Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Mar 6;12(3):e11021.
doi: 10.15252/emmm.201911021. Epub 2020 Jan 14.

Dysregulated Mesenchymal PDGFR-β Drives Kidney Fibrosis

Affiliations
Free PMC article

Dysregulated Mesenchymal PDGFR-β Drives Kidney Fibrosis

Eva M Buhl et al. EMBO Mol Med. .
Free PMC article

Abstract

Kidney fibrosis is characterized by expansion and activation of platelet-derived growth factor receptor-β (PDGFR-β)-positive mesenchymal cells. To study the consequences of PDGFR-β activation, we developed a model of primary renal fibrosis using transgenic mice with PDGFR-β activation specifically in renal mesenchymal cells, driving their pathological proliferation and phenotypic switch toward myofibroblasts. This resulted in progressive mesangioproliferative glomerulonephritis, mesangial sclerosis, and interstitial fibrosis with progressive anemia due to loss of erythropoietin production by fibroblasts. Fibrosis induced secondary tubular epithelial injury at later stages, coinciding with microinflammation, and aggravated the progression of hypertensive and obstructive nephropathy. Inhibition of PDGFR activation reversed fibrosis more effectively in the tubulointerstitium compared to glomeruli. Gene expression signatures in mice with PDGFR-β activation resembled those found in patients. In conclusion, PDGFR-β activation alone is sufficient to induce progressive renal fibrosis and failure, mimicking key aspects of chronic kidney disease in humans. Our data provide direct proof that fibrosis per se can drive chronic organ damage and establish a model of primary fibrosis allowing specific studies targeting fibrosis progression and regression.

Keywords: PDGFR; anemia; chronic kidney disease; fibroblasts; progression.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Figures

Figure 1
Figure 1. PDGFR‐β expression is upregulated in fibrotic human and murine kidneys

Immunoblot detection and its quantification of PDGFR‐β in kidney cortex lysates of healthy and diseased human kidneys show twice as much PDGFR‐β abundance in diseased tissue. A stronger signal for phosphorylated PDGFR‐β indicates the active status of the receptor. The high α‐SMA content in the diseased kidneys shows that they are affected by fibrosis. Bar graphs show means ± SD, healthy n = 5, diseased n = 6. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05. Exact P‐values are provided in Appendix Table S4.

PDGFR‐β staining is detectable in mesangium of healthy (a) and diseased (b–d) glomeruli of human kidneys. PDGFR‐β‐positive mesangial cells expand in mesangioproliferating and scarring glomerular changes. Similarly, the population of PDGFR‐β‐positive interstitial cells is discrete in healthy kidney cortex (e) but expands in fibrosis (f‐h). Arrows point to selected areas of positive PDGFR‐β staining. Scale bar = 50 µm.

GFP expression under Pdgfrb promoter in mice is present in mesangial cells of glomeruli and interstitial cells of cortex and medulla, but not in tubular cells. Nuclei are stained with DAPI (blue). Scale bar = 50 µm.

Tissue clearing with Scale and 3D reconstruction of Pdgfrb‐GFP reporter mice in a healthy and fibrotic kidney (UUO day 5) shows the expansion of Pdgfrb‐expressing cells in fibrosis.

Source data are available online for this figure.
Figure 2
Figure 2. PDGFR‐β activation leads to mesenchymal proliferation in vivo and in vitro

Activation of PDGFR‐β in renal mesenchymal cells, denoted here as Foxd1Cre::Pdgfrb +/J mice, was achieved by crossbreeding the Foxd1‐Cre mouse line (FoxD1 GC) with a heterozygous knock‐in mouse line with constitutively active Pdgfrb mutant (J) allele (Pdgfrb (S)J) instead of the wt Pdgfrb allele.

Representative Ki67 immunofluorescence staining (green) in wt and Foxd1Cre::Pdgfrb +/J mice, showing increased proliferation in the transgenic mice. Glomeruli are outlined with circles, and arrowheads point to Ki67‐positive interstitial cells. Nuclei are stained with DAPI (blue). Scale bar = 50 µm.

Quantification of proliferating Ki67‐positive cells specifically in glomeruli (C), interstitium (D), and tubules (E) in Foxd1Cre::Pdgfrb +/J mice (black bars) and wt mice (white bars) 6, 14, 25, and 35 weeks of age. Foxd1Cre::Pdgfrb +/J mice exhibited increased proliferation of glomerular and interstitial cells, whereas tubular epithelial cell proliferation was not altered. Data in (C–E) are shown as means ± SD of n = 3 animals per group. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05 compared to wt of the same time point. Exact P‐values are provided in Appendix Table S4.

FoxD1 reporter mice (Foxd1::tdTomato) confirmed the expansion of mesenchymal cells in 14‐week‐old Foxd1Cre::Pdgfrb +/J mice in both glomeruli and interstitium. Circles outline glomeruli. Scale bar = 50 µm.

Quantification of FoxD1‐Tomato‐positive cells in the cortical interstitium in 14‐week‐old Foxd1Cre::Pdgfrb +/J and wt mice confirmed the significantly increased expansion of mesenchymal cells by 42% in Foxd1Cre::Pdgfrb +/J mice. Cells were counted in six view fields at 40× magnification. Bar graphs show means ± SD of n = 3 mice per group. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05 compared to wt of the same time point. Exact P‐values are provided in Appendix Table S4.

Isolated primary fibroblasts and mesangial cells from Foxd1Cre::Pdgfrb +/J mice have higher proliferation rates assessed by bromodeoxyuridine (BrdU) incorporation assay compared to cells from wt mice. Bar graphs show means ± SD, n = 4 per group, Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05. Exact P‐values are provided in Appendix Table S4.

Source data are available online for this figure.
Figure EV1
Figure EV1. Constitutive PDGFR‐β activation in renal mesenchyme in transgenic Foxd1Cre::Pdgfrb +/J mice

Genotyping PCR in Foxd1Cre::Pdgfrb +/J mice shows a wild‐type band at 346 bp and the mutated Pdgfrb p.V536A (J) band with at 160 bp.

Foxd1‐driven Cre recombinase activity is proven by PCR detection with primers detecting the Pdgfrb p.V546A allele with cutout STOP sequence. The corresponding band has a size of 200 bp.

(C) Western blots of PDGFR‐β and its downstream signaling molecules in 25‐week‐old wt and Foxd1Cre::Pdgfrb +/J mice and their respective densitometric evaluations (D) show regulations on the level of total protein and on phosphorylation status for PDGFR‐β and downstream molecules Akt and p38. Phospho‐levels are normalized to the respective unphosphorylated protein forms. kD = kilodalton. Bar graphs represent means ± SD; wt n = 8, Foxd1Cre::Pdgfrb +/J n = 7. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05. Exact P‐values are provided in Appendix Table S4.

Figure 3
Figure 3. Activation of PDGFR‐β in mesenchymal cells in Foxd1Cre::Pdgfrb +/J mice results in progressive mesangioproliferative glomerulonephritis and mesangial glomerulosclerosis

PAS staining of glomeruli of 25‐week‐old wt and of 6‐ and 35‐week‐old Foxd1Cre::Pdgfrb +/J mice. Glomeruli of Foxd1Cre::Pdgfrb+/J mice show pathological changes.

Glomerular tuft size of Foxd1Cre::Pdgfrb +/J mice was significantly increased compared to wt mice during the time course.

(C) Immunohistological staining of α‐SMA and its histomorphometric quantification (D) reveal expansion of activated mesangial cells in the glomerular tuft of the transgenic mice in the time course.

(E) Immunohistological staining of collagen I and its histomorphometric quantification (F) show that collagen I deposits in an increasing manner in Foxd1Cre::Pdgfrb +/J mice during the time course.

Data information: Scale bars = 50 µm. Bar graphs show mean ± SD of n = 3 animals per group. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05 compared to wt of the same time point. Exact P‐values are provided in Appendix Table S4.Source data are available online for this figure.
Figure 4
Figure 4. Activation of PDGFR‐β in mesenchymal cells in Foxd1Cre::Pdgfrb +/J mice results in progressive interstitial fibrosis, inflammation, and tubular damage

(A) Histological stainings and its histomorphometric quantification (B) of kidney cortex for α‐SMA of wt (25 weeks) and Foxd1Cre::Pdgfrb +/J mice 6 and 35 weeks of age show expansion of myofibroblasts in the interstitium of the transgenic mice.

(C) Collagen I staining and its histomorphometric quantification (D) show the progression of interstitial fibrosis in the time course.

The tubular stress marker lipocalin‐2 (LCN2) is not expressed in early time points but increases in the time course in late stages when fibrosis is more prominent, as seen on histological staining (E) and mRNA level (F).

A slight, but significant, increase in interstitial ErHr3‐positive stained macrophages was observed, particularly at later stages (G). For evaluation, ErHr3‐positive cells were counted on 20 view fields on 40× magnification (H).

Data information: Scale bars = 50 µm. Bar graphs show means ± SD of n = 3 animals per group. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05 compared to wt of the same time point. Exact P‐values are provided in Appendix Table S4.Source data are available online for this figure.
Figure 5
Figure 5. Reduced kidney function in Foxd1Cre::Pdgfrb +/J mice

Creatinine clearance was measured in mice 10, 14, 25, and 35 weeks of age and decreased continuously in Foxd1Cre::Pdgfrb +/J mice, with a predicted complete loss of renal function 62 weeks in transgenic mice using linear regression analyses as described previously (Steiger et al, 2018). Ten weeks n = 5 wt and n = 4 Foxd1Cre::Pdgfrb +/J, 14 weeks n = 2 per group, 25 weeks n = 8 wt and n = 8 Foxd1Cre::Pdgfrb +/J, 35 weeks n = 3 per group.

The blood urea nitrogen (BUN) concentrations were higher in Foxd1Cre::Pdgfrb +/J mice compared to wt mice, apart from the 25‐week‐old group. Bar graphs show means ± SD; 10 weeks n = 7 wt and n = 5 Foxd1Cre::Pdgfrb +/J, 14 weeks n = 3 per group, 25 weeks n = 11 wt and n = 12 Foxd1Cre::Pdgfrb +/J, 35 weeks n = 3 per group. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05 compared to wt of the same time point. Exact P‐values are provided in Appendix Table S4.

Figure 6
Figure 6. The Foxd1Cre::Pdgfrb +/J mice develop progressive renal anemia and are unable to regulate erythropoiesis under pathological stimuli

Erythrocyte numbers (A), hematocrit (B), and hemoglobin content (C) decrease in Foxd1Cre::Pdgfrb +/J mice continuously. (D) EPO levels shown as delta from respective wt littermates. Horizontal lines are means ± SD of n = 3 per group. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05 compared to wt of the same time point. Exact P‐values are provided in Appendix Table S4.

RNA in situ hybridization (RNAscope) of Epo (green) and Pdgfrb (purple) in renal kidney cortex of 35‐week‐old mice shows that EPO‐producing cells co‐express Pdgfrb. Foxd1Cre::Pdgfrb +/J mice have less EPO‐producing cells (arrows) than wt mice. Scale bar = 25 µm.

After induction of anemia by blood taking, the erythrocyte numbers recovered after 1 week in wt mice, whereas Foxd1Cre::Pdgfrb +/J mice were not able to recover their erythrocytes numbers at all. Data are means ± SD of n = 5 per group. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05 compared to wt of the same time point. Exact P‐values are provided in Appendix Table S4.

Source data are available online for this figure.
Figure EV2
Figure EV2. Blood pressure and renin production are not changed by mesenchymal PDGFR‐β activation

The blood pressure of Foxd1Cre::Pdgfrb +/J and wt mice remains normal during the whole time course. Data represent means ± SD of n = 3 animals.

The population of renin‐expressing cells is also similar between the groups, as shown using the renin staining. Scale bar = 50 µm.

Figure 7
Figure 7. PDGFR‐β activation aggravates the course of renal disease models, whereas blocking partially reverses the fibrotic phenotype

Five days after the induction of unilateral ureteral obstruction (UUO), a model of obstructive nephropathy and interstitial fibrosis, we assessed fibrosis (collagen I, α‐SMA) and inflammation (F4/80), the classical readout parameters in this model. Compared to wt mice with UUO, Foxd1Cre::Pdgfrb +/J mice developed significantly stronger fibrosis and inflammation (A). Representative pictures of the histological stainings are shown in (B). Scale bar = 50 µm. All data in (A) and (B) show means ± SD; wt n = 4, Foxd1Cre::Pdgfrb +/J n = 3. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05. Exact P‐values are provided in Appendix Table S4.

Angiotensin II infusion via an osmotic pump induced hypertension similarly in wt and Foxd1Cre::Pdgfrb +/J mice (C). In this model of hypertensive injury, the kidney function and even the histopathology were largely unaffected in wt mice, whereas the Foxd1Cre::Pdgfrb +/J mice showed significantly increased BUN values and prominent glomerulopathy after 28 days compared to the starting values (day 0), to wt animals after 28 days, and to age‐matched Foxd1Cre::Pdgfrb +/J mice without angiotensin II infusion (D). Consistently, the creatinine clearance per g bodyweight decreased in the hypertensive Foxd1Cre::Pdgfrb +/J mice (E). (F) PAS staining showed more prominent mesangial expansion and intratubular protein casts indicative of proteinuria only in hypertensive Foxd1Cre::Pdgfrb +/J mice. (G) Hypertension did not induce mesangial expansion and activation in wt mice, but led to a significant increase in hypertensive Foxd1Cre::Pdgfrb +/J mice as visualized by α‐SMA staining. (H) Histomorphometric analysis of α‐SMA‐positive percentage of the glomerular tuft area. All data in (C–H) show means ± SD, wt n = 7, Foxd1Cre::Pdgfrb +/J n = 5. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P ≤ 0.05 compared to wt of the same time point, § P ≤ 0.05 compared to d0, # P ≤ 0.05 compared to same‐aged non‐hypertensive mice. Exact P‐values are provided in Appendix Table S4.

Twenty‐two‐week‐old Foxd1Cre::Pdgfrb +/J mice were treated with imatinib (daily gavage, 50 mg/kg bodyweight for 21 days; Foxd1Cre::Pdgfrb +/J + imatinib) and were compared to control Foxd1Cre::Pdgfrb +/J mice receiving vehicle (water; Foxd1Cre::Pdgfrb +/J + water). Age‐matched wt mice are indicated as dashed lines. (I) Histomorphometric quantifications showed a significant reduction of α‐SMA abundance in the glomerular tufts of imatinib‐treated Foxd1Cre::Pdgfrb +/J mice. (J) Glomerular cellularity evaluated by counting the total number of cells per glomerular tuft, normalized to the area, was significantly reduced in imatinib‐treated Foxd1Cre::Pdgfrb +/J mice compared to control mice. (K, L) Histomorphometric quantifications of collagen I showed no effect of imatinib treatment on mesangial sclerosis in the glomeruli (K), whereas reduced interstitial fibrosis in imatinib‐treated Foxd1Cre::Pdgfrb +/J mice compared to controls was found (L). (M) Hemoglobin serum levels were significantly improved in imatinib‐treated mice. (N) Representative pictures of the histological stainings of α‐SMA (left) and collagen I (middle: glomeruli; right: cortical interstitium). Scale bar = 50 µm. (O, P) Total PDGFR‐β protein levels were decreased in Foxd1Cre::Pdgfrb +/J mice after imatinib treatment, as shown using Western blot (O) and its densitometric evaluation (P). Bar graphs show means ± SD, n = 5 per group. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P ≤ 0.05 Foxd1Cre::Pdgfrb +/J + water versus Foxd1Cre::Pdgfrb +/J + imatinib.

Source data are available online for this figure.
Figure 8
Figure 8. Foxd1Cre::Pdgfrb +/J mice show similar gene expression profiles like human patients with diabetic or hypertensive nephropathy

Gene expression arrays were performed in 6‐week‐old Foxd1Cre::Pdgfrb +/J mice compared to wt mice. Listed are genes with log2 fold changes higher or lower than 0.6. Out of these 46 genes, 17 are IFN‐regulated genes, five are involved in autophagy, and five are involved in matrix remodeling and fibrosis. The same genes were analyzed in microdissected glomeruli (G) and tubulointerstitium (T) of patients with diabetic (DN, G: n = 14, T: n = 18) and hypertensive (HT, G: n = 15, T: n = 21) nephropathy and living donors (LD, G: n = 42, T: n = 42) as controls, revealing a similar expression profile.

Heatmap depicting pathway activities in Foxd1Cre::Pdgfrb +/J mice and human CKD patients according to PROGENy.

Enrichment analyses of nodes in the inferred causal regulatory signaling networks of Foxd1Cre::Pdgfrb +/J mice and human CKD patients using the Biocarta database.

Based on the gene array data sets, transcription factor activities were estimated by using DoRothEA. Depicted are the most regulated ones regulated both in Foxd1Cre::Pdgfrb +/J mice and in human CKD patients.

Source data are available online for this figure.
Figure EV3
Figure EV3. Single‐cell RNA sequome in fibrotic murine and human kidney shows mesenchymal allocation of genes regulated in Foxd1Cre::Pdgfrb +/J mice

Genes regulated in our array in Foxd1Cre::Pdgfrb +/J mice were reanalyzed in a single‐cell RNA sequome data set of a fibrotic murine kidney subjected to UUO for 14 days. Shown are the genes that could be allocated to mesenchymal cell populations (activated fibroblasts = Act Fib1 and Act Fib2; or juxtaglomerular cells/renin cells = JGA). Pod = podocytes; EC = endothelial cells; PT(S1‐3) = proximal tubules; Dediff. PT = dedifferentiated proximal tubular cells; Prolif. PT = proliferating proximal tubular cells. DL and tAL = descending loop of henle and thin ascending loop; TAL = thick ascending loop; DCT = distal convoluted tubules; CNT = connecting tubules; CD‐PC = principal cells of collecting duct; IC = intercalated cells of collecting duct; Mø = macrophages.

The same expression comparison was done in a data set of a human renal graft rejected kidney. LOH (DL) = descending loop of henle; LOH (AL) = ascending loop of henle; CD = collecting duct; Mono1 and 2 = monocytes.

Figure EV4
Figure EV4. Glomerular pathology in Sox2Cre::Pdgfrb +/K mice resembled the findings in Foxd1Cre::Pdgfrb +/J but suggested no role for STAT1
Glomeruli stained for PAS and immunohistological stainings for collagen III and α‐SMA of 3‐week‐old Sox2Cre::Pdgfrb +/K mice with intact STAT1 signaling (Stat1 +/−) and with deletion of STAT1 signaling (Stat1−/−). At this very early stage, hardly any sclerosis was found on PAS or collagen III staining; however, the mesangial expression of α‐SMA was reduced in mice lacking STAT1. Sox2Cre::Pdgfrb +/K mice with intact STAT1 signaling (Stat1 +/−) were not viable at 9 weeks and could not be analyzed. However, the Sox2Cre::Pdgfrb +/K mice lacking STAT1 signaling survived up till later age and showed a progression of mesangioproliferative glomerulonephritis and also mesangial sclerosis, resembling closely the findings in Foxd1Cre::Pdgfrb +/J. Mice without activating Pdgfrb mutation and lacking STAT1 (Sox2Cre::Pdgfrb +/+ ::Stat1 −/−) do not show any pathological phenotype. Pictures are chosen as representatives from n = 3 each group. Scale bar = 50 µm.Source data are available online for this figure.
Figure EV5
Figure EV5. Developmental knockout of Pdgfrb in renal mesenchyme leads to defective glomerulogenesis due to lack of mesangium and resulting pathological frailty of glomeruli

Knockout of Pdgfrb in renal mesenchyme was achieved by crossbreeding the Foxd1‐Cre mouse line (FoxD1GC) with a mouse line with a homozygous floxed Pdgfrb gene (129S4/SvJae‐Pdgfrbtm11Sor/J), resulting in Foxd1Cre::Pdgfrb fl/fl mice.

Immunohistochemical staining for PDGFR‐β shows positivity in mesangial cells of wild‐type (wt) mice but not of Foxd1Cre::Pdgfrb fl/fl mice.

Six‐week‐old wt mice (a–c) have normal developed kidney cortex with fully developed glomeruli in outer cortex (a) as well as in the inner cortex (b). Foxd1Cre::Pdgfrb fl/fl mice (d–h) exhibit severe pathological changes in the kidney with aberrant glomeruli, dilated tubules, and interstitial fibrosis. The glomeruli of the outer cortex, i.e., the younger glomeruli (e, f), lack mesangium (arrows point to expected mesangial areas), whereas the ones of the inner cortex, i.e., the older glomeruli, show focal (g) or global (h) sclerosis (*) and also segmental rupture of capillaries with extracellular proliferates (crescents) (arrowheads).

Histological staining of mesangial cell marker NG2 in 1‐week‐old mice shows that in Foxd1Cre::Pdgfrb fl/fl mice, only in the glomeruli of the inner cortex a mesangial tree developed, whereas it is lacking in the glomeruli of the outer cortex, leading to undeveloped glomeruli. Arrowheads point to positive stained mesangial cells.

(E) Decreased creatinine clearance per gram bodyweight and (F) elevated serum creatinine levels in Foxd1Cre::Pdgfrb fl/fl mice reflected renal insufficiency in these mice.

Kidney dysfunction in Foxd1Cre::Pdgfrb fl/fl mice results in hypertension.

Data information: Bar graphs show mean ± SD; wt n = 11, Foxd1Cre::Pdgfrb fl/fl n = 9. Statistical analysis was performed by unpaired two‐tailed Student's t‐test. *P < 0.05. Exact P‐values are provided in Appendix Table S4.Source data are available online for this figure.

Similar articles

See all similar articles

Cited by 1 article

  • PDGFR-β and kidney fibrosis.
    Ortiz A. Ortiz A. EMBO Mol Med. 2020 Mar 6;12(3):e11729. doi: 10.15252/emmm.201911729. Epub 2020 Feb 18. EMBO Mol Med. 2020. PMID: 32072759 Free PMC article.

References

    1. Alpers CE, Seifert RA, Hudkins KL, Johnson RJ, Bowen‐Pope DF (1992) Developmental patterns of PDGF B‐chain, PDGF‐receptor, and alpha‐actin expression in human glomerulogenesis. Kidney Int 42: 390–399 - PubMed
    1. Andrae J, Gallini R, Betsholtz C (2008) Role of platelet‐derived growth factors in physiology and medicine. Genes Dev 22: 1276–1312 - PMC - PubMed
    1. Arar M, Xu YC, Elshihabi I, Barnes JL, Choudhury GG, Abboud HE (2000) Platelet‐derived growth factor receptor beta regulates migration and DNA synthesis in metanephric mesenchymal cells. J Biol Chem 275: 9527–9533 - PubMed
    1. Arts FA, Chand D, Pecquet C, Velghe AI, Constantinescu S, Hallberg B, Demoulin JB (2016) PDGFRB mutants found in patients with familial infantile myofibromatosis or overgrowth syndrome are oncogenic and sensitive to imatinib. Oncogene 35: 3239–3248 - PubMed
    1. Asada N, Takase M, Nakamura J, Oguchi A, Asada M, Suzuki N, Yamamura K, Nagoshi N, Shibata S, Rao TN et al (2011) Dysfunction of fibroblasts of extrarenal origin underlies renal fibrosis and renal anemia in mice. J Clin Investig 121: 3981–3990 - PMC - PubMed

Associated data

LinkOut - more resources

Feedback