Parathyroid hormone-related protein induces fibronectin up-regulation in rat mesangial cells through reactive oxygen species/Src/EGFR signaling

Parathyroid hormone-related protein (PTHrP) is known to be up-regulated in both glomeruli and tubules in patients with diabetic kidney disease (DKD), but its role remains unclear. Previous studies show that PTHrP-induced hypertrophic response in mesangial cells (MCs) and epithelial-mesenchymal transition (EMT) in tubuloepithelial cells can be mediated by TGF-β1. In the present study, although long-term PHTrP (1–34) treatment increased the mRNA and protein level of TGF-β1 in primary rat MCs, fibronectin up-regulation occurred earlier, suggesting that fibronectin induction is independent of TGF-β1/Smad signaling. We thus evaluated the involvement of epidermal growth factor receptor (EGFR) signaling and found that nicotinamide adenine dinucleotide phosphate oxidase-derived reactive oxygen species mediates PTHrP (1–34)-induced Src kinase activation. Src phosphorylates EGFR at tyrosine 845 and then transactive EGFR. Subsequent PI3K activation mediates Akt and ERK1/2 activation. Akt and ERK1/2 discretely lead to excessive protein synthesis of fibronectin. Our study thus demonstrates the new role of PTHrP in fibronectin up-regulation for the first time in glomerular MCs. These data also provided new insights to guide development of therapy for glomerular sclerosis.


Introduction
Parathyroid hormone-related protein (PTHrP) was first discovered as the factor responsible for humoral hypercalcemia of malignancy. Different from the endocrine regulator parathyroid hormone (PTH), PTHrP is produced and secreted from nonmalignant fetal and adult tissues and plays a critical role in the development and/or growth regulation of bone, heart, mammary glands, and other tissues [1,2]. In the adult kidney, PTHrP is abundant in the glomeruli, tubules, and intrarenal arterial tree [3,4]. PTHrP exerts a modulatory action on renal function including renal plasma flow, glomerular filtration rate, etc [5].
PTHrP is known to be up-regulated in both glomeruli and tubules in patients with diabetic kidney disease (DKD) [6], but its role remains unclear. Izquierdo et al. developed a transgenic mouse model characterized by PTHrP overexpression in the renal proximal tubule. They found that chronic PTHrP overexpression showed no morphologic or functional alterations in basal conditions. But in streptozotocin (STZ)-induced experimental DKD, PTHrP transgenic mouse developed increased renal hypertrophy, a higher urinary albumin excretion (UAE), and lower total plasma protein levels than control mice [7]. It has been reported that PTHrP might promote epithelial-mesenchymal transition (EMT) through interaction with vascular endothelial growth factor (VEGF), TGF-β1, and epidermal growth factor (EGF) in renal tubuloepithelial cells [8,9]. In glomerular mesangial cells (MCs), continuous expression of or incubation with PTHrP induces a proliferative effect (24 h) followed by hypertrophy at 72 h [10][11][12], and PTHrP-induced hypertrophic response is probably mediated by TGF-β1 [12]. Our previous study showed that PTHrP could induce fibronectin up-regulation in rat MCs [13]. However, whether PTHrP-induced extracellular matrix (ECM) up-regulation is mediated by TGF-β1 is still not clear. The present study found that PTHrP  peptide-induced fibronectin up-regulation in rat MCs could be independent of TGF-β/Smad signaling. PTHrP  exposure induced reactive oxygen species (ROS)-dependent, Src kinase-mediated epidermal growth factor receptor (EGFR) transactivation and subsequent Akt and ERK1/2 activation, which involve in fibronectin up-regulation. It could help to make certain the role of PTHrP in the accumulation of ECM and provide new thought for the therapeutic strategy of glomerular sclerosis.

Cell culture and treatments
After removal of kidneys from two to four Sprague-Dawley rats, trim the perirenal fat in PBS in a petri dish and then quickly remove the capsule. Cut in half lengthwise and remove the medulla. Then mush through the 250, 106, and 75 μm sieve with PBS sequencially. Wash with cold PBS to collect all the glomeruli and digest with collagenase for 15 min at 37 • C. Next, place glomeruli in 100 mm plates with 8 ml of low-glucose DMEM (5.6 mM glucose) containing 20% FBS (Invitrogen, Carlsbad, U.S.A.), change the medium twice per week, and then subculture cell when confluent. Freeze down at each passage after passage 2. Characterize MCs by RT-PCR for vimentin, keratin, and desmin. MCs are positive for vimentin, negative for keratin. MCs were cultured in low-glucose DMEM containing 20% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37 • C in a 5% CO 2 atmosphere. All of the experiments were performed between passages 6 and 18. In vitro studies have established that the amino-terminal peptide fragments are sufficient for the actions of PTHrP, as PTHrP  and PTHrP (1-36) peptide display high-affinity receptor binding and efficient receptor activation [14]. Based on these data, we substituted 100 nM PTHrP (1-34) peptide (Bachem, Bubendorf, Swiss) for PTHrP. Pharmacologic inhibitors were added at the indicated concentrations and durations before PTHrP

Quantitative real-time PCR
Quantitative PCR (qPCR) was performed using RNA extracted from rat MCs. Total RNA was isolated using RNA Extraction Kit (Qiagen, Germany). cDNA was reverse transcribed using Reverse Transcription kit (GeneCopoeia, Rockville, U.S.A.). Quantitative PCR was performed in duplicate using qPCR Kit (GeneCopoeia). Negative controls of cDNA were included for each gene set in all reactions to detect contamination. The primer sequences are shown as follows: GAPDH, sense, 5 -TGCACCACCAACTGCTTAGC-3 , antisense, 5 -GGCATGGACTGTGGTCATGAG-3 ; TGF-β1, sense, 5 -AAACGGAAGCGCATCGAA-3 , antisense, 5 -GGGACTGGCGAGCCTTAGTT-3 . The thermo-cycle program was performed in MiniOpticon (Bio-Rad, Hercules, U.S.A.), and was set as 5 min at 95 • C, followed by 30 cycles of at 95 • C for 30 s, 60 • C for 30 s, and 72 • C for 1 min. Gene expression level was calculated using the Ct method relative to GAPDH.

Measurement of ROS generation by flow cytometry
MCs cultured in a 24-well plate were made quiescent in serum-free medium for 24 h and incubated with 5 μM 2 ,7 -dichlorodihydrofluorescin diacetate (DCFH-DA, Beyotime Biotechnology, Shanghai, China) or dihydroethidium (DHE, Beyotime) at 37 • C for 30 min, followed by washing three times with PBS. The cells were then left untreated or treated with 100 μM apocynin for 30 min or 10 μM NAC for 10 min before addition of 100 nM PTHrP (1-34) for 5 min, followed by washing three times with PBS. Next, the cells were trypsinized and resuspended in PBS. The fluorescence intensity (DCFH-DA: excitation wavelength 488 nm and emission wavelength 535 nm; DHE: excitation wavelength 300 nm and emission wavelength 610 nm) was measured by using flow cytometry (CytoFLEX, Beckman Coulter) and analyzed with FlowJo 7.6.1 software.

Determination of nicotinamide adenine dinucleotide phosphate oxidase activity
Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX) activity was determined by measuring the NADPH-dependent superoxide dismutase (SOD)-inhibitable cytochrome C reduction. The measurement was performed according to the manufacturers' instruction (GENMED, Shanghai, China). Briefly, 900 μl of the reaction buffer containing NOX substrate (NADPH) and oxidized cytochrome C in a quartz cuvette was preincubated at 30 • C for 3 min. Next, 100 μl supernatant from MC lysate of different groups was added to the reaction mixture and incubated at 30 • C for 15 min. The absorbance at 550 nm was read by a spectrophotometer. The NOX activity was calculated as SOD-inhibitable cytochrome C reduction and expressed as O 2 − in nmol/min/mg.

Diabetic rat model
The Center for Animal Experiment in Wuhan University approved all experiments and all experiments were performed according to Chinese Ethics Community Guidelines. Male Sprague-Dawley rats (weighing 200-225 g) were injected with STZ by tail vein (55 mg/kg body weight, freshly prepared in 0.1 mol/l citrate buffer, pH 4.5) or vehicle alone to induce diabetes. Total 2 days later, hyperglycemia (blood glucose > 20 mM) was confirmed using a reflectance meter (One Touch, Lifescan, Milpetas, U.S.A.). Diabetic rats were injected with different dosages of PTHrP (1-34) peptide (Bachem) by subcutaneous (s.c.) injection (40, 80, or 160 μg/kg body weight, respectively). Injections were once daily for 5 days per week for 3 months [16]. Blood glucose levels were monitored and systolic blood pressure was determined weekly in all diabetic rats. Urine was collected for 24 h in metabolic cages at 3 months. Rats were then anesthetised for kidney removal. Kidney samples were rapidly excised, weighed, and frozen in liquid nitrogen. At the time of killing, renal hypertrophy was assessed by kidney to body weight ratio (mg/g).

Biochemical analysis
Serum calcium and phosphate, serum creatinine, urinary creatinine, and albumin were determined by the clinical laboratory of Zhongnan Hospital, Wuhan University. Creatinine clearance, Ccr (ml · min −1 · 100 g body weight −1 ) was calculated as urine creatinine × urine volume (ml · min −1 )/serum creatinine/100 g body weight.

Immunohistochemistry
The renal cortex was fixed by 4% paraformaldehyde, embedded in paraffin and then cut into slices with a thickness of 4 μm. Next, renal cortical sections were routinely deparaffinized, rehydrated, and subjected to heat-induced antigen retrieval. After blocking endogenous peroxidase activity and rinsing, these sections were blocked with 5% BSA in PBS and incubated with mouse monoclonal fibronectin antibody (1:500, BD Biosciences, San Jose, U.S.A.) at 4 • C overnight. The sections were subsequently incubated with anti-mouse secondary antibody for 1 h and then stained by 3, 3 -diaminobenzidine (DAB) to produce brown colorization. Hematoxylin was used for counterstaining. Negative control was performed without incubation within primary antibody. Positive staining signal percentage was quantitated using the image analysis system (Image-Pro Plus 7.0). Total ten high-power microscope fields (400×) were selected randomly.

Statistical analysis
All data are presented as means + − S.E.M. Experimental repetition times (n) were given in figure legends. Results analysis was performed by one-way ANOVA with Turkey Honestly Significant Difference (HSD) for post hoc analysis (SPSS 20.0 for Windows). A P value <0.05 is considered statistical significant.

PTHrP (1-34)-induced fibronectin up-regulation is independent of TGF-β1 signaling in rat MCs
It has been reported that prolonged PTHrP exposure induces a hypertrophic response mediated by TGF-β1 in cultured human MCs [12]. We further observed whether the TGF-β1 system involves in PTHrP-induced fibronectin up-regulation in rat MCs. PTHrP (1-34) increased the mRNA and protein expression of TGF-β1 after 24 h ( Figure  1A,B), and an increase in TβRII protein expression was also induced at 72 h ( Figure 1B), which is consistent with previous findings [12]. We then investigated the effect of PTHrP (1-34) exposure on fibronectin protein level. As shown in Figure 1C, fibronectin protein expression markedly increased from 12 h and continued to 48 h. However, Smad2/3 phosphorylation was not affected by PTHrP (1-34) until 72 h (Supplementary Figure S1A). We next used TGF-β receptor inhibitor SB431542 to block TGF-β1 signaling. It has no effect on PTHrP (1-34)-induced fibronectin up-regulation ( Figure 1D). These results strongly indicate that PTHrP might induce fibronectin up-regulation independent of TGF-β1/Smad signaling.

PTHrP (1-34) induces Src-dependent EGFR transactivation and Akt phosphorylation
Given the well-known effect of PKA and PKC as downstream effects of PTH/PTHrP signaling [17], we evaluated their involvement in PTHrP-induced fibronectin up-regulation by using the PKA inhibitor H-89 and the PKC inhibitor bisindolylmaleimide I. Neither inhibitor affected PTHrP (1-34)-induced FN up-regulation (Supplementary Figure  S1B,C), indicating that PTHrP-induced fibronectin up-regulation is probably independent of PKA and PKC signaling. EGFR is known to aid in transmitting signals for diverse nonligand mediated stimuli in a process known as transactivation [18]. Stimulation of PTH type 1 receptor (PTH1R), common to PTH and PTHrP, leads to EGFR transactivation in human embryonic kidney cells HEK-293, murine osteoblasts and renal tubule cells [9,19,20]. We then tested whether the activation of PTH1R by PTHrP (1-34) treatment induced EGFR transactivation in rat MCs. As seen in Figure 2A, PTHrP (1-34) induced sustained EGFR transactivation, as determined by phosphorylation at tyrosine 845 (Y845), but not Y1173. PI3K/Akt pathway is usually the downstream of EGFR signaling, so we investigated whether PTHrP (1-34) induced Akt activation. As a result, Akt phosphorylation at serine 473 (S473), an indicator of Akt activity, was significantly increased at 10 min by PTHrP (1-34) treatment ( Figure 2A). Moreover, Akt activation triggered by PTHrP (1-34) was prevented when we treated MCs with the specific EGFR kinase inhibitor gefitinib or AG1478 ( Figure 2B), suggesting that EGFR activation is an important upstream event in PI3K/Akt pathway.
Several studies confirmed that PTHrP (107-111) peptide (known as osteostatin domain) may activate c-Src (Src) kinase in osteoblastic cells [21,22]. Whether PTHrP (1-34) peptide mediates Src kinase activation is not known. We thus examined the phosphorylation level of Src kinase at tyrosine 416 (Y416), which represents the activation of Src kinase [23]. In response to PTHrP (1-34), Src phosphorylation on Y416 increased at 5 min, and lasted until 3 h, the last time point tested ( Figure 2C). The increase in phosphorylation level of Src is earlier than that of EGFR phosphorylation, indicating that Src kinase might mediate EGFR transactivation in MCs. We thus took advantage of specific Src kinase activity inhibitor PP1 and SU6656 to investigate the effects of Src kinase on EGFR transactivation and subsequent Akt activation, and found that Src kinase inhibition blocked PTHrP (1-34)-induced EGFR and Akt phosphorylation ( Figure 2D). Taken together, these results suggested that PTHrP (1-34)-induced EGFR transactivation and Akt activation require Src kinase.
On the other hand, G protein-coupled receptors (GPCRs), including PTH1R, are known to transactivate EGFR by soluble EGF-like ligand (such as heparin-binding EGF, HB-EGF) release and binding in a variety of cell types [9,19,20]. Although PTH (1-36) rapidly (within 5 min) and transiently induces EGFR transactivation through PKC and matrix metalloproteinases (MMPs)-mediated proteolytic processing of EGFR ligands in murine cortical tubule cells [9], either GM6001, a pan-specific MMPs inhibitor or CRM197, the HB-EGF inhibitor could not prevent Src, EGFR, and Akt phosphorylation by PTHrP  in rat MCs ( Figure 4A,B), suggesting that PTHrP (1-34)-induced sustained Src activation and EGFR transactivation are independent of MMPs-mediated HB-EGF cleavage and release.

ROS-mediated Src/EGFR/Akt signaling mediates PTHrP (1-34)-induced fibronectin up-regulation
Next, we investigated whether ROS-mediated Src/EGFR/Akt signaling involves in PTHrP (1-34)-induced fibronectin up-regulation in rat MCs. Both NOX inhibitor and Src kinase inhibitor blocked fibronectin up-regulation in response to PTHrP (1-34) exposure ( Figure 5A,B). Inhibition of EGFR signaling by gefitinib, or inhibition of Akt signaling with the specific PI3K inhibitor, LY294002, also prevented fibronectin up-regulation ( Figure 5C,D). These results indicated that PTHrP (1-34)-induced fibronectin up-regulation is dependent on NOX-derived ROS in rat MCs, which activate Src kinase and downstream EGFR/PI3K/Akt signaling. To verify this, we further observed the effect of long-term PTHrP injection on STZ-induced diabetic rats. During the 3 months of PTHrP (1-34) treatment, there was no change in the serum calcium and phosphate levels amongst groups. Diabetic rats had remarkably higher UAE compared with controls, and this was unaffected by PTHrP (1-34). Diabetic rats developed renal hypertrophy, while administration of all three doses of PTHrP (1-34) aggravated renal hypertrophy ( Table 1). Immunohistochemistry of cortical sections for fibronectin showed significantly increased staining both in the glomeruli and tubular cells in diabetic rats at 3 months compared with control. Fibronectin staining was stronger in 160 μg/kg PTHrP (1-34) group compared with diabetic rats (Figure 6A), indicating that long-term treatment with PTHrP promote fibronectin up-regulation in the renal cortices of diabetic rats. We also tested the protein level of fibronectin in renal cortex by western blot. Same as immunohistochemistry, increased fibronectin protein level was also observed in PTHrP (1-34)-injected diabetic rat kidney at 3 months ( Figure 6B).
An important source of ROS in the diabetic glomeruli is NOX1 and NOX2 [27,28]. The cytosolic p47 phox subunit is a key regulator of NOX1 and NOX2 [29,30]. Since PTHrP (1-34) induced NOX-dependent ROS generation in rat MCs, we investigated the role of PTHrP  in the expression of p47 phox . Increased p47 phox protein level was seen in diabetic renal cortex, and p47 phox protein level was even higher in 160 μg/kg PTHrP (1-34) group ( Figure 6C). We also observed whether PTHrP (1-34) treatment has an effect on EGFR and Akt activation in diabetic rat kidney. Consistent with the in vitro findings, the phosphorylation levels of EGFR and Akt were markedly increased in renal  cortex in 160 μg/kg PTHrP (1-34)-injected diabetic rats ( Figure 6C). These results collectively suggested that PTHrP (1-34) induced ROS production and increased EGFR/Akt phosphorylation in diabetic rat kidney.

Discussion
In the present study, we found for the first time that PTHrP (1-34) peptide-induced fibronectin up-regulation could be independent of TGF-β/Smad signaling in rat MCs. PTHrP (1-34) exposure induces ROS-dependent, Src kinase-mediated EGFR transactivation and subsequent Akt and ERK1/2 activation, which lead to increased fibronectin expression. In STZ-induced diabetic rat, long-term treatment of PTHrP  peptide was found to result in increased renal hypertrophy, higher expression of p47 phox , phosphorylation of EGFR, Akt and ERK1/2, and higher expression of fibronectin in renal cortex. TGF-β/Smad signaling has been recognized as a key mediator in ECM accumulation and renal fibrosis [32,33]. PTHrP might promote EMT through interaction with TGFβ1 in renal tubule cells [8,9]. Our previous study showed that PTHrP could induce fibronectin up-regulation in rat MCs [13]. In the present work, we tested the role of TGF-β1