A novel role of kallikrein-related peptidase 8 in the pathogenesis of diabetic cardiac fibrosis

Abstract

Rationale: Among all the diabetic complications, diabetic cardiomyopathy, which is characterized by myocyte loss and myocardial fibrosis, is the leading cause of mortality and morbidity in diabetic patients. Tissue kallikrein-related peptidases (KLKs) are secreted serine proteases, that have distinct and overlapping roles in the pathogenesis of cardiovascular diseases. However, whether KLKs are involved in the development of diabetic cardiomyopathy remains unknown.The present study aimed to determine the role of a specific KLK in the initiation of endothelial-to-mesenchymal transition (EndMT) during the pathogenesis of diabetic cardiomyopathy. Methods and Results-By screening gene expression profiles of KLKs, it was found that KLK8 was highly induced in the myocardium of mice with streptozotocin-induced diabetes. KLK8 deficiency attenuated diabetic cardiac fibrosis, and rescued the impaired cardiac function in diabetic mice. Small interfering RNA (siRNA)-mediated KLK8 knockdown significantly attenuated high glucose-induced endothelial damage and EndMT in human coronary artery endothelial cells (HCAECs). Diabetes-induced endothelial injury and cardiac EndMT were significantly alleviated in KLK8-deficient mice. In addition, transgenic overexpression of KLK8 led to interstitial and perivascular cardiac fibrosis, endothelial injury and EndMT in the heart. Adenovirus-mediated overexpression of KLK8 (Ad-KLK8) resulted in increases in endothelial cell damage, permeability and transforming growth factor (TGF)-β1 release in HCAECs. KLK8 overexpression also induced EndMT in HCAECs, which was alleviated by a TGF-β1-neutralizing antibody. A specificity protein-1 (Sp-1) consensus site was identified in the human KLK8 promoter and was found to mediate the high glucose-induced KLK8 expression. Mechanistically, it was identified that the vascular endothelial (VE)-cadherin/plakoglobin complex may associate with KLK8 in HCAECs. KLK8 cleaved the VE-cadherin extracellular domain, thus promoting plakoglobin nuclear translocation. Plakoglobin was required for KLK8-induced EndMT by cooperating with p53. KLK8 overexpression led to plakoglobin-dependent association of p53 with hypoxia inducible factor (HIF)-1α, which further enhanced the transactivation effect of HIF-1α on the TGF-β1 promoter. KLK8 also induced the binding of p53 with Smad3, subsequently promoting pro-EndMT reprogramming via the TGF-β1/Smad signaling pathway in HCAECs. The in vitro and in vivo findings further demonstrated that high glucose may promote plakoglobin-dependent cooperation of p53 with HIF-1α and Smad3, subsequently increasing the expression of TGF-β1 and the pro-EndMT target genes of the TGF-β1/Smad signaling pathway in a KLK8-dependent manner. Conclusions: The present findings uncovered a novel pro-EndMT mechanism during the pathogenesis of diabetic cardiac fibrosis via the upregulation of KLK8, and may contribute to the development of future KLK8-based therapeutic strategies for diabetic cardiomyopathy.

Keywords: KLK8; cardiac fibrosis; diabetic cardiomyopathy; endothelial-to-mesenchymal transition; plakoglobin.

Figure 1

KLK (kallikrein-related peptidase) 8 expression is significantly increased in diabetic myocardium. A, The mRNA levels of KLK family members in the hearts after 12 and 24 weeks of diabetes. B-E, Immunohistochemistry (IHC) staining (B) and quantification (C) showed increased level of KLK8 in both cardiomyocytes and endothelial cells and masson's trichrome staining (D) and quantification (E) showed increased collagen deposition in both interstitial and perivascular regions in the hearts after 24 weeks of diabetes as compared to control group (scale bar = 50 µm). F, Immunoblots of KLK8 in the hearts after 12 and 24 weeks of diabetes. Data are expressed as means ± SEM (n = 8). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 2

KLK (kallikrein-related peptidase) 8 deficiency attenuates diabetic cardiac fibrosis. A-E, Heart tissues were obtained from the KLK8-deficient (KLK8^-/-) and KLK8^+/+ mice after 24 weeks of diabetes. A, Immunoblots of KLK8 in the hearts. B and C, Masson's trichrome staining (B) and quantification (C) showed decreased collagen deposition in both interstitial and perivascular regions in the hearts obtained from KLK8^-/- diabetic mice as compared to KLK8^+/+ diabetic mice (scale bar = 50 µm). D-F, Cardiac fibrosis was also quantified by determination of collagen-I (D), hydroxyproline (E), and TGF-β1 (F) levels. Data are expressed as means ± SEM (n = 7). *p < 0.05,.**p < 0.01, ***p < 0.001, ****p < 0.0001. G, Representative M-mode images of echocardiographic analysis exhibited smaller left ventricular end-diastolic and end-systolic dimension in KLK8^-/- diabetic mice as compared to KLK8^+/+ diabetic mice after 12 and 24 weeks of diabetes.

Figure 3

KLK (kallikrein-related peptidase) 8 deficiency attenuates diabetic cardiac diastolic dysfunction. Both wild-type (KLK8^+/+) and KLK8-deficient (KLK8^-/-) mice were injected with STZ to construct diabetic mice. Cardiac diastolic function was assessed by using echocardiographic analysis at 16 weeks after STZ injection. A, Representative pulse doppler images of echocardiographic analysis. B-D, KLK8 deficiency partially rescued the reduced E/A ratio (B), and reduced E-wave deceleration time (C) and IVRT (D) in STZ-induced diabetic mice. E/A, ratio of E wave to A wave amplitude; Deceleration Time, E-wave deceleration time; IVRT, isovolumic relaxation time. Data are expressed as means ± SEM (n = 7). *p < 0.05, **p < 0.01, ****p < 0.0001.

Figure 4

KLK (kallikrein-related peptidase) 8 knockdown alleviates high glucose-induced endothelial dysfunction and endothelial-to-mesenchymal transition in human coronary artery endothelial cells (HCAECs). HCAECs were transfected with control siRNA or KLK8 siRNAs, and were cultured with or without high glucose (HG, 25 mM) treatment for 5 days. A-C, Levels of thrombomodulin (A), von Willebrand factor (B) and E-selectin (C) as markers of endothelial damage/dysfunction and activation in the culture medium. D, MTT assay showed that KLK8 knockdown alleviates HG-induced cell injury in HCAECs. E, KLK8 knockdown alleviates HG-induced TGF-β1 release in the supernatant of HCAECs. F & G, Immunoblots of the endothelial markers (VE-cadherin, CD31) and mesenchymal markers (vimentin, α-SMA). The representative protein bands (F) and the corresponding histograms (G) showed that HG-induced loss of CD31 and VE-cadherin was largely prevented, whereas the acquisition of α-SMA and vimentin was decreased by KLK8 knockdown. Data are expressed as means ± SEM (n = 4). **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 5

KLK (kallikrein-related peptidase) 8 deficiency attenuates diabetic endothelial damage and endothelial-to-mesenchymal transition. Plasma samples and heart tissues were obtained from the KLK8-deficient (KLK8^-/-) and KLK8^+/+ mice after 24 weeks of diabetes. A-C, Plasma levels of thrombomodulin (A), von Willebrand factor (B) and E-selectin (C). D & E, Immunoblots of VE-cadherin, CD31, vimentin and α-SMA in the hearts. The representative protein bands (D) and the corresponding histograms (E) showed that the loss of CD31 and VE-cadherin in diabetic heart tissues was largely prevented, whereas the acquisition of α-SMA and vimentin was dramatically decreased in KLK^-/- mice. F & G, Immunofluorescent staining showed decreased colocalization of α-SMA (F, green), FSP-1 (G, green) and CD31 (red) in the hearts obtained from KLK8^-/- diabetic mice as compared to KLK8^+/+ diabetic mice. Nuclei were counterstained with DAPI (blue), scale bar = 50 µm. Data are expressed as means ± SEM (n = 7). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 6

KLK (kallikrein-related peptidase) 8 overexpression leads to cardiac fibrosis and endothelial damage. A, Masson's trichrome staining showed collagen deposition in the perivascular and intramyocardial regions in the hearts obtained from 12-weeks-old Tg-KLK8 rats (scale bar = 50 µm). B-D, Cardiac fibrosis was also quantified by determination of collagen-I (B), hydroxyproline (C), and TGF-β (D) levels in heart tissues obtained from 6-weeks-old and 12-weeks-old control and Tg-KLK8 rats. E-G, Plasma levels of thrombomodulin (E), von Willebrand factor (F) and E-selectin (G) as markers of endothelial damage/dysfunction and activation in 6-weeks-old and 12-weeks-old control and KLK8 transgenic (Tg-KLK8) rats. Data are expressed as means ± SEM (n = 7). ***p < 0.001, ****p < 0.0001. H-K, HCAECs were infected with increasing doses of KLK8 adenovirus (Ad-KLK8) at a multiplicity of infection (MOI) of 1, 3, or 10 for 72 h. H-J, Levels of thrombomodulin (H), von Willebrand factor (I) and E-selectin (J) in the culture medium. K, MTT assay showed that Ad-KLK8 induced cell injury in a dose- and time-dependent manner in human coronary artery endothelial cells (HCAECs). L, The permeability of a confluent HCAECs monolayer measured by FITC-dextran flux assay showed that Ad-KLK8 (MOI 10) treatment for 72 h led to significantly increased permeability of HCAECs. Data are expressed as means ± SEM (n = 4). *p < 0.05, **p < 0.01, ****p < 0.0001, ##p < 0.01, ####p < 0.0001 versus Ad-vector Day 3; $$$$p < 0.0001 versus Ad-vector Day 5.

Figure 7

KLK (kallikrein-related peptidase) 8 overexpression induces endothelial-to-mesenchymal transition in the myocardium and human coronary artery endothelial cells (HCAECs). A, Immunofluorescent staining showed increased colocalization of α-SMA (green) and FSP-1 (green) and CD31 (red) in the hearts obtained from 12-weeks-old Tg-KLK8 rats as compared to age-matched control rats. Nuclei were counterstained with DAPI (blue), scale bar = 50 µm. B, Representative z-stack image analysis showed specific overlay of double immunostaining. CD31⁺/α-SMA⁺ and CD31⁺/FSP-1⁺ cells in specific ordinate were analyzed in z-stack with optimal interval range of 0.8 µm. Nuclei were counterstained with DAPI (blue), scale bar = 50 µm. C & D, Immunoblots of the VE-cadherin, CD31, vimentin, and α-SMA in the hearts obtained from 12-weeks-old control and Tg-KLK8 rats. The representative protein bands (C) and the corresponding histograms (D) showed that transgenic KLK8 overexpression significantly upregulated vimentin and α-SMA expression, whereas downregulated VE-cadherin and CD31 expression in heart tissues. E & F, Immunoblots of the endothelial markers (VE-cadherin, CD31) and mesenchymal markers (vimentin, α-SMA) in HCAECs infected with increasing doses of KLK8 adenovirus (Ad-KLK8) at a multiplicity of infection (MOI) of 1, 3, or 10 for 72 h. The representative protein bands (E) and the corresponding histograms (F) showed that Ad-KLK8 dose-dependently increased protein expressions of the mesenchymal markers, whereas decreased the endothelial markers in HCAECs. G, Infection of Ad-KLK8 at an MOI of 10 for 72 h led to an increase of TGF-β1 release in the supernatant of HCAECs. H & I, Immunoblots of the endothelial markers (VE-cadherin, CD31) and mesenchymal markers (vimentin, α-SMA). The representative protein bands (H) and the corresponding histograms (I) showed that the KLK8-induced loss of CD31 and VE-cadherin was largely prevented, whereas the acquisition of α-SMA and vimentin was decreased by TGF-β1 neutralizing antibody. Data are expressed as means ± SEM (n = 4). *p < 0.05, **p < 0.01, ****p < 0.0001.

Figure 8

Sp-1 mediates high glucose-induced upregulation of KLK8 in human coronary artery endothelial cells (HCAECs). A-C, HCAECs were treated with increasing concentration of glucose (15 and 25 mM) for 5 days. A, The mRNA level of KLK8 in HCAECs. B and C, Immunoblots of KLK8 (B) and Sp-1 (C). The representative protein bands of KLK8 (B) and Sp-1 (C) were presented on the top of the corresponding histograms. D and E, High glucose-induced mRNA (D) and protein (E) expressions of KLK8 were abolished by Sp-1 inhibitor plicamyin (1 nM) in HCAECs. F, HCAECs were transfected with human KLK8 promoter-luciferase reporter plasmid pGL3-KLK8 or mutated plasmid pGL3-KLK8△Sp1, and then were exposed to high glucose (25 mM) for 48 h. Promoter activity was analyzed using a dual-luciferase reporter assay. Data are expressed as means ± SEM (n = 4). **p < 0.01, ****p < 0.0001.

Figure 9

KLK (kallikrein-related peptidase) 8 degrades VE-cadherin, thus promoting plakoglobin nuclear translocation. A and B, Association of plakoglobin with VE-cadherin and KLK8 was tested by reciprocal immunoprecipitations in rat myocardium and human coronary artery endothelial cells (HCAECs). IgG was controlled for nonspecific interaction. C, HCAECs were treated with Ad-vector or Ad-KLK8 in serum-free medium for 72 h. Immunoblots showed the appearance of a ~30 kDa N-terminal VE-cadherin fragment in the medium. WB, western blot. D, The N-terminal sequence of the ~30 kDa protein band was determined by Edman assay. The sequence as indicated was belonging to VE-cadherin sequence. E, Schematic representation of the extracellular part of VE-cadherin cleaved by KLK8. VE-cadherin extracellular domain is constituted by 5 cadherin domains numbered 1-5 from the N-terminus. KLK8 cleaved VE-cadherin before amino-acid 266 in the extracellular cadherin domain 3. F, Immunofluorescent staining showed that infection of Ad-KLK8 for 72 h resulted in significant loss of both plakoglobin (red) and VE-cadherin (green) in the plasma membrane, whereas caused nuclear translocation of plakoglobin in HCAECs. Lentivirus-mediated VE-cadherin (Lv-VE-cadherin) overexpression reduced nuclear whereas increased membrane and cytosol plakoglobin levels. Nuclei were counterstained with DAPI (blue), scale bar = 100 µm. G, Immunoblots of cellular fractionations of plasma membrane, cytosol and nucleus for VE-cadherin and plakoglobin in HCAECs infected with or without Ad-KLK8 and Lv-VE-cadherin for 72 h. The representative protein bands (H) and the corresponding histograms (I) were presented. Data are expressed as means ± SEM (n = 4). **p < 0.01, ****p < 0.0001.

Figure 10

Plakoglobin is required for KLK (kallikrein-related peptidase) 8-induced endothelial-to-mesenchymal transition by cooperating with p53. A & B, Immunoblots of plakoglobin-knockdown human coronary artery endothelial cells (HCAECs) under KLK8 adenovirus (Ad-KLK8) treatment for 72 h. The representative protein bands (A) and the corresponding histograms (B) were presented. C, Association of plakoglobin with TCF-4 and p53 was observed by immunoprecipitation in HCAECs, whereas Ad-KLK8 treatment enhanced the interactions. IgG was controlled for nonspecific interaction. D, The KLK8 overexpression-induced mRNA expression of TGF-β1 was reduced by plakoglobin knockdown or HIF-1α inhibitor echinomycin (20 nM) in HCAECs. E, ChIP assay showed that the KLK8 overexpression-induced HIF-1α binding to TGF-β1 promoter was blocked by plakoglobin knockdown or p53 inhibitor pifithrin-α (20 μM). F, Immunoprecipitation assay showed that Ad-KLK8 treatment enhanced the association of p53 with HIF-1α and Smad3 in HCAECs, which was reduced by plakoglobin knockdown. G and H, The KLK8 overexpression-induced mRNA levels of the pro-EndMT target genes of TGF-β1/Smad pathway (Snail, Slug, Zeb1, Zeb2 and Twist) were reduced by plakoglobin knockdown (G) or p53 inhibitor pifithrin-α (H, 20 μM). Data are expressed as means ± SEM (n = 4). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 11

High glucose promotes plakoglobin-dependent cooperation of p53 with HIF-1α and Smad, subsequently increasing the expression of TGF-β1 and its pro-EndMT target genes in a KLK (kallikrein-related peptidase) 8-dependent manner. A, Immunoprecipitation assay showed that high glucose (HG, 25 mM glucose) treatment enhanced the association of p53 with plakoglobin, HIF-1α and Smad3 in HCAECs, which was reduced by KLK8 knockdown. B, Immunoprecipitation assay showed that HG-induced association of p53 with HIF-1α and Smad3 was reduced by plakoglobin knockdown in HCAECs. C and D, ChIP assay showed that the HG-induced HIF-1α binding to TGF-β1 promoter was blocked by knockdown of KLK8 or plakoglobin (C) or p53 inhibitor pifithrin-α (D, 20 μM). Data are expressed as means ± SEM (n = 4). NG indicates normal glucose. E-G, Heart tissues were obtained from the KLK8-deficient (KLK8^-/-) and KLK8^+/+ mice after 24 weeks of diabetes. The mRNA levels of the pro-EndMT target genes of TGF-β1/Smad pathway (Snail, Slug, Zeb1, Zeb2 and Twist) were decreased in the hearts obtained from KLK8^-/- diabetic mice, as compared to KLK8^+/+ diabetic mice (E). F, Immunoprecipitation assay showed that the association of p53 with plakoglobin, HIF-1α and Smad3 was reduced in KLK8^-/- diabetic mice as compared to KLK8^+/+ diabetic mice. G, ChIP assay showed that KLK8 deficiency suppressed HIF-1α binding to TGF-β1 promoter in diabetic myocardium. Data are expressed as means ± SEM (n = 7). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. H, Schematic diagram of the mechanism by which KLK8 promotes EndMT and cardiac fibrosis in diabetic cardiomyopathy. Hyperglycemia upregulates KLK8 expression in endothelial cells, which cleavages the VE-cadherin extracellular domain and promotes plakoglobin nuclear translocation and its cooperation with p53. The plakoglobin-dependent cooperation of p53 with HIF-1α and Smad3 subsequently increased the expression of TGF-β1 and the pro-EndMT target genes of TGF-β1/Smad pathway, which finally promotes the differentiation of endothelial cells into mesenchymal cells and the pathogenesis of cardiac fibrosis.