a dense capillary network with branches toward the intersectional channel. On Day 12 of culture, DKK3 (10 µg/mL) was added to the endothelial culture medium. Forty-eight hours later the capillary network was abolished and the RMVECs underwent mesenchymal transition, as judged by the appearance of α-SMA-GFP, represented as the yellow areas created by colocalization of the red CellTracker and the GFP signal (n = 3; scale bar 150 µm). (B–E) Quantification of the α-SMA-GFP fluorescence intensity, the vessels percentage area, total number of junctions and total vessels length intensity on Day 14 (*P < 0.05). FIGURE 8: View largeDownload slide DKK3 disrupts capillary network formation. (A) Representative images of cocultured RMVECs before and after exposure to DKK3. The central channel harboring RMVECs is shown on the right side of the image. The intersectional channel is shown on the left side of the image. RMVECs prelabeled with a CellTracker are shown in red. Green = α-SMA-GFP. Black dots indicate beads conjugated with CD31 antibody. RMVECs were cocultured with renal fibroblasts for 12 days (left panel). RMVECs formed a dense capillary network with branches toward the intersectional channel. On Day 12 of culture, DKK3 (10 µg/mL) was added to the endothelial culture medium. Forty-eight hours later the capillary network was abolished and the RMVECs underwent mesenchymal transition, as judged by the appearance of α-SMA-GFP, represented as the yellow areas created by colocalization of the red CellTracker and the GFP signal (n = 3; scale bar 150 µm). (B–E) Quantification of the α-SMA-GFP fluorescence intensity, the vessels percentage area, total number of junctions and total vessels length intensity on Day 14 (*P < 0.05). DKK3-eGFP mice with adriamycin-induced nephropathy and sulindac sulfide treatment To facilitate monitoring of DKK3 expression in the kidney, we made use of reporter DKK3-eGFP × Col3.6-GFPcyan mice. Under basal conditions, DKK3 expression, as judged by GFP fluorescence, was minimal. Induction of adriamycin-induced nephropathy in these mice resulted in robust GFP signals emanating from the tubular epithelial cells, parietal epithelial cells and, to a lesser degree, the vascular endothelium (Figure 9). These findings further expand the list of potential sources of DKK3 in pathologic situations and suggest that the tubular and glomerular epithelia may contribute to the DKK3 burden of the endothelial secretome. These findings are in concert with the previous demonstration that the tubular DKK3 promotes the development of renal tubular atrophy and fibrosis [20]. FIGURE 9: View largeDownload slide Sulindac sulfide attenuates renal fibrosis. (A) Representative images of DKK3-GFP fluorescence intensity and sirius red staining after adriamycin-induced nephropathy with and without sulindac sulfide treatment (n = 4; scale bar for ×10, ×150 µm and for ×40, ×30 µm). (B–C) Quantification of the DKK3-GFP fluorescence intensity and the tissue fibrosis percentage area. Mice treated with sulindac sulfide showed suppressed DKK3-GFP fluorescence and reduced interstitial fibrosis in the adriamycin-induced nephropathy model (*P < 0.05 versus vehicle). FIGURE 9: View largeDownload slide Sulindac sulfide attenuates renal fibrosis. (A) Representative images of DKK3-GFP fluorescence intensity and sirius red staining after adriamycin-induced nephropathy with and without sulindac sulfide treatment (n = 4; scale bar for ×10, ×150 µm and for ×40, ×30 µm). (B–C) Quantification of the DKK3-GFP fluorescence intensity and the tissue fibrosis percentage area. Mice treated with sulindac sulfide showed suppressed DKK3-GFP fluorescence and reduced interstitial fibrosis in the adriamycin-induced nephropathy model (*P < 0.05 versus vehicle). Having screened a library of compounds with the ability to suppress the Wnt pathway, Bravi et al. [23] demonstrated that sulindac sulfide is the most efficient inhibitor. For these reasons we chronically treated the DKK3-eGFP × Col3.6-GFPcyan reporter mice subjected to adriamycin-induced nephropathy with sulindac sulfide. After 7 days of treatment the intensity of fluorescent signals decreased, while the extent of renal fibrosis was significantly diminished (Figure 9). Additional in vitro studies with RMVECs isolated from DKK3-eGFP × Col3.6-GFPcyan reporter mice showed that the DKK3-GFP expression increased significantly after exposure to TGF-β, and especially in combination with Sirt1-inhibition (Figure 10). Sirt1 inhibition alone did not significantly increase the DKK3 expression, which is in accordance with the mass spectrometry results. Treatment with sulindac sulfide reduced DKK3 expression in all the groups, further emphasizing its ability to suppress the Wnt pathway. In addition, renal fibroblasts isolated from α-SMA-GFP mice showed reductions in TCF-1 transcripts, a Wnt downstream target gene, after treatment with sulindac sulfide (Figure 10C). FIGURE 10: View largeDownload slide Sulindac sulfide reduces DKK3 expression in RMVECs. (A) Representative images of DKK3-GFP fluorescence intensity in RMVECs isolated from DKK3-eGFP × Col3.6-GFPcyan mice 48 h after treatment with the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of the DKK3-GFP fluorescence intensity. DKK3 expression increased after TGF-β treatment and especially in combination with Sirt1 inhibitor in RMVECs. Sulindac sulfide effectively reduced DKK3 expression (*P < 0.05; #, indicates the control group and Sirt1 inhibitor group to be significantly different from the TGF-β group and Sirt1 inhibitor + TGF-β group). (C) Quantitative polymerase chain reaction of TCF-1 transcripts in renal fibroblasts isolated from α-SMA-GFP mice (n = 4; *P < 0.05). FIGURE 10: View largeDownload slide Sulindac sulfide reduces DKK3 expression in RMVECs. (A) Representative images of DKK3-GFP fluorescence intensity in RMVECs isolated from DKK3-eGFP × Col3.6-GFPcyan mice 48 h after treatment with the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of the DKK3-GFP fluorescence intensity. DKK3 expression increased after TGF-β treatment and especially in combination with Sirt1 inhibitor in RMVECs. Sulindac sulfide effectively reduced DKK3 expression (*P < 0.05; #, indicates the control group and Sirt1 inhibitor group to be significantly different from the TGF-β group and Sirt1 inhibitor + TGF-β group). (C) Quantitative polymerase chain reaction of TCF-1 transcripts in renal fibroblasts isolated from α-SMA-GFP mice (n = 4; *P < 0.05). In order to investigate DKK3-GFP fluorescence in a model of endothelial dysfunction, we created Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice. Endothelial Sirt1 deficiency has been shown to cause the development of microvascular rarefaction, fibrosis and a decline in the endothelial glycocalyx, to name a few, already at a basal state [7, 10, 19]. As demonstrated in Figure 11A, wild-type 10-week-old mice with intact endothelial Sirt1 showed no DKK3-GFP fluorescence. In contrast, endothelial Sirt1 deficiency was associated with a significant increase in DKK3-GFP fluorescence. Costaining with antibodies against CD31 allowed us to colocalize DKK3 with the endothelial compartment, in addition to DKK3 expression by epithelial and interstitial cells (Figure 11B). These findings support the hypothesis that endothelial dysfunction causes an increase in DKK3. FIGURE 11: View largeDownload slide Endothelial Sirt1 deficiency increases DKK3-GFP fluorescence at a basal state. (A) Representative images of DKK3-GFP fluorescence intensity of Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice versus wild-type × DKK3-eGFP × Col3.6-GFPcyan mice at a basal state (n = 3; red = CD31, green = DKK3-GFP, blue = DAPI; scale bar for ×10, 150 µm and for ×40, 30 µm). (B) Representative images of DKK3-GFP fluorescence intensity of Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice at a basal state (n = 3; red = CD31, green = DKK3-GFP, blue = DAPI; scale bar 30 µm). (C) Quantification of the DKK3-GFP fluorescence intensity. Sirt1endo−/− mice revealed increased DKK3-GFP fluorescence at a basal state as compared with wild-type mice (*P < 0.05). FIGURE 11: View largeDownload slide Endothelial Sirt1 deficiency increases DKK3-GFP fluorescence at a basal state. (A) Representative images of DKK3-GFP fluorescence intensity of Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice versus wild-type × DKK3-eGFP × Col3.6-GFPcyan mice at a basal state (n = 3; red = CD31, green = DKK3-GFP, blue = DAPI; scale bar for ×10, 150 µm and for ×40, 30 µm). (B) Representative images of DKK3-GFP fluorescence intensity of Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice at a basal state (n = 3; red = CD31, green = DKK3-GFP, blue = DAPI; scale bar 30 µm). (C) Quantification of the DKK3-GFP fluorescence intensity. Sirt1endo−/− mice revealed increased DKK3-GFP fluorescence at a basal state as compared with wild-type mice (*P < 0.05). DISCUSSION A remarkable finding of proteomic screening of RMVEC secretomes consists of the detection of DKK3 exclusively in the profibrogenic cells derived from Sirt1endo−/− mice and stimulated with TGF-β. Of note, Yin et al. [6] detected 123 proteins in the secretome of nonstimulated human umbilical vein endothelial cells. Somewhat similar findings were reported in the secretome of human adipose stromal vascular fraction cells during adipogenesis [24]. The salient feature of our study consists of juxtaposing proteomic findings in the pro- and antifibrogenic secretomes. Any differences in their respective proteomes could potentially be responsible for contrasting phenotypic properties of microvascular endothelial cells. The Wnt pathway is critical for the maintenance of stem cells and tissue renewal [25]; however, chronic activation of this pathway may have detrimental effects. Persistent activation of the Wnt pathway has been implicated in the development of renal, pulmonary and liver fibrosis [26]. While DKK1 is a well-known inhibitor of canonical Wnt/β-catenin signaling, the functions of DKK3 are much less examined, its ability to modulate Wnt signaling is controversial [27] and it is not certain whether it interacts, as DKK1 does, with Frizzled and its coreceptor LRP5/6 [28]. One study stipulates that DKK3 is a prosurvival signal that positively modulates Wnt signaling [27]. According to another study, DKK3 acts as a cytokine and induces differentiation of stem cells toward smooth muscle lineage via activating transcription factor 6 signaling [29], as well as acting as a tumor suppressor [30] or as an inhibitor of VEGFR2/Akt/mTOR signaling [31]. Recently, tubular epithelial DKK3 has been found to promote renal fibrosis [20]. Similarly, overexpression of Wnt in tubular epithelia was found to induce sterile fibrosis [32]. The data presented herein implicate DKK3 in antagonizing inhibitory effects of DKK1 on canonical Wnt/β-catenin pathway–induced conversion of renal fibroblasts to myofibroblasts, especially when occurring on the background of activated TGF-β. The TGF-β contribution may be explained by the fact that TGF-β itself leads to rapid secretion of Wnt by fibroblasts, an observation that has been implicated in cardiac myofibroblast formation and myocardial fibrosis [33]. We employ α-SMA-GFP mice to isolate renal fibroblasts and microvascular endothelial cells to be used as readouts for cell acquisition of the activated mesenchymal phenotype, as was reported previously [34, 35]. As expected, TGF-β or Wnt1 activates fibroblasts to express the marker of myofibroblasts, α-SMA-GFP. Similarly, application of DKK3 also activates fibroblasts. When DKK3 is combined with TGF-β or Wnt1, the extent of fibroblast-to-myofibroblast conversion shows no additive effect. In contrast, when DKK3 is coapplied with DKK1 in the presence of TGFβ or Wnt1, it appears to counteract the antagonistic effect of DKK1, thus suggesting that its mode of action may consist in competition with DKK1 for binding to LRP5/6. Of note, our shotgun proteomic findings in RMVEC secretomes did not detect other components of these pathways, such as Wnt, DKK1, soluble Frizzled receptor proteins or Wnt inhibitory factor, which could, in turn, interfere with endogenous Wnt binding to the β-propeller domain of LRP5/6 [36]. Our finding that DKK3 is a natural antagonist of DKK1 may explain its profibrogenic action. Yet, there should be an additional mechanism of DKK3 action to induce myofibroblast formation from renal fibroblasts since its effects are detected even in the absence of DKK1 or TGF-β. Actions of DKK3 overexpressed in the aberrant secretome of dysfunctional endothelial cells are not limited to the activation of neighboring fibroblasts. DKK3 autocrine effects on RMVECs themselves consist of suppressed capillary cords formation on matrigel and halted angiogenesis in microfluidic devices while promoting endothelial–mesenchymal transition. These findings are consistent with the previous demonstration of DKK3 interfering with VEGFR2/Akt signaling [31]. The novel finding that DKK3 promotes endothelial–mesenchymal transition is of singular importance. This transformation is characterized by acquisition of the mesenchymal phenotype with the loss of vasodilatory, antithrombogenic, antiplatelet and leukocyte adhesion and enhanced synthesis of fibrogenic extracellular matrix proteins [4]. It is an essential pathogenic step in microvascular rarefaction or malfunction associated with tissue fibrosis [37–39], atherosclerosis [40], cavernous cerebral malformations [41], heterotopic ossification [42], transplant arteriopathy [43] and cancer [44]. Thus endothelial–mesenchymal transition has a broad footprint in the pathogenesis of diverse conditions. The schema presented in Figure 12 summarizes findings presented herein and supplemented with the published relevant data. Dysfunctional endothelial cells generate excessive amounts of DKK3, potentially enhancing Wnt/β-catenin signaling in the interstitial fibroblasts and possibly pericytes. DKK1 produced by the epithelial cells [32] counteracts the effects of DKK3, except when its concentration is reduced by the distance from its source. Federico et al. [20] demonstrated that DKK3 can be also produced by the stressed tubular epithelia, and our findings in DKK3-GFP mice are consistent with that observation. In addition, Zhou et al. [21] demonstrated secretion of tubule-derived Wnt as a prerequisite to fibroblast activation and fibrosis. These overlapping gradients of pro- and antifibrogenic signals play an important role in instructing fibroblasts to undergo conversion to myofibroblasts. FIGURE 12: View largeDownload slide Cross talk of different cell lines through DKK1/DKK3/Wnt/β-catenin in the kidney. Schematic presentation of the DKK1/DKK3/Wnt/β-catenin interactions between dysfunctional endothelial cells, interstitial fibroblasts, pericytes and epithelial cells. FIGURE 12: View largeDownload slide Cross talk of different cell lines through DKK1/DKK3/Wnt/β-catenin in the kidney. Schematic presentation of the DKK1/DKK3/Wnt/β-catenin interactions between dysfunctional endothelial cells, interstitial fibroblasts, pericytes and epithelial cells. The key conclusion of this study is that microenvironmental signals derived from dysfunctional endothelial cells instruct (myo)fibroblast fate determination. Endothelial cells defective in Sirt1 secrete DKK3 and coerce fibroblast-to-myofibroblast conversion. In conjunction with the previous findings on Wnt and DKK3 secretion from stressed renal tubular epithelial cells [20, 32, 45], it appears that renal microvascular endothelium acts as a ‘gatekeeper’ for Wnt agonists and antagonists reaching renal fibroblasts, as well as endothelial cells themselves. Collectively, these findings support our hypothesis on endothelial contribution to the ‘third pathway’ of fibrogenesis [18]. By intercepting persistent DKK3 signals, shown to be pathologic, using the Wnt pathway inhibitor sulindac sulfide, it becomes possible to achieve therapeutic effects by reducing fibrogenesis in a model of chronic kidney disease. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS The authors are grateful to Dr Yujiro Kida for his initial help in isolating renal microvascular endothelial cells and collecting their conditioned media. FUNDING These studies have been supported in part by the Dr. Werner Jackstädt Foundation (to M.L.), the Westchester Community Trust Foundation – Renal Research Fund and NIH grants DK54602, DK052783 and DK45462 (to M.S.G.) and R21 AR055750 (to D.W.R.). N.L.J. was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT & Future Planning (2015R1A2A1A09005662 and NRF-2016R1A4A1010796). AUTHORS’ CONTRIBUTIONS M.L., H.D., G.A.M. and M.S.G. designed the study. M.L., H.D. and S.D. carried out experiments. M.L., H.D., B.B.R. analyzed the data. M.L. and H.D. made the figures. M.L., H.D., G.A.M. and M.S.G. drafted and revised the manuscript. N.L.J. provided the microfluidic devices. D.W.R. provided the DKK3-eGFP × Col3.6-GFPcyan mice. All authors approved the final version of the manuscript. CONFLICT OF INTEREST STATEMENT None declared. REFERENCES 1 Butler JM , Kobayashi H , Rafii S. Instructive role of the vascular niche in promoting tumour growth and tissue repair by angiocrine factors . Nat Rev Cancer 2010 ; 10 : 138 – 146 Google Scholar CrossRef Search ADS PubMed 2 Hemmen K , Reinl T , Buttler K et al. High-resolution mass spectrometric analysis of the secretome from mouse lung endothelial progenitor cells . Angiogenesis 2011 ; 14 : 163 – 172 Google Scholar CrossRef Search ADS PubMed 3 Pula G , Mayr U , Evans C et al. Proteomics identifies thymidine phosphorylase as a key regulator of the angiogenic potential of colony-forming units and endothelial progenitor cell cultures . Circ Res 2009 ; 104 : 32 – 40 Google Scholar CrossRef Search ADS PubMed 4 Dejana E , Hirschi KK , Simons M. The molecular basis of endothelial cell plasticity . Nat Commun 2017 ; 8 : 14361 Google Scholar CrossRef Search ADS PubMed 5 Rafii S , Butler JM , Ding BS. Angiocrine functions of organ-specific endothelial cells . Nature 2011 ; 529 : 316 – 325 Google Scholar CrossRef Search ADS 6 Yin X , Bern M , Xing Q et al. Glycoproteomic analysis of the secretome of human endothelial cells . Mol Cell Proteomics 2013 ; 12 : 956 – 978 Google Scholar CrossRef Search ADS PubMed 7 Vasko R , Xavier S , Chen J et al. Endothelial sirtuin 1 deficiency perpetrates nephrosclerosis through downregulation of matrix metalloproteinase-14: relevance to fibrosis of vascular senescence . J Am Soc Nephrol 2014 ; 25 : 276 – 291 Google Scholar CrossRef Search ADS PubMed 8 Xavier S , Vasko R , Matsumoto K et al. Curtailing endothelial TGF-β signaling is sufficient to reduce endothelial-mesenchymal transition and fibrosis in CKD . J Am Soc Nephrol 2015 ; 26 : 817 – 829 Google Scholar CrossRef Search ADS PubMed 9 Potente M , Ghaeni L , Baldessari D et al. SIRT1 controls endothelial angiogenic functions during vascular growth . Genes Dev 2007 ; 21 : 2644 – 2658 Google Scholar CrossRef Search ADS PubMed 10 Chen J , Xavier S , Moskowitz-Kassai E et al. Cathepsin cleavage of sirtuin 1 in endothelial progenitor cells mediates stress-induced premature senescence . Am J Pathol 2012 ; 180 : 973 – 983 Google Scholar CrossRef Search ADS PubMed 11 Kida Y , Goligorsky MS. Sirtuins, cell senescence, and vascular aging . Can J Cardiol 2016 ; 32 : 634 – 641 Google Scholar CrossRef Search ADS PubMed 12 Ziyadeh FN , Hoffman BB , Han DC et al. Long-term prevention of renal insufficiency, excess matrix gene expression, and glomerular mesangial matrix expansion by treatment with monoclonal antitransforming growth factor-beta antibody in db/db diabetic mice . Proc Natl Acad Sci USA 2000 ; 97 : 8015 – 8020 Google Scholar CrossRef Search ADS PubMed 13 Chan LL , Huang J , Hagiwara Y et al. Discriminating multiplexed GFP reporters in primary articular chondrocyte cultures using image cytometry . J Fluoresc 2014 ; 24 : 1041 – 1053 Google Scholar CrossRef Search ADS PubMed 14 Mori Y , Adams D , Hagiwara Y et al. Identification of a progenitor cell population destined to form fracture fibrocartilage callus in Dickkopf-related protein 3-green fluorescent protein reporter mice . J Bone Miner Metab 2016 ; 34 : 606 – 614 Google Scholar CrossRef Search ADS PubMed 15 Yokota T , Kawakami Y , Nagai Y et al. Bone marrow lacks a transplantable progenitor for smooth muscle type alpha-actin-expressing cells . Stem Cells 2006 ; 24 : 13 – 22 Google Scholar CrossRef Search ADS PubMed 16 Kalajzic Z , Li H , Wang LP et al. Use of an alpha-smooth muscle actin GFP reporter to identify an osteoprogenitor population . Bone 2008 ; 43 : 501 – 510 Google Scholar CrossRef Search ADS PubMed 17 Kim S , Lee H , Chung M et al. Engineering of functional, perfusable 3D microvascular networks on a chip . Lab Chip 2013 ; 13 : 1489 – 1500 Google Scholar CrossRef Search ADS PubMed 18 Lipphardt M , Song JW , Matsumoto K et al. The third path of tubulointerstitial fibrosis: aberrant endothelial secretome . Kidney Int 2017 ; 92 : 558 – 568 Google Scholar CrossRef Search ADS PubMed 19 Lipphardt M , Song JW , Ratliff BB et al. Endothelial dysfunction is a super-inducer of syndecan 4: fibrogenic role of its ectodomain . Am J Physiol Heart Circ Physiol 2018 ; 314 : H484 – H496 Google Scholar CrossRef Search ADS PubMed 20 Federico G , Meister M , Mathow D et al. Tubular Dickkopf-3 promotes the development of renal atrophy and fibrosis . JCI Insight 2016 ; 1 : e84916 Google Scholar CrossRef Search ADS PubMed 21 Zhou D , Fu H , Zhang L et al. Tubule-derived wnts are required for fibroblast activation and kidney fibrosis . J Am Soc Nephrol 2017 ; 28 : 2322 – 2336 Google Scholar CrossRef Search ADS PubMed 22 Karamariti E , Margariti A , Winkler B et al. Smooth muscle cells differentiated from reprogrammed embryonic lung fibroblasts through DKK3 signaling are potent for tissue engineering of vascular grafts . Circ Res 2013 ; 112 : 1433 – 1443 Google Scholar CrossRef Search ADS PubMed 23 Bravi L , Rudini N , Cuttano R et al. Sulindac metabolites decrease cerebrovascular malformations in CCM3-knockout mice . Proc Natl Acad Sci USA 2015 ; 112 : 8421 – 8426 Google Scholar CrossRef Search ADS PubMed 24 Kim J , Choi YS , Lim S et al. Comparative analysis of the secretory proteome of human adipose stromal vascular fraction cells during adipogenesis . Proteomics 2010 ; 10 : 394 – 405 Google Scholar CrossRef Search ADS PubMed 25 Clevers H , Loh KM , Nusse R. Stem cell signaling. An integral program for tissue renewal and regeneration: Wnt signaling and stem cell control . Science 2014 ; 346 : 1248012 Google Scholar CrossRef Search ADS PubMed 26 Miao CG , Yang YY , He X et al. Wnt signaling in liver fibrosis: progress, challenges and potential directions . Biochimie 2013 ; 95 : 2326 – 2335 Google Scholar CrossRef Search ADS PubMed 27 Nakamura RE , Hackam AS. Analysis of Dickkopf3 interactions with Wnt signaling receptors . Growth Factors 2010 ; 28 : 232 – 242 Google Scholar CrossRef Search ADS PubMed 28 Krupnik VE , Sharp JD , Jiang C et al. Functional and structural diversity of the human Dickkopf gene family . Gene 1999 ; 238 : 301 – 313 Google Scholar CrossRef Search ADS PubMed 29 Wang X , Karamariti E , Simpson R et al. Dickkopf homolog 3 induces stem cell differentiation into smooth muscle lineage via ATF6 signalling . J Biol Chem 2015 ; 290 : 19844 – 19852 Google Scholar CrossRef Search ADS PubMed 30 Veeck J , Dahl E. Targeting the Wnt pathway in cancer: the emerging role of Dickkopf-3 . Biochim Biophys Acta 2012 ; 1825 : 18 – 28 Google Scholar PubMed 31 Kim BR , Lee EJ , Seo SH et al. Dickkopf-3 (DKK-3) obstructs VEGFR-2/Akt/mTOR signaling cascade by interacting of β2-microglobulin (β2M) in ovarian tumorigenesis . Cell Signal 2015 ; 27 : 2150 – 2159 Google Scholar CrossRef Search ADS PubMed 32 Maarouf OH , Aravamudhan A , Rangarajan D et al. Paracrine Wnt1 drives interstitial fibrosis without inflammation by tubulointerstitial cross-talk . J Am Soc Nephrol 2016 ; 27 : 781 – 790 Google Scholar CrossRef Search ADS PubMed 33 Blyszczuk P , Müller-Edenborn B , Valenta T et al. Transforming growth factor-β-dependent Wnt secretion controls myofibroblast formation and myocardial fibrosis progression in experimental autoimmune myocarditis . Eur Heart J 2017 ; 38 : 1413 – 1425 Google Scholar PubMed 34 Kida Y , Zullo JA , Goligorsky MS. Endothelial sirtuin 1 inactivation enhances capillary rarefaction and fibrosis following kidney injury through Notch activation . Biochem Biophys Res Commun 2016 ; 478 : 1074 – 1079 Google Scholar CrossRef Search ADS PubMed 35 Matsumoto K , Xavier S , Chen J et al. Instructive role of the microenvironment in preventing renal fibrosis . Stem Cells Transl Med 2017 ; 6 : 992 – 1005 Google Scholar CrossRef Search ADS PubMed 36 Bourhis E , Wang W , Tam C et al. Wnt antagonists bind through a short peptide to the first β-propeller domain of LRP5/6 . Structure 2011 ; 19 : 1433 – 1442 Google Scholar CrossRef Search ADS PubMed 37 Zeisberg EM , Tarnavski O , Zeisberg M et al. Endothelial-to-mesenchymal transition contributes to cardiac fibrosis . Nat Med 2007 ; 13 : 952 – 961 Google Scholar CrossRef Search ADS PubMed 38 O'Riordan E , Mendelev N , Patschan S et al. Chronic NOS inhibition actuates endothelial-mesenchymal transformation . Am J Physiol Heart Circ Physiol 2007 ; 292 : H285 – H294 Google Scholar CrossRef Search ADS PubMed 39 Piera-Velazquez S , Li Z , Jimenez SA. Role of endothelial-mesenchymal transition (EndoMT) in the pathogenesis of fibrotic disorders . Am J Pathol 2011 ; 179 : 1074 – 1080 Google Scholar CrossRef Search ADS PubMed 40 Chen PY , Qin L , Baeyens N et al. Endothelial-to-mesenchymal transition drives atherosclerosis progression . J Clin Invest 2015 ; 125 : 4514 – 4528 Google Scholar CrossRef Search ADS PubMed 41 Maddaluno L , Rudini N , Cuttano R et al. EndMT contributes to the onset and progression of cerebral cavernous malformations . Nature 2013 ; 498 : 492 – 496 Google Scholar CrossRef Search ADS PubMed 42 Medici D , Shore EM , Lounev VY et al. Conversion of vascular endothelial cells into multipotent stem-like cells . Nat Med 2010 ; 16 : 1400 – 1406 Google Scholar CrossRef Search ADS PubMed 43 Tellides G , Pober JS. Inflammatory and immune responses in the arterial media . Circ Res 2015 ; 116 : 312 – 322 Google Scholar CrossRef Search ADS PubMed 44 Zeisberg EM , Potenta S , Xie L et al. Discovery of endothelial to mesenchymal transition as a source for carcinoma-associated fibroblasts . Cancer Res 2007 ; 67 : 10123 – 10128 Google Scholar CrossRef Search ADS PubMed 45 Gröne EF , Federico G , Nelson PJ et al. The hormetic functions of Wnt pathways in tubular injury . Pflugers Arch 2017 ; 469 : 899 – 906 Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nephrology Dialysis Transplantation Oxford University Press

Dickkopf-3 in aberrant endothelial secretome triggers renal fibroblast activation and endothelial–mesenchymal transition

Loading next page...
 
/lp/ou_press/dickkopf-3-in-aberrant-endothelial-secretome-triggers-renal-fibroblast-bUUl4uhlX1
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
ISSN
0931-0509
eISSN
1460-2385
D.O.I.
10.1093/ndt/gfy100
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Our laboratory has previously demonstrated that Sirt1endo−/− mice show endothelial dysfunction and exaggerated renal fibrosis, whereas mice with silenced endothelial transforming growth factor beta (TGF-β) signaling are resistant to fibrogenic signals. Considering the fact that the only difference between these mutant mice is confined to the vascular endothelium, this indicates that secreted substances contribute to these contrasting responses. Methods We performed an unbiased proteomic analysis of the secretome of renal microvascular endothelial cells (RMVECs) isolated from these two mutants. We cultured renal fibroblasts and RMVECs and used microfluidic devices for coculturing. Results Dickkopf-3 (DKK3), a putative ligand of the Wnt/β-catenin pathway, was present exclusively in the fibrogenic secretome. In cultured fibroblasts, DKK3 potently induced myofibroblast activation. In addition, DKK3 antagonized effects of DKK1, a known inhibitor of the Wnt pathway, in conversion of fibroblasts to myofibroblasts. In RMVECs, DKK3 induced endothelial–mesenchymal transition and impaired their angiogenic competence. The inhibition of endothelial outgrowth, enhanced myofibroblast formation and endothelial–mesenchymal transition were confirmed in coculture. In reporter DKK3-eGFP × Col3.6-GFPcyan mice, DKK3 was marginally expressed under basal conditions. Adriamycin-induced nephropathy resulted in upregulation of DKK3 expression in tubular and, to a lesser degree, endothelial compartments. Sulindac sulfide was found to exhibit superior Wnt pathway-suppressive action and decreased DKK3 signals and the extent of renal fibrosis. Conclusions In conclusion, this unbiased proteomic screen of the profibrogenic endothelial secretome revealed DKK3 acting as an agonist of the Wnt pathway, enhancing formation of myofibroblasts and endothelial–mesenchymal transition and impairing angiogenesis. A potent inhibitor of the Wnt pathway, sulindac sulfide, suppressed nephropathy-induced DKK3 expression and renal fibrosis. endothelial–fibroblast cocultures, endothelial secretome, microfluidic device, renal fibrosis, Wnt pathway INTRODUCTION It has been recognized that endothelial cells normally produce diverse paracrine-tropic and angiocrine factors necessary for differentiation and regeneration of tissues as dissimilar as pancreatic acini, neurons, hematopoietic precursors, hepatocytes and alveolar epithelia [1]. For instance, senescent endothelial cells [2, 3] or endothelial cells undergoing mesenchymal transition [4] not only disrupt the function of the endothelial lining of the vessels but also desert their instructive angiocrine functions and affect the neighboring cells by their secretome. The term ‘secretome’ comprises all proteins secreted from the plasma membrane and released with exosomes and other cell organellar products harvested from the extracellular compartment [5]. The recent analysis of the secretome of cultured human umbilical vein endothelial cells disclosed 123 secreted proteins [6]. We have recently generated two strains of mutant mice with endothelial cells either lacking exon 4 encoding for the deacetylase activity of sirtuin 1 (Sirt1endo−/−) or deficient in transforming growth factor beta (TGF-β) receptor II (TGFβRIIendo+/−); the former mutant is prone to develop fibrosis and the latter is protected from fibrosis [7, 8]. Since endothelial Sirt1 deficiency is a consistent companion of a range of chronic cardiovascular, metabolic and renal diseases [9–11], clearly the Sirt1endo−/− mouse model has broad applicability. On the other hand, silencing of TGF-β signaling, as in our TGFβRIIendo+/− model, is one of the goals of pharmaceutical research for preventing fibrosis [8, 12]. Sirt1-deficient endothelial cells have all the signs of dysfunction: reduced endothelium-dependent vasorelaxation, microvascular rarefaction, impaired migration, sprouting, matrilytic activity, loss of glycocalyx, premature senescence and exaggerated fibrotic response [7–11]. In contrast, the TGF-βRIIendo+/− model has no distinctive phenotype under basal conditions but develops much reduced fibrosis after induction of unilateral ureteral obstruction or the chronic phase of folic acid nephropathy [8]. These contrasting strains not only emphasize the fact that messages from the endothelium are instructive for fibroblasts but also represent a useful tool to juxtapose endothelial secretomes as the only differentially expressed variables potentially explaining the distinct end effects. Hence one of the goals of the present study was to perform mining of the profibrogenic endothelial secretome. The detection of Dickkopf-3 (DKK3) uniquely present in this aberrant secretome prompted the undertaking of studies in cultured renal microvascular endothelial cells (RMVECs) and fibroblasts, in their cocultures and in whole animals to elucidate the functional significance of this finding. MATERIALS AND METHODS Mouse models used in the study All animal experiments were performed in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and was approved by the Institutional Animal Care and Use Committee. Sirt1endo−/− mice. The endothelial Sirt1-deleted mouse model was established by cross-breeding B6 129-Sirt1tm1Ygu/J, which harbors a deletion of exon 4 encoding for the deacetylase catalytic domain, with Tie2-Cre transgenic mice expressing cre-recombinase in vascular endothelial cells (both from Jackson Laboratory, Bar Harbor, ME, USA) [7]. The resulting Sirt1endo+/− mice were mated with Sirt1Flox/Flox mice to produce endothelial-deleted Sirt1endo−/− mutant mice. TGFβRIIendo+/− mice. These mice were generated by Xavier et al. [8] as detailed previously. DKK3-eGFP × Col3.6-GFPcyan mice. These mice on a CD1 background were generated by Dr. David W. Rowe [13, 14]. Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice. These were established by cross-breeding Sirt−/− mice with DKK3-eGFP × Col3.6-GFPcyan mice. α-Smooth muscle actin–green fluorescent protein mice. α-Smooth muscle actin (SMA)–green fluorescent protein (GFP) transgenic mice on a C57BL/6 J background were generously supplied by Dr. Ivo Kalajzic (University of Connecticut Health Center, Farmington, CT, USA). Mice were originally developed by Dr. Jen-Yue Tsai (National Eye Institute, NIH) and carried a regulatory sequence of α-SMA gene spanning 1074 bp of the 5′-flanking region, the transcription start site, 48 bp of exon 1, the 2.5-kbp intron 1 and the 15-bp exon 2 of mouse α-SMA. GFP is specifically expressed in both vascular and nonvascular smooth muscle cells [15, 16]. Mice were housed in the animal care facility of New York Medical College (25°C, 50% humidity and 12-h dark/light cycle) with free access to food and water. Microfluidic device studies of spatially separated RMVEC fibroblast cocultures We performed all microfluidic device experiments in accordance with Kim et al.’s protocol [17]. Briefly, a fibrinogen solution was prepared by dissolving 2.5 mg/mL bovine fibrinogen in phosphate-buffered saline (Corning, Corning, NY, USA). RMVECs, preloaded with a CellTracker-red (Fisher Scientific, Pittsburgh, PA, USA), and renal fibroblasts, all obtained from α-SMA-GFP mice, were individually suspended in the fibrinogen solution. The cell solutions were mixed with thrombin (0.5 U/mL) and then promptly pipetted into the central and peripheral channels leaving the spaces between endothelial cells and fibroblasts for the respective ingrowth. After clotting for 3–5 min at room temperature, the inlet reservoirs of the cell culture medium channels were loaded with endothelial cell culture medium and then a vacuum was gently applied at the outlet reservoirs to fill the hydrophobic channels. The microfluidic devices were incubated at 37°C in a humidified 5% carbon dioxide atmosphere. The cell culture medium was refilled with fresh endothelial cell culture medium every 48 h. Statistical analysis Values are given as the mean ± standard error of the mean unless stated otherwise. Data were analyzed using independent t-test or analysis of variance (ANOVA) with post hoc analysis for multiple group comparisons using the Bonferroni method. A P-value <0.05 was considered statistically significant. All statistical analyses were performed with NCSS 10 (NCSS, Kaysville, UT, USA). RESULTS Mass spectrometry of the RMVEC secretome CD31 magnetically isolated high-purity populations of RMVEC from control wild-type, Sirt1endo−/− or TGF-βRIIendo+/− mice were expanded and exposed to a vehicle or TGF-β and conditioned medium was collected 48 h later. The latter was analyzed using unbiased, nontargeted tandem mass spectrometry as detailed in the Supplementary Methods section. The tryptic digests were analyzed using the Thermo Scientific Q Exactive Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) and the tandem mass spectra were extracted. Using this technique we were able to detect 332 nonredundant proteins that belong to diverse categories, as reported previously [18, 19]. We argue that the differential signatures of proteins secreted exclusively by Sirt1endo−/− versus TGF-βRIIendo+/− could theoretically contain protein culprits for pro- and antifibrogenic phenotypes of the respective mice. Among those differentially expressed secreted proteins, DKK3 was present exclusively in the fibrogenic secretome (Figure 1). Whether DKK3 secreted by dysfunctional endothelial cells may interfere with DKK1 in regulating Wnt signaling in surrounding cells is entirely unknown. For these reasons, we decided to explore the effects of DKK3 on the Wnt pathway in renal fibroblasts and RMVECs. FIGURE 1: View largeDownload slide DKK3 is elevated in the secretome of Sirt1endo−/− mice. RMVEC from control wild-type and Sirt1endo−/− mice were expanded, exposed to TGF-β (5 ng/mL) and conditioned media were collected and subjected to mass spectrometry analyses. Among the differentially regulated proteins, DKK3 was highly enriched in the secretome of RMVECs obtained from Sirt1endo−/− mice upon stimulation with TGF-β (***P < 0.05). FIGURE 1: View largeDownload slide DKK3 is elevated in the secretome of Sirt1endo−/− mice. RMVEC from control wild-type and Sirt1endo−/− mice were expanded, exposed to TGF-β (5 ng/mL) and conditioned media were collected and subjected to mass spectrometry analyses. Among the differentially regulated proteins, DKK3 was highly enriched in the secretome of RMVECs obtained from Sirt1endo−/− mice upon stimulation with TGF-β (***P < 0.05). Paracrine effects of DKK3 in renal fibroblasts Fibroblasts were isolated from α-SMA-GFP mouse kidneys using positive (enrichment of PDGFR+ cells) and negative (depletion of CD31+cells) selection with magnetic beads according to the previously established protocol as detailed in the Methods section. Under resting conditions, very few fibroblasts expressed α-SMA-GFP. Application of DKK3 was combined with the known agonist of the Wnt pathway, Wnt1, or a known antagonist of the Wnt pathway, DKK1, each of these components with or without the application of TGF-β to fibroblasts in culture and monitoring at the time of appearance of GFP fluorescence. We expected that this protocol might reveal modulatory effects of DKK3 on the time and extent of fibroblast-to-myofibroblast conversion. Data indicated that DKK3 (10 µg/mL) alone induced a myofibroblastic phenotype (Figure 2); however, it did not further alter TGF-β-induced responses. Application of Wnt1 (100 ng/mL) alone similarly induced GFP expression in α-SMA-GFP renal fibroblasts, whereas DKK1 coapplication reduced fibroblast activation. When DKK3 was applied in combination with Wnt1 and DKK1, it resulted in the antagonistic signal from DKK1 being weakly overridden by DKK3. The presence of TGF-β has not perceptibly altered the directionality of these responses but induced a strong counter-regulation by DKK3 of DKK1-induced inhibition of myofibroblasts formation. Hence data generated in renal fibroblasts is consistent with the notion that DKK3 acts as an agonist of the Wnt pathway and is capable of counteracting antagonistic effects of DKK1, especially in the presence of TGF-β. FIGURE 2: View largeDownload slide DKK3 induces fibroblast-to-myofibroblast transition. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured renal fibroblasts isolated from α-SMA-GFP reporter mice 24 h after exposure to the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Data indicate that DKK3 (10 µg/mL) alone induced the myofibroblastic phenotype; however, it did not further enhance responses to TGF-β. *, indicates the control group to be statistically significantly different from all other groups; #, indicates the Wnt/DKK3 group to be statistically significantly different from all other groups except the Wnt group (P < 0.05). (C) Quantification of α-SMA-GFP fluorescence intensity. DKK3 did not further enhance responses to TGF-β. *, indicates the control group to be statistically significantly different from all other groups; #, indicates the TGFβ/Wnt/DKK1 group to be statistical significantly different from all other groups as well (P < 0.05). FIGURE 2: View largeDownload slide DKK3 induces fibroblast-to-myofibroblast transition. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured renal fibroblasts isolated from α-SMA-GFP reporter mice 24 h after exposure to the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Data indicate that DKK3 (10 µg/mL) alone induced the myofibroblastic phenotype; however, it did not further enhance responses to TGF-β. *, indicates the control group to be statistically significantly different from all other groups; #, indicates the Wnt/DKK3 group to be statistically significantly different from all other groups except the Wnt group (P < 0.05). (C) Quantification of α-SMA-GFP fluorescence intensity. DKK3 did not further enhance responses to TGF-β. *, indicates the control group to be statistically significantly different from all other groups; #, indicates the TGFβ/Wnt/DKK1 group to be statistical significantly different from all other groups as well (P < 0.05). Next, we examined in depth (Figure 3) the dose–response relations between DKK1 and DKK3. Data showed a complex relationship: DKK3 is able to convert fibroblasts to myofibroblasts only at low levels of DKK1. The sharp increase in sensitivity of fibroblasts to DKK3 at the lower spectrum of DKK1 concentrations would argue that at distant sites from DKK3-secreting dysfunctional endothelial cells, DKK1 effects predominate and Wnt signaling is partially counteracted. At the sites nearest to the dysfunctional endothelial cells, such as pericytes, the gradients of DKK3 and DKK1 favor the former and therefore these cells could be most influenced by the endothelium-secreted DKK3. On the other hand, based on studies from Federico et al. [20], interstitial fibroblasts may be more influenced by the epithelium-secreted DKK3. In addition, Wnt secretion by the damaged epithelium has been shown by Zhou et al. [21], thus placing interstitial fibroblasts under the cooperative influence of both DKK3 and Wnt to convert them to myofibroblasts. This intricate gradient field of profibrogenic messengers focusing on pericytes and interstitial fibroblasts, still insufficiently investigated, may play a substantial role in conversion of these cell types into myofibroblasts. FIGURE 3: View largeDownload slide Dose response of DKK1 and DKK3 on renal fibroblasts. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured renal fibroblasts isolated from α-SMA-GFP reporter mice 48 h after exposure to the indicated proteins and an additional 48 h after exposure to TGF-β(n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Data indicate that DKK3 in a dose-responsive manner is able to convert fibroblasts to myofibroblasts only at the low levels of DKK1. (C) Quantification of α-SMA-GFP fluorescence intensity. TGF-β did not further enhance responses to DKK3. FIGURE 3: View largeDownload slide Dose response of DKK1 and DKK3 on renal fibroblasts. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured renal fibroblasts isolated from α-SMA-GFP reporter mice 48 h after exposure to the indicated proteins and an additional 48 h after exposure to TGF-β(n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Data indicate that DKK3 in a dose-responsive manner is able to convert fibroblasts to myofibroblasts only at the low levels of DKK1. (C) Quantification of α-SMA-GFP fluorescence intensity. TGF-β did not further enhance responses to DKK3. Autocrine effects of the secretome and how does DKK3 affect the angiogenesis of endothelial cells Since DKK3 released to the extracellular milieu acquires the functions of a cytokine [22], it is plausible that it may exert not only paracrine effects on mesenchymal cells but also autocrine effects. DKK3 autocrine actions as possible angiogenic modulators were studied using RMVECs obtained from α-SMA-GFP mouse kidneys, again employed as a reporter system. Application of DKK3 was performed in the presence or absence of Wnt1 (to induce the ‘on’ state of frizzled receptors), DKK1 (to confirm its well-described inhibition of Wnt actions) and TGF-β. Results demonstrated (Figure 4) that 48 h after these treatments, control cells gradually underwent endothelial–mesenchymal transition (as judged by the acquisition of α-SMA-GFP signal). Moreover, Wnt1 and DKK3 further promoted endothelial–mesenchymal transition, but their effects were nonadditive. As expected, the addition of DKK1 to Wnt1 resulted in a significant reduction of α-SMA-GFP fluorescence. Remarkably, the addition of DKK3 to the combined treatment with Wnt1 and DKK1 reduced the inhibitory effect of DKK1 on endothelial–mesenchymal transition. When the above-mentioned treatments were applied in the presence of TGF-β, acquisition of α-SMA-GFP fluorescence by RMVECs was also inhibited by DKK1. However, under those conditions, DKK3 afforded even higher antagonistic activity against DKK1. FIGURE 4: View largeDownload slide Autocrine effects of DKK3. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured RMVECs isolated from α-SMA-GFP reporter mice 48 h after exposure to the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Control cells gradually underwent endothelial–mesenchymal transition (as judged by the acquisition of α-SMA-GFP signal). Moreover, Wnt1 and DKK3 further promoted endothelial–mesenchymal transition, but their effects were nonadditive. As expected, the addition of DKK1 to Wnt1 resulted in a significant reduction in α-SMA-GFP fluorescence. *, labeled groups are statistically significantly different from the nonlabeled groups (P < 0.05). (C) Quantification of α-SMA-GFP fluorescence intensity. *, indicates the control group and TGFβ/Wnt/DKK1/DKK3 group to be significantly different from all other groups; #, indicates the TGFβ/Wnt/DKK1 group to be significantly different from all other groups (P < 0.05). FIGURE 4: View largeDownload slide Autocrine effects of DKK3. (A) Representative images of α-SMA-GFP fluorescence intensity of cultured RMVECs isolated from α-SMA-GFP reporter mice 48 h after exposure to the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of α-SMA-GFP fluorescence intensity. Control cells gradually underwent endothelial–mesenchymal transition (as judged by the acquisition of α-SMA-GFP signal). Moreover, Wnt1 and DKK3 further promoted endothelial–mesenchymal transition, but their effects were nonadditive. As expected, the addition of DKK1 to Wnt1 resulted in a significant reduction in α-SMA-GFP fluorescence. *, labeled groups are statistically significantly different from the nonlabeled groups (P < 0.05). (C) Quantification of α-SMA-GFP fluorescence intensity. *, indicates the control group and TGFβ/Wnt/DKK1/DKK3 group to be significantly different from all other groups; #, indicates the TGFβ/Wnt/DKK1 group to be significantly different from all other groups (P < 0.05). Having demonstrated the capability of DKK3 to induce endothelial–mesenchymal transition, we explored whether DKK3 influences the angiogenesis of RMVECs. Therefore we plated RMVECs on matrigel, exposed them to DKK3 and examined the ability to form a capillary network. We quantified angiogenic parameters using NIH-developed Angiotool software. The total length, percentage area occupied by capillary-like structures and the number of bifurcations were all significantly reduced by DKK3 (Figure 5). FIGURE 5: View largeDownload slide Angiogenesis assay of cultured RMVECs on matrigel. (A) Representative images of cultured RMVECs on matrigel with and without DKK3 (n = 5; scale bar 150 µm). (B–D) Quantification of the vessels percentage area, total number of junctions and total vessels length. RMVECs treated with DKK3 experienced a profound decrease in vessels percentage area, total number of branches and total vessels length after 96 h of treatment (*P < 0.05). FIGURE 5: View largeDownload slide Angiogenesis assay of cultured RMVECs on matrigel. (A) Representative images of cultured RMVECs on matrigel with and without DKK3 (n = 5; scale bar 150 µm). (B–D) Quantification of the vessels percentage area, total number of junctions and total vessels length. RMVECs treated with DKK3 experienced a profound decrease in vessels percentage area, total number of branches and total vessels length after 96 h of treatment (*P < 0.05). Microfluidic device cocultures of renal endothelial and fibroblastic cells Since the above-mentioned two-dimensional (2D) cultures of endothelial cells and fibroblasts lack the complexity of their biological three-dimensional topography and interactions, we established a coculture system using microfluidic devices according to the previously published protocols [17]. The microfluidic devices were fabricated using Sylgard 184 (Dow Corning, Corning, NY, USA) by lithography and replica modeling as described previously [17] and illustrated in Figure 6A. RMVECs and renal fibroblasts were isolated from kidneys of α-SMA-GFP mice. The RMVECs were prelabeled with red CellTracker and introduced into nanochannels in the fibrin gel (fibrinogen and thrombin) alongside fibroblasts. DKK3 was added to the endothelial culture medium and the outgrowth of RMVECs toward fibroblasts as well as the proportion of RMVECs undergoing endothelial–mesenchymal transition and the proportion of fibroblasts becoming α-SMA-GFP positive were monitored by intravital microscopy and quantified. Analysis of angiogenic parameters showed a reduction in the total length, percentage area occupied by capillary-like structures and number of bifurcations with the addition of DKK3 (Figure 6B–I). Furthermore, we documented that with the addition of DKK3 to the culture medium, the extent of endothelial–mesenchymal transition and fibroblast–myofibroblast transition was significantly increased (Figure 7). FIGURE 6: View largeDownload slide Angiogenesis assay of microfluidic cocultures of renal endothelial and fibroblastic cells. (A) Illustration of a microfluidic device with five nanochannels (one central, two outer channels and two intersectional channels). (B) Representative bright-field images of cocultured RMVECs and renal fibroblasts with and without DKK3 at Day 4 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. Under control conditions, RMVECs began forming a capillary network inside the central channel with early outgrowth toward the intersectional channel, whereas the addition of DKK3 decreased the vessels percentage, the total number of branches and the total vessels length (n = 6; scale bar 150 µm). (C–E) Quantification of the vessels percentage area, total number of junctions and total vessels length on Day 4 (*P < 0.05). (F) Representative bright-field images of cocultured RMVECs and renal fibroblasts with and without DKK3 on Day 9 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. By Day 9, RMVECs formed a capillary network in the intersectional channel under control conditions. A capillary network or branching could not be recognized with the addition of DKK3 (n = 6; scale bar 150 µm). (G–I) Quantification of the vessels percentage area, total number of junctions and total vessels length on Day 9 (*P < 0.05). FIGURE 6: View largeDownload slide Angiogenesis assay of microfluidic cocultures of renal endothelial and fibroblastic cells. (A) Illustration of a microfluidic device with five nanochannels (one central, two outer channels and two intersectional channels). (B) Representative bright-field images of cocultured RMVECs and renal fibroblasts with and without DKK3 at Day 4 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. Under control conditions, RMVECs began forming a capillary network inside the central channel with early outgrowth toward the intersectional channel, whereas the addition of DKK3 decreased the vessels percentage, the total number of branches and the total vessels length (n = 6; scale bar 150 µm). (C–E) Quantification of the vessels percentage area, total number of junctions and total vessels length on Day 4 (*P < 0.05). (F) Representative bright-field images of cocultured RMVECs and renal fibroblasts with and without DKK3 on Day 9 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. By Day 9, RMVECs formed a capillary network in the intersectional channel under control conditions. A capillary network or branching could not be recognized with the addition of DKK3 (n = 6; scale bar 150 µm). (G–I) Quantification of the vessels percentage area, total number of junctions and total vessels length on Day 9 (*P < 0.05). FIGURE 7: View largeDownload slide DKK3 induced endothelial–mesenchymal and fibroblast-to-myofibroblast transition in microfluidic device cocultures of renal endothelial and fibroblastic cells. (A–C) Representative images of cocultured RMVECs and renal fibroblasts with and without DKK3 on Days 1, 3 and 8 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. RMVECs prelabeled with a CellTracker are shown in red. Green = α-SMA-GFP (n = 6; scale bar 150 µm). (D–F) Quantification of the α-SMA-GFP fluorescence intensity on Days 1, 3, and 8 (*P < 0.05). The addition of DKK3 to the culture medium caused more intense endothelial–mesenchymal and fibroblast-to-myofibroblast transition. FIGURE 7: View largeDownload slide DKK3 induced endothelial–mesenchymal and fibroblast-to-myofibroblast transition in microfluidic device cocultures of renal endothelial and fibroblastic cells. (A–C) Representative images of cocultured RMVECs and renal fibroblasts with and without DKK3 on Days 1, 3 and 8 of culture. The central channel harboring RMVECs is shown on the left side of the image. One outer channel harboring renal fibroblasts is shown on the right side of the image. The intersectional channel is shown in the center of the image. RMVECs prelabeled with a CellTracker are shown in red. Green = α-SMA-GFP (n = 6; scale bar 150 µm). (D–F) Quantification of the α-SMA-GFP fluorescence intensity on Days 1, 3, and 8 (*P < 0.05). The addition of DKK3 to the culture medium caused more intense endothelial–mesenchymal and fibroblast-to-myofibroblast transition. In another approach, we cocultured RMVECs and renal fibroblasts for 12 days in endothelial cell culture medium. We then added DKK3 (10 µg/mL) to the medium and cultured for 2 additional days. As shown in Figure 8, the RMVECs formed a dense capillary network inside the central channel with bifurcations reaching out to the intersectional channel after 12 days. After the addition of DKK3, the capillary network collapsed as judged by measuring the total length, percentage area occupied by capillary-like structures and number of bifurcations. Moreover, endothelial–mesenchymal transition was enhanced after the addition of DKK3 as judged by the intensity of α-SMA-GFP fluorescence. Collectively these findings confirm our results of the 2D cultures establishing DKK3 as an inducer of endothelial–mesenchymal/fibroblast–myofibroblast transition and as a suppressor of angiogenesis. FIGURE 8: View largeDownload slide DKK3 disrupts capillary network formation. (A) Representative images of cocultured RMVECs before and after exposure to DKK3. The central channel harboring RMVECs is shown on the right side of the image. The intersectional channel is shown on the left side of the image. RMVECs prelabeled with a CellTracker are shown in red. Green = α-SMA-GFP. Black dots indicate beads conjugated with CD31 antibody. RMVECs were cocultured with renal fibroblasts for 12 days (left panel). RMVECs formed a dense capillary network with branches toward the intersectional channel. On Day 12 of culture, DKK3 (10 µg/mL) was added to the endothelial culture medium. Forty-eight hours later the capillary network was abolished and the RMVECs underwent mesenchymal transition, as judged by the appearance of α-SMA-GFP, represented as the yellow areas created by colocalization of the red CellTracker and the GFP signal (n = 3; scale bar 150 µm). (B–E) Quantification of the α-SMA-GFP fluorescence intensity, the vessels percentage area, total number of junctions and total vessels length intensity on Day 14 (*P < 0.05). FIGURE 8: View largeDownload slide DKK3 disrupts capillary network formation. (A) Representative images of cocultured RMVECs before and after exposure to DKK3. The central channel harboring RMVECs is shown on the right side of the image. The intersectional channel is shown on the left side of the image. RMVECs prelabeled with a CellTracker are shown in red. Green = α-SMA-GFP. Black dots indicate beads conjugated with CD31 antibody. RMVECs were cocultured with renal fibroblasts for 12 days (left panel). RMVECs formed a dense capillary network with branches toward the intersectional channel. On Day 12 of culture, DKK3 (10 µg/mL) was added to the endothelial culture medium. Forty-eight hours later the capillary network was abolished and the RMVECs underwent mesenchymal transition, as judged by the appearance of α-SMA-GFP, represented as the yellow areas created by colocalization of the red CellTracker and the GFP signal (n = 3; scale bar 150 µm). (B–E) Quantification of the α-SMA-GFP fluorescence intensity, the vessels percentage area, total number of junctions and total vessels length intensity on Day 14 (*P < 0.05). DKK3-eGFP mice with adriamycin-induced nephropathy and sulindac sulfide treatment To facilitate monitoring of DKK3 expression in the kidney, we made use of reporter DKK3-eGFP × Col3.6-GFPcyan mice. Under basal conditions, DKK3 expression, as judged by GFP fluorescence, was minimal. Induction of adriamycin-induced nephropathy in these mice resulted in robust GFP signals emanating from the tubular epithelial cells, parietal epithelial cells and, to a lesser degree, the vascular endothelium (Figure 9). These findings further expand the list of potential sources of DKK3 in pathologic situations and suggest that the tubular and glomerular epithelia may contribute to the DKK3 burden of the endothelial secretome. These findings are in concert with the previous demonstration that the tubular DKK3 promotes the development of renal tubular atrophy and fibrosis [20]. FIGURE 9: View largeDownload slide Sulindac sulfide attenuates renal fibrosis. (A) Representative images of DKK3-GFP fluorescence intensity and sirius red staining after adriamycin-induced nephropathy with and without sulindac sulfide treatment (n = 4; scale bar for ×10, ×150 µm and for ×40, ×30 µm). (B–C) Quantification of the DKK3-GFP fluorescence intensity and the tissue fibrosis percentage area. Mice treated with sulindac sulfide showed suppressed DKK3-GFP fluorescence and reduced interstitial fibrosis in the adriamycin-induced nephropathy model (*P < 0.05 versus vehicle). FIGURE 9: View largeDownload slide Sulindac sulfide attenuates renal fibrosis. (A) Representative images of DKK3-GFP fluorescence intensity and sirius red staining after adriamycin-induced nephropathy with and without sulindac sulfide treatment (n = 4; scale bar for ×10, ×150 µm and for ×40, ×30 µm). (B–C) Quantification of the DKK3-GFP fluorescence intensity and the tissue fibrosis percentage area. Mice treated with sulindac sulfide showed suppressed DKK3-GFP fluorescence and reduced interstitial fibrosis in the adriamycin-induced nephropathy model (*P < 0.05 versus vehicle). Having screened a library of compounds with the ability to suppress the Wnt pathway, Bravi et al. [23] demonstrated that sulindac sulfide is the most efficient inhibitor. For these reasons we chronically treated the DKK3-eGFP × Col3.6-GFPcyan reporter mice subjected to adriamycin-induced nephropathy with sulindac sulfide. After 7 days of treatment the intensity of fluorescent signals decreased, while the extent of renal fibrosis was significantly diminished (Figure 9). Additional in vitro studies with RMVECs isolated from DKK3-eGFP × Col3.6-GFPcyan reporter mice showed that the DKK3-GFP expression increased significantly after exposure to TGF-β, and especially in combination with Sirt1-inhibition (Figure 10). Sirt1 inhibition alone did not significantly increase the DKK3 expression, which is in accordance with the mass spectrometry results. Treatment with sulindac sulfide reduced DKK3 expression in all the groups, further emphasizing its ability to suppress the Wnt pathway. In addition, renal fibroblasts isolated from α-SMA-GFP mice showed reductions in TCF-1 transcripts, a Wnt downstream target gene, after treatment with sulindac sulfide (Figure 10C). FIGURE 10: View largeDownload slide Sulindac sulfide reduces DKK3 expression in RMVECs. (A) Representative images of DKK3-GFP fluorescence intensity in RMVECs isolated from DKK3-eGFP × Col3.6-GFPcyan mice 48 h after treatment with the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of the DKK3-GFP fluorescence intensity. DKK3 expression increased after TGF-β treatment and especially in combination with Sirt1 inhibitor in RMVECs. Sulindac sulfide effectively reduced DKK3 expression (*P < 0.05; #, indicates the control group and Sirt1 inhibitor group to be significantly different from the TGF-β group and Sirt1 inhibitor + TGF-β group). (C) Quantitative polymerase chain reaction of TCF-1 transcripts in renal fibroblasts isolated from α-SMA-GFP mice (n = 4; *P < 0.05). FIGURE 10: View largeDownload slide Sulindac sulfide reduces DKK3 expression in RMVECs. (A) Representative images of DKK3-GFP fluorescence intensity in RMVECs isolated from DKK3-eGFP × Col3.6-GFPcyan mice 48 h after treatment with the indicated proteins (n = 4; scale bar 150 µm). (B) Quantification of the DKK3-GFP fluorescence intensity. DKK3 expression increased after TGF-β treatment and especially in combination with Sirt1 inhibitor in RMVECs. Sulindac sulfide effectively reduced DKK3 expression (*P < 0.05; #, indicates the control group and Sirt1 inhibitor group to be significantly different from the TGF-β group and Sirt1 inhibitor + TGF-β group). (C) Quantitative polymerase chain reaction of TCF-1 transcripts in renal fibroblasts isolated from α-SMA-GFP mice (n = 4; *P < 0.05). In order to investigate DKK3-GFP fluorescence in a model of endothelial dysfunction, we created Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice. Endothelial Sirt1 deficiency has been shown to cause the development of microvascular rarefaction, fibrosis and a decline in the endothelial glycocalyx, to name a few, already at a basal state [7, 10, 19]. As demonstrated in Figure 11A, wild-type 10-week-old mice with intact endothelial Sirt1 showed no DKK3-GFP fluorescence. In contrast, endothelial Sirt1 deficiency was associated with a significant increase in DKK3-GFP fluorescence. Costaining with antibodies against CD31 allowed us to colocalize DKK3 with the endothelial compartment, in addition to DKK3 expression by epithelial and interstitial cells (Figure 11B). These findings support the hypothesis that endothelial dysfunction causes an increase in DKK3. FIGURE 11: View largeDownload slide Endothelial Sirt1 deficiency increases DKK3-GFP fluorescence at a basal state. (A) Representative images of DKK3-GFP fluorescence intensity of Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice versus wild-type × DKK3-eGFP × Col3.6-GFPcyan mice at a basal state (n = 3; red = CD31, green = DKK3-GFP, blue = DAPI; scale bar for ×10, 150 µm and for ×40, 30 µm). (B) Representative images of DKK3-GFP fluorescence intensity of Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice at a basal state (n = 3; red = CD31, green = DKK3-GFP, blue = DAPI; scale bar 30 µm). (C) Quantification of the DKK3-GFP fluorescence intensity. Sirt1endo−/− mice revealed increased DKK3-GFP fluorescence at a basal state as compared with wild-type mice (*P < 0.05). FIGURE 11: View largeDownload slide Endothelial Sirt1 deficiency increases DKK3-GFP fluorescence at a basal state. (A) Representative images of DKK3-GFP fluorescence intensity of Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice versus wild-type × DKK3-eGFP × Col3.6-GFPcyan mice at a basal state (n = 3; red = CD31, green = DKK3-GFP, blue = DAPI; scale bar for ×10, 150 µm and for ×40, 30 µm). (B) Representative images of DKK3-GFP fluorescence intensity of Sirt1endo−/− × DKK3-eGFP × Col3.6-GFPcyan mice at a basal state (n = 3; red = CD31, green = DKK3-GFP, blue = DAPI; scale bar 30 µm). (C) Quantification of the DKK3-GFP fluorescence intensity. Sirt1endo−/− mice revealed increased DKK3-GFP fluorescence at a basal state as compared with wild-type mice (*P < 0.05). DISCUSSION A remarkable finding of proteomic screening of RMVEC secretomes consists of the detection of DKK3 exclusively in the profibrogenic cells derived from Sirt1endo−/− mice and stimulated with TGF-β. Of note, Yin et al. [6] detected 123 proteins in the secretome of nonstimulated human umbilical vein endothelial cells. Somewhat similar findings were reported in the secretome of human adipose stromal vascular fraction cells during adipogenesis [24]. The salient feature of our study consists of juxtaposing proteomic findings in the pro- and antifibrogenic secretomes. Any differences in their respective proteomes could potentially be responsible for contrasting phenotypic properties of microvascular endothelial cells. The Wnt pathway is critical for the maintenance of stem cells and tissue renewal [25]; however, chronic activation of this pathway may have detrimental effects. Persistent activation of the Wnt pathway has been implicated in the development of renal, pulmonary and liver fibrosis [26]. While DKK1 is a well-known inhibitor of canonical Wnt/β-catenin signaling, the functions of DKK3 are much less examined, its ability to modulate Wnt signaling is controversial [27] and it is not certain whether it interacts, as DKK1 does, with Frizzled and its coreceptor LRP5/6 [28]. One study stipulates that DKK3 is a prosurvival signal that positively modulates Wnt signaling [27]. According to another study, DKK3 acts as a cytokine and induces differentiation of stem cells toward smooth muscle lineage via activating transcription factor 6 signaling [29], as well as acting as a tumor suppressor [30] or as an inhibitor of VEGFR2/Akt/mTOR signaling [31]. Recently, tubular epithelial DKK3 has been found to promote renal fibrosis [20]. Similarly, overexpression of Wnt in tubular epithelia was found to induce sterile fibrosis [32]. The data presented herein implicate DKK3 in antagonizing inhibitory effects of DKK1 on canonical Wnt/β-catenin pathway–induced conversion of renal fibroblasts to myofibroblasts, especially when occurring on the background of activated TGF-β. The TGF-β contribution may be explained by the fact that TGF-β itself leads to rapid secretion of Wnt by fibroblasts, an observation that has been implicated in cardiac myofibroblast formation and myocardial fibrosis [33]. We employ α-SMA-GFP mice to isolate renal fibroblasts and microvascular endothelial cells to be used as readouts for cell acquisition of the activated mesenchymal phenotype, as was reported previously [34, 35]. As expected, TGF-β or Wnt1 activates fibroblasts to express the marker of myofibroblasts, α-SMA-GFP. Similarly, application of DKK3 also activates fibroblasts. When DKK3 is combined with TGF-β or Wnt1, the extent of fibroblast-to-myofibroblast conversion shows no additive effect. In contrast, when DKK3 is coapplied with DKK1 in the presence of TGFβ or Wnt1, it appears to counteract the antagonistic effect of DKK1, thus suggesting that its mode of action may consist in competition with DKK1 for binding to LRP5/6. Of note, our shotgun proteomic findings in RMVEC secretomes did not detect other components of these pathways, such as Wnt, DKK1, soluble Frizzled receptor proteins or Wnt inhibitory factor, which could, in turn, interfere with endogenous Wnt binding to the β-propeller domain of LRP5/6 [36]. Our finding that DKK3 is a natural antagonist of DKK1 may explain its profibrogenic action. Yet, there should be an additional mechanism of DKK3 action to induce myofibroblast formation from renal fibroblasts since its effects are detected even in the absence of DKK1 or TGF-β. Actions of DKK3 overexpressed in the aberrant secretome of dysfunctional endothelial cells are not limited to the activation of neighboring fibroblasts. DKK3 autocrine effects on RMVECs themselves consist of suppressed capillary cords formation on matrigel and halted angiogenesis in microfluidic devices while promoting endothelial–mesenchymal transition. These findings are consistent with the previous demonstration of DKK3 interfering with VEGFR2/Akt signaling [31]. The novel finding that DKK3 promotes endothelial–mesenchymal transition is of singular importance. This transformation is characterized by acquisition of the mesenchymal phenotype with the loss of vasodilatory, antithrombogenic, antiplatelet and leukocyte adhesion and enhanced synthesis of fibrogenic extracellular matrix proteins [4]. It is an essential pathogenic step in microvascular rarefaction or malfunction associated with tissue fibrosis [37–39], atherosclerosis [40], cavernous cerebral malformations [41], heterotopic ossification [42], transplant arteriopathy [43] and cancer [44]. Thus endothelial–mesenchymal transition has a broad footprint in the pathogenesis of diverse conditions. The schema presented in Figure 12 summarizes findings presented herein and supplemented with the published relevant data. Dysfunctional endothelial cells generate excessive amounts of DKK3, potentially enhancing Wnt/β-catenin signaling in the interstitial fibroblasts and possibly pericytes. DKK1 produced by the epithelial cells [32] counteracts the effects of DKK3, except when its concentration is reduced by the distance from its source. Federico et al. [20] demonstrated that DKK3 can be also produced by the stressed tubular epithelia, and our findings in DKK3-GFP mice are consistent with that observation. In addition, Zhou et al. [21] demonstrated secretion of tubule-derived Wnt as a prerequisite to fibroblast activation and fibrosis. These overlapping gradients of pro- and antifibrogenic signals play an important role in instructing fibroblasts to undergo conversion to myofibroblasts. FIGURE 12: View largeDownload slide Cross talk of different cell lines through DKK1/DKK3/Wnt/β-catenin in the kidney. Schematic presentation of the DKK1/DKK3/Wnt/β-catenin interactions between dysfunctional endothelial cells, interstitial fibroblasts, pericytes and epithelial cells. FIGURE 12: View largeDownload slide Cross talk of different cell lines through DKK1/DKK3/Wnt/β-catenin in the kidney. Schematic presentation of the DKK1/DKK3/Wnt/β-catenin interactions between dysfunctional endothelial cells, interstitial fibroblasts, pericytes and epithelial cells. The key conclusion of this study is that microenvironmental signals derived from dysfunctional endothelial cells instruct (myo)fibroblast fate determination. Endothelial cells defective in Sirt1 secrete DKK3 and coerce fibroblast-to-myofibroblast conversion. In conjunction with the previous findings on Wnt and DKK3 secretion from stressed renal tubular epithelial cells [20, 32, 45], it appears that renal microvascular endothelium acts as a ‘gatekeeper’ for Wnt agonists and antagonists reaching renal fibroblasts, as well as endothelial cells themselves. Collectively, these findings support our hypothesis on endothelial contribution to the ‘third pathway’ of fibrogenesis [18]. By intercepting persistent DKK3 signals, shown to be pathologic, using the Wnt pathway inhibitor sulindac sulfide, it becomes possible to achieve therapeutic effects by reducing fibrogenesis in a model of chronic kidney disease. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS The authors are grateful to Dr Yujiro Kida for his initial help in isolating renal microvascular endothelial cells and collecting their conditioned media. FUNDING These studies have been supported in part by the Dr. Werner Jackstädt Foundation (to M.L.), the Westchester Community Trust Foundation – Renal Research Fund and NIH grants DK54602, DK052783 and DK45462 (to M.S.G.) and R21 AR055750 (to D.W.R.). N.L.J. was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT & Future Planning (2015R1A2A1A09005662 and NRF-2016R1A4A1010796). AUTHORS’ CONTRIBUTIONS M.L., H.D., G.A.M. and M.S.G. designed the study. M.L., H.D. and S.D. carried out experiments. M.L., H.D., B.B.R. analyzed the data. M.L. and H.D. made the figures. M.L., H.D., G.A.M. and M.S.G. drafted and revised the manuscript. N.L.J. provided the microfluidic devices. D.W.R. provided the DKK3-eGFP × Col3.6-GFPcyan mice. All authors approved the final version of the manuscript. CONFLICT OF INTEREST STATEMENT None declared. REFERENCES 1 Butler JM , Kobayashi H , Rafii S. Instructive role of the vascular niche in promoting tumour growth and tissue repair by angiocrine factors . Nat Rev Cancer 2010 ; 10 : 138 – 146 Google Scholar CrossRef Search ADS PubMed 2 Hemmen K , Reinl T , Buttler K et al. High-resolution mass spectrometric analysis of the secretome from mouse lung endothelial progenitor cells . Angiogenesis 2011 ; 14 : 163 – 172 Google Scholar CrossRef Search ADS PubMed 3 Pula G , Mayr U , Evans C et al. Proteomics identifies thymidine phosphorylase as a key regulator of the angiogenic potential of colony-forming units and endothelial progenitor cell cultures . Circ Res 2009 ; 104 : 32 – 40 Google Scholar CrossRef Search ADS PubMed 4 Dejana E , Hirschi KK , Simons M. The molecular basis of endothelial cell plasticity . Nat Commun 2017 ; 8 : 14361 Google Scholar CrossRef Search ADS PubMed 5 Rafii S , Butler JM , Ding BS. Angiocrine functions of organ-specific endothelial cells . Nature 2011 ; 529 : 316 – 325 Google Scholar CrossRef Search ADS 6 Yin X , Bern M , Xing Q et al. Glycoproteomic analysis of the secretome of human endothelial cells . Mol Cell Proteomics 2013 ; 12 : 956 – 978 Google Scholar CrossRef Search ADS PubMed 7 Vasko R , Xavier S , Chen J et al. Endothelial sirtuin 1 deficiency perpetrates nephrosclerosis through downregulation of matrix metalloproteinase-14: relevance to fibrosis of vascular senescence . J Am Soc Nephrol 2014 ; 25 : 276 – 291 Google Scholar CrossRef Search ADS PubMed 8 Xavier S , Vasko R , Matsumoto K et al. Curtailing endothelial TGF-β signaling is sufficient to reduce endothelial-mesenchymal transition and fibrosis in CKD . J Am Soc Nephrol 2015 ; 26 : 817 – 829 Google Scholar CrossRef Search ADS PubMed 9 Potente M , Ghaeni L , Baldessari D et al. SIRT1 controls endothelial angiogenic functions during vascular growth . Genes Dev 2007 ; 21 : 2644 – 2658 Google Scholar CrossRef Search ADS PubMed 10 Chen J , Xavier S , Moskowitz-Kassai E et al. Cathepsin cleavage of sirtuin 1 in endothelial progenitor cells mediates stress-induced premature senescence . Am J Pathol 2012 ; 180 : 973 – 983 Google Scholar CrossRef Search ADS PubMed 11 Kida Y , Goligorsky MS. Sirtuins, cell senescence, and vascular aging . Can J Cardiol 2016 ; 32 : 634 – 641 Google Scholar CrossRef Search ADS PubMed 12 Ziyadeh FN , Hoffman BB , Han DC et al. Long-term prevention of renal insufficiency, excess matrix gene expression, and glomerular mesangial matrix expansion by treatment with monoclonal antitransforming growth factor-beta antibody in db/db diabetic mice . Proc Natl Acad Sci USA 2000 ; 97 : 8015 – 8020 Google Scholar CrossRef Search ADS PubMed 13 Chan LL , Huang J , Hagiwara Y et al. Discriminating multiplexed GFP reporters in primary articular chondrocyte cultures using image cytometry . J Fluoresc 2014 ; 24 : 1041 – 1053 Google Scholar CrossRef Search ADS PubMed 14 Mori Y , Adams D , Hagiwara Y et al. Identification of a progenitor cell population destined to form fracture fibrocartilage callus in Dickkopf-related protein 3-green fluorescent protein reporter mice . J Bone Miner Metab 2016 ; 34 : 606 – 614 Google Scholar CrossRef Search ADS PubMed 15 Yokota T , Kawakami Y , Nagai Y et al. Bone marrow lacks a transplantable progenitor for smooth muscle type alpha-actin-expressing cells . Stem Cells 2006 ; 24 : 13 – 22 Google Scholar CrossRef Search ADS PubMed 16 Kalajzic Z , Li H , Wang LP et al. Use of an alpha-smooth muscle actin GFP reporter to identify an osteoprogenitor population . Bone 2008 ; 43 : 501 – 510 Google Scholar CrossRef Search ADS PubMed 17 Kim S , Lee H , Chung M et al. Engineering of functional, perfusable 3D microvascular networks on a chip . Lab Chip 2013 ; 13 : 1489 – 1500 Google Scholar CrossRef Search ADS PubMed 18 Lipphardt M , Song JW , Matsumoto K et al. The third path of tubulointerstitial fibrosis: aberrant endothelial secretome . Kidney Int 2017 ; 92 : 558 – 568 Google Scholar CrossRef Search ADS PubMed 19 Lipphardt M , Song JW , Ratliff BB et al. Endothelial dysfunction is a super-inducer of syndecan 4: fibrogenic role of its ectodomain . Am J Physiol Heart Circ Physiol 2018 ; 314 : H484 – H496 Google Scholar CrossRef Search ADS PubMed 20 Federico G , Meister M , Mathow D et al. Tubular Dickkopf-3 promotes the development of renal atrophy and fibrosis . JCI Insight 2016 ; 1 : e84916 Google Scholar CrossRef Search ADS PubMed 21 Zhou D , Fu H , Zhang L et al. Tubule-derived wnts are required for fibroblast activation and kidney fibrosis . J Am Soc Nephrol 2017 ; 28 : 2322 – 2336 Google Scholar CrossRef Search ADS PubMed 22 Karamariti E , Margariti A , Winkler B et al. Smooth muscle cells differentiated from reprogrammed embryonic lung fibroblasts through DKK3 signaling are potent for tissue engineering of vascular grafts . Circ Res 2013 ; 112 : 1433 – 1443 Google Scholar CrossRef Search ADS PubMed 23 Bravi L , Rudini N , Cuttano R et al. Sulindac metabolites decrease cerebrovascular malformations in CCM3-knockout mice . Proc Natl Acad Sci USA 2015 ; 112 : 8421 – 8426 Google Scholar CrossRef Search ADS PubMed 24 Kim J , Choi YS , Lim S et al. Comparative analysis of the secretory proteome of human adipose stromal vascular fraction cells during adipogenesis . Proteomics 2010 ; 10 : 394 – 405 Google Scholar CrossRef Search ADS PubMed 25 Clevers H , Loh KM , Nusse R. Stem cell signaling. An integral program for tissue renewal and regeneration: Wnt signaling and stem cell control . Science 2014 ; 346 : 1248012 Google Scholar CrossRef Search ADS PubMed 26 Miao CG , Yang YY , He X et al. Wnt signaling in liver fibrosis: progress, challenges and potential directions . Biochimie 2013 ; 95 : 2326 – 2335 Google Scholar CrossRef Search ADS PubMed 27 Nakamura RE , Hackam AS. Analysis of Dickkopf3 interactions with Wnt signaling receptors . Growth Factors 2010 ; 28 : 232 – 242 Google Scholar CrossRef Search ADS PubMed 28 Krupnik VE , Sharp JD , Jiang C et al. Functional and structural diversity of the human Dickkopf gene family . Gene 1999 ; 238 : 301 – 313 Google Scholar CrossRef Search ADS PubMed 29 Wang X , Karamariti E , Simpson R et al. Dickkopf homolog 3 induces stem cell differentiation into smooth muscle lineage via ATF6 signalling . J Biol Chem 2015 ; 290 : 19844 – 19852 Google Scholar CrossRef Search ADS PubMed 30 Veeck J , Dahl E. Targeting the Wnt pathway in cancer: the emerging role of Dickkopf-3 . Biochim Biophys Acta 2012 ; 1825 : 18 – 28 Google Scholar PubMed 31 Kim BR , Lee EJ , Seo SH et al. Dickkopf-3 (DKK-3) obstructs VEGFR-2/Akt/mTOR signaling cascade by interacting of β2-microglobulin (β2M) in ovarian tumorigenesis . Cell Signal 2015 ; 27 : 2150 – 2159 Google Scholar CrossRef Search ADS PubMed 32 Maarouf OH , Aravamudhan A , Rangarajan D et al. Paracrine Wnt1 drives interstitial fibrosis without inflammation by tubulointerstitial cross-talk . J Am Soc Nephrol 2016 ; 27 : 781 – 790 Google Scholar CrossRef Search ADS PubMed 33 Blyszczuk P , Müller-Edenborn B , Valenta T et al. Transforming growth factor-β-dependent Wnt secretion controls myofibroblast formation and myocardial fibrosis progression in experimental autoimmune myocarditis . Eur Heart J 2017 ; 38 : 1413 – 1425 Google Scholar PubMed 34 Kida Y , Zullo JA , Goligorsky MS. Endothelial sirtuin 1 inactivation enhances capillary rarefaction and fibrosis following kidney injury through Notch activation . Biochem Biophys Res Commun 2016 ; 478 : 1074 – 1079 Google Scholar CrossRef Search ADS PubMed 35 Matsumoto K , Xavier S , Chen J et al. Instructive role of the microenvironment in preventing renal fibrosis . Stem Cells Transl Med 2017 ; 6 : 992 – 1005 Google Scholar CrossRef Search ADS PubMed 36 Bourhis E , Wang W , Tam C et al. Wnt antagonists bind through a short peptide to the first β-propeller domain of LRP5/6 . Structure 2011 ; 19 : 1433 – 1442 Google Scholar CrossRef Search ADS PubMed 37 Zeisberg EM , Tarnavski O , Zeisberg M et al. Endothelial-to-mesenchymal transition contributes to cardiac fibrosis . Nat Med 2007 ; 13 : 952 – 961 Google Scholar CrossRef Search ADS PubMed 38 O'Riordan E , Mendelev N , Patschan S et al. Chronic NOS inhibition actuates endothelial-mesenchymal transformation . Am J Physiol Heart Circ Physiol 2007 ; 292 : H285 – H294 Google Scholar CrossRef Search ADS PubMed 39 Piera-Velazquez S , Li Z , Jimenez SA. Role of endothelial-mesenchymal transition (EndoMT) in the pathogenesis of fibrotic disorders . Am J Pathol 2011 ; 179 : 1074 – 1080 Google Scholar CrossRef Search ADS PubMed 40 Chen PY , Qin L , Baeyens N et al. Endothelial-to-mesenchymal transition drives atherosclerosis progression . J Clin Invest 2015 ; 125 : 4514 – 4528 Google Scholar CrossRef Search ADS PubMed 41 Maddaluno L , Rudini N , Cuttano R et al. EndMT contributes to the onset and progression of cerebral cavernous malformations . Nature 2013 ; 498 : 492 – 496 Google Scholar CrossRef Search ADS PubMed 42 Medici D , Shore EM , Lounev VY et al. Conversion of vascular endothelial cells into multipotent stem-like cells . Nat Med 2010 ; 16 : 1400 – 1406 Google Scholar CrossRef Search ADS PubMed 43 Tellides G , Pober JS. Inflammatory and immune responses in the arterial media . Circ Res 2015 ; 116 : 312 – 322 Google Scholar CrossRef Search ADS PubMed 44 Zeisberg EM , Potenta S , Xie L et al. Discovery of endothelial to mesenchymal transition as a source for carcinoma-associated fibroblasts . Cancer Res 2007 ; 67 : 10123 – 10128 Google Scholar CrossRef Search ADS PubMed 45 Gröne EF , Federico G , Nelson PJ et al. The hormetic functions of Wnt pathways in tubular injury . Pflugers Arch 2017 ; 469 : 899 – 906 Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Nephrology Dialysis TransplantationOxford University Press

Published: May 3, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off