Syndecan-4 knockout leads to reduced extracellular transglutaminase-2 and protects against tubulointerstitial fibrosis

Abstract

Transglutaminase type 2 (TG2) is an extracellular matrix crosslinking enzyme with a pivotal role in kidney fibrosis. The interaction of TG2 with the heparan sulfate proteoglycan syndecan-4 (Sdc4) regulates the cell surface trafficking, localization, and activity of TG2 in vitro but remains unstudied in vivo. We tested the hypothesis that Sdc4 is required for cell surface targeting of TG2 and the development of kidney fibrosis in CKD. Wild-type and Sdc4-null mice were subjected to unilateral ureteric obstruction and aristolochic acid nephropathy (AAN) as experimental models of kidney fibrosis. Analysis of renal scarring by Masson trichrome staining, kidney hydroxyproline levels, and collagen immunofluorescence demonstrated progressive fibrosis associated with increases in extracellular TG2 and TG activity in the tubulointerstitium in both models. Knockout of Sdc-4 reduced these effects and prevented AAN-induced increases in total and active TGF-β1. In wild-type mice subjected to AAN, extracellular TG2 colocalized with Sdc4 in the tubular interstitium and basement membrane, where TG2 also colocalized with heparan sulfate chains. Heparitinase I, which selectively cleaves heparan sulfate, completely abolished extracellular TG2 in normal and diseased kidney sections. In conclusion, the lack of Sdc4 heparan sulfate chains in the kidneys of Sdc4-null mice abrogates injury-induced externalization of TG2, thereby preventing profibrotic crosslinking of extracellular matrix and recruitment of large latent TGF-β1. This finding suggests that targeting the TG2-Sdc4 interaction may provide a specific interventional strategy for the treatment of CKD.

Copyright © 2014 by the American Society of Nephrology.

Figures

Figure 1.

Sdc4-KO protects against the development of renal fibrosis in the UUO model of CKD. Paraffin sections from WT and Sdc4 KO kidneys (control and 21 days after UUO) were stained with MT (A–D), collagen I (E–H), collagen III (I–J), and collagen IV (M–P). Collagen staining (red) and nuclei staining (blue). Representative images at ×200 magnification are shown. Detail of collagen IV staining at ×400 magnification is shown (M).

Figure 2.

Sdc4-KO protects against the development of renal fibrosis in the AAN model of CKD. Paraffin sections from WT and Sdc4 KO kidneys (control and AAN at 12 weeks) were stained with MT (A–D), collagen I (E–H), collagen III (I–L), and collagen IV (M–P). Collagens staining (green) and nuclei staining (blue). Representative images at ×200 magnification are shown.

Figure 3.

Sdc4-KO prevents ECM protein increases in the UUO and AAN model of CKD. Graphical presentation of kidney fibrosis in the UUO (A–D) and AAN (E–H) fibrotic lesions. The fibrosis score was calculated by multiphase analysis of MT-stained kidney sections as the ratio blue/green (collagen)/pink/red (tissue area) (A and E). Deposition of collagen I (B and F), collagen III (C and G), and collagen IV (D and H) was measured by multiphase analysis of immunofluorescence-stained kidney sections (collagens/DAPI). All data were normalized by the WT control at the lower time point (day 7 for UUO, week 9 for AAN). Raw data at these time points were the following: MT UUO, 0.02; MT AAN, 0.001; collagen I UUO, 0.01; collagen I AAN, 0.03; collagen III UUO, 0.04; collagen III AAN, 0.06; collagen IV UUO, 0.03; collagen IV AAN, 0.08. *P<0.05 versus control; †P<0.05 versus Sdc4-KO UUO or AAN; aP<0.05 versus WT day 7 UUO or week 9 AAN; bP<0.05 versus Sdc4-KO day 7 UUO or week 9 AAN; cP<0.05 versus WT day 14 UUO; dP<0.05 versus Sdc-KO day 14 UUO.

Figure 4.

Sdc4-KO decreases extracellular TG2 and TG activity in the UUO model of CKD. Extracellular TG2 and TG in situ activity (ISA) were detected on cryostat sections of WT and Sdc4 KO kidneys from the UUO model through immunofluorescence (A and B). Representative images from control and UUO kidneys (×200 magnification) are shown. The levels of TG2 (C) and TG ISA (D) were quantified by multiphase analysis; TG2 was quantified by dividing the TG2 signal with DAPI, while for TG ISA by dividing incorporated Texas red cadaverine by tissue area (green autofluorescence). All data were normalized to the WT control at day 7 UUO; the original values were 0.098 for TG2 and 3.564 for TG ISA. *P<0.05 versus control; †P<0.05 versus Sdc4 KO UUO.

Figure 5.

Sdc4-KO decreases extracellular TG2 and TG activity in the AAN model of CKD. Extracellular TG2 was detected using cryostat sections of WT and Sdc4-KO kidneys from the AAN model through immunofluorescence (A). TG in situ activity was measured by incorporation of biointylated cadaverine and revealed by TexasRed-labeled streptavidin (B). Quantification of levels of TG2 (C) and TG in situ activity (D) were performed using multiphase image analysis by dividing the TG2 fluorescence or the incorporated cadaverine fluorescence by tissue area (green autofluorescence). All data were normalized by the WT control at week 9; the original values were 0.04 for TG2 and 0.03 for TG in situ activity. *P<0.05 compared to control; †P<0.05 compared to Sdc4-KO AAN; achanges from 9 and 12 weeks in the WT AAN groups.

Figure 6.

The expression of TG2 in kidney is similar in the WT and Sdc4-KO genotype. TG2 expression was evaluated by Western blot in control kidneys and fibrotic kidneys (total homogenates) from WT and Sdc4 KO mice. (A) AAN model. (B) UUO model. Cyclophilin A (Cyp A) was used as loading control. Twenty-five micrograms of kidney proteins was used, and 100 ng guinea pig liver TG2 was loaded as a control (M). TG2 protein level is expressed as mean±SEM of TG2/Cyp A (TG2 level); data are normalized for WT control. Differences in TG2 between the two genotypes and treatments were nonsignificant (P>0.05).

Figure 7.

Sdc4-KO reduces the cell membrane targeting of TG2. TG2 expression was quantified by Western blot in the AAN kidneys from WT and Sdc4 KO mice (A–D). Kidney homogenates were fractionated as described in the Supplemental Methods. Forty-micrograms of kidney proteins was loaded. Cyclophilin A (Cyp A) was used as loading control, and β-tubulin and Na+/K+ ATPase were used as cytosolic and membrane markers, respectively. TG2 protein level is expressed as mean±SEM of TG2/Cyp A (TG2 level); data are normalized for WT membrane fraction. *P<0.05. (D). Extracellular TG2 was detected in cryosections and intracellular TG2 was immunostained in paraffin sections by immunofluorescence. Details of representative images of extracellular TG2 staining (Supplemental Figure 3) and representative images of intracellular TG2 overlapped to phase contrast at day 21 after UUO are shown at ×400 and ×630 magnification, respectively (E). Scale, 50 μm.

Figure 8.

TG2 partially colocalizes with Sdc4 core protein in tubulointerstitial fibrotic lesions. Sdc4 and TG2 immunostaining of cryostat sections were performed by using rabbit polyclonal anti-Sdc4 and mouse monoclonal anti-TG2 IA12 antibody followed by, respectively, donkey antirabbit AlexaFluor 488 and goat antimouse DyLight 594. Representative confocal images of Sdc4, TG2, and DAPI stained sections are shown separately and merged (with and without DAPI staining) for control and AAI-treated kidneys (two fibrotic lesions are shown, field 1 and field 2). Basolateral membrane localization of Sdc4 and TG2 after AAI is shown at higher magnification (details a, b, and c). Dual staining controls were carried out in kidney sections from TG2 KO and Sdc4 KO mice (Supplemental Figure 3). Scale bars are shown under each column of images.

Figure 9.

TG2 largely colocalizes with HS chains in tubulointerstitial fibrotic lesions. HS and TG2 immunostainings of cryostat sections were performed using mouse monoclonal anti-HS antibody and rabbit polyclonal anti-TG2 antibody, respectively, followed by goat anti-mouse (IgM) FITC and donkey anti-rabbit IgG AlexaFluor 568. Representative pictures of HS, TG2, and DAPI-stained sections are shown separately and merged (with and without DAPI staining) for control and AAI-treated kidneys (two fibrotic lesions are shown). Basolateral membrane and interstitial localization of HS and TG2 after AAI are shown at higher magnification (details a, b, and c). Dual staining controls were carried out in TG2 KO kidney sections (Supplemental Figure 3). Scale bars are shown under each column of images.

Figure 10.

TG2 extracellular location depends on the HS chains of HSPGs. Kidney cryosections were treated with 50 mU/ml protease-free heparitinase I (Hep-I) for 2 hours at 37°C. Extracellular TG2 was immunolabeled using mouse anti-TG2 IA12 antibody followed by goat anti-mouse DyLight 594. Images were obtained by confocal microscopy (A) and the level of extracellular TG2 was quantified by ImageJ intensity analysis in four kidneys per treatment (B). Three fields are shown per treatment, representative of control and AAN fibrotic kidneys (A).

Figure 11.

Sdc4-KO lowers TGF-β1 in the AAN model of CKD. Active TGF-β (A), total TGF-β (B), and percentage of activated TGF-β (C) were evaluated in WT and Sdc4-KO kidneys using the mink lung TGF-β bioassay, as described in the Concise Methods. Total TGF-β was converted to a biologically active form for analysis by acid activation. The percentage of activation was calculated by expressing the level of active TGF-β as a percentage of total TGF-β. (D) Recombinant TGF-β standard curve. Values are the mean of four kidneys±SEM, each assessed in triplicate. RLU, relative light unit. *P<0.05; **P<0.01; ***P<0.001 compared with the control or compared with Sdc4 KO AAN.

Figure 12.

Interplay between TG2 and Sdc4/HS in TG2 externalization and matrix crosslinking during the development of fibrosis. Transmembrane Sdc4 traps TG2 at the cell surface through its HS chains, facilitating TG2 externalization. It is unclear how TG2 crosses the plasma membrane. The HS chains of secreted HSPG could facilitate diffusion of TG2 via adjacent binding sites, thus allowing TG2 to “slide” in the matrix (a). Engagement with protein substrates (collagen, fibronectin) leads to activation of TG2 transamidation, resulting in matrix stabilization by crosslinking (b). Furthermore, TG2 promotes TGF-β1 large latent complex (LLC) deposition into the ECM by covalently linking the latent TGF-β1–binding protein (LTBP) to matrix components (c),,, leading to its activation in the fibrotic kidney., TGF-β1 and the latency-associated peptide (LAP) are proteolytically separated at the site indicated by the arrowhead. Sdc4 could trap LLC via HS chains and/or contribute to TGF-β1 activation directly or indirectly through other Sdc4-linked pathways (d).