Overexpression of heterogeneous nuclear ribonucleoprotein F stimulates renal Ace-2 gene expression and prevents TGF-β1-induced kidney injury in a mouse model of diabetes

Aims/hypothesis We investigated whether heterogeneous nuclear ribonucleoprotein F (hnRNP F) stimulates renal ACE-2 expression and prevents TGF-β1 signalling, TGF-β1 inhibition of Ace-2 gene expression and induction of tubulo-fibrosis in an Akita mouse model of type 1 diabetes. Methods Adult male Akita transgenic (Tg) mice overexpressing specifically hnRNP F in their renal proximal tubular cells (RPTCs) were studied. Non-Akita littermates and Akita mice served as controls. Immortalised rat RPTCs stably transfected with plasmid containing either rat Hnrnpf cDNA or rat Ace-2 gene promoter were also studied. Results Overexpression of hnRNP F attenuated systemic hypertension, glomerular filtration rate, albumin/creatinine ratio, urinary angiotensinogen (AGT) and angiotensin (Ang) II levels, renal fibrosis and profibrotic gene (Agt, Tgf-β1, TGF-β receptor II [Tgf-βrII]) expression, stimulated anti-profibrotic gene (Ace-2 and Ang 1–7 receptor [MasR]) expression, and normalised urinary Ang 1–7 level in Akita Hnrnpf-Tg mice as compared with Akita mice. In vitro, hnRNP F overexpression stimulated Ace-2 gene promoter activity, mRNA and protein expression, and attenuated Agt, Tgf-β1 and Tgf-βrII gene expression. Furthermore, hnRNP F overexpression prevented TGF-β1 signalling and TGF-β1 inhibition of Ace-2 gene expression. Conclusions/interpretation These data demonstrate that hnRNP F stimulates Ace-2 gene transcription, prevents TGF-β1 inhibition of Ace-2 gene transcription and induction of kidney injury in diabetes. HnRNP F may be a potential target for treating hypertension and renal fibrosis in diabetes. Electronic supplementary material The online version of this article (doi:10.1007/s00125-015-3700-y) contains peer-reviewed but unedited supplementary material, which is available to authorised users.

We previously reported that high glucose milieu enhances expression of angiotensinogen (AGT, the sole precursor of all angiotensins) through generation of reactive oxygen species (ROS) in cultured rat renal proximal tubular cells (RPTCs) [17,18]. Rat AGT overexpression in RPTCs leads to hypertension, albuminuria and RPTC hypertrophy, and enhances TGF-β1 expression in diabetic AGT-transgenic (Tg) mice [19,20]. Conversely, RPTC-selective overexpression of catalase or pharmacological blockade of the renin-angiotensin system (RAS) attenuates hypertension, ROS generation, kidney injury and normalised RPTC ACE-2 expression in mouse models of diabetes [21][22][23][24]. Taken together, these observations indicate that oxidative stress-induced upregulation of AGT expression and downregulation of ACE-2 expression in RPTCs, resulting in higher angiotensin (Ang)II/Ang 1-7 ratio, may be key determinants of development of hypertension and nephropathy in diabetes.
We reported that insulin inhibits high glucose stimulation of rat renal Agt gene expression via two nuclear proteins-heterogeneous nuclear ribonucleoproteins F and K (hnRNP F, hnRNP K)-that interact with the insulin-responsive element (IRE) in the Agt gene promoter [25][26][27][28], and that hnRNP F overexpression in RPTCs inhibits Agt gene expression and kidney hypertrophy in Akita Hnrnpf-Tg mice [29]. Here, we report that overexpression of hnRNP F stimulates Ace-2 gene transcription and suppresses profibrotic gene (Tgf-β1, Tgf-βrII) expression in RPTCs of Akita Hnrnpf-Tg mice. We have confirmed these changes by in vitro studies in rat RPTCs. We also show that hnRNP F overexpression prevents TGF-β1 signalling and inhibition of Ace-2 gene expression in RPTCs. Finally, we identified the putative DNA response elements (REs) in the Ace-2 gene promoter that are responsive to hnRNP F and TGF-β1.

Methods
Chemicals and constructs Active human recombinant TGF-β1 was obtained from R&D Systems (Minneapolis, MN, USA). SB431542 (a TGF-β receptor I [RI] inhibitor) and other chemicals were purchased from Sigma-Aldrich (Oakville, ON, Canada). The antibodies used in the present study are listed in electronic supplementary material (ESM) Table 1. The pKAP2 plasmid containing the kidney-specific androgen-regulated protein (KAP) promoter was a gift from C. D. Sigmund (University of Iowa, Iowa City, IA, USA) [30]. Full-length rat Hnrnpf cDNA fused with HA tag (encoding amino acid residues 98-106 [YPYDVPDYA] of human influenza virus hemagglutinin) was inserted into pKAP2 plasmid at the NotI site at both 5′ and 3′ termini [25,29]. pGL4.20 vector containing Luciferase reporter was obtained from Promega (Sunnyvale, CA, USA). Rat Ace-2 gene promoter (N-1,091/+83) was cloned from rat genomic DNA with specific primers (ESM Table 2), as described by Milsted et al [31] and then inserted into pGL4.20 plasmid at HindIII and KpnI restriction sites. Scrambled Silencer Negative Control no. 1 small interfering RNA (siRNA) and Hnrnpf siRNA were bought from Ambion (Austin, TX, USA). QuickChange II Site-Directed Mutagenesis Kit and LightShift Chemiluminescent electrophoretic mobility shift assay (EMSA) Kit were procured from Agilent Technologies (Santa Clara, CA, USA) and Thermo Scientific (Life Technologies, Burlington, ON, Canada), respectively. The primer biotinlabelling kit was purchased from Integrated DNATechnologies (Coralville, IA, USA).
Physiological studies Adult male heterozygous Akita mice (Mus musculus) with a mutated Ins2 gene (C57BL6-Ins2 Akita /J) were purchased from Jackson Laboratories (Bar Harbor, ME, USA: http://jaxmice.jax.org). Akita Tg mice (C57Bl/6 background) overexpressing rat hnRNP F-HA in RPTCs (line 937) were created in our laboratory (by J. S. D. Chan) [29]. Male adult non-Tg and non-Akita littermates served as wild-type (WT) controls, and were tested along with Hnrnpf-Tg, Akita and Akita Hnrnpf-Tg mice. All animals were housed individually in metabolic cages for 24 h before euthanasia at age 20 weeks. All animals were fed standard mouse chow and water ad libitum. Animal care and procedures were approved by the CRCHUM Animal Care Committee and followed the Principles of Laboratory Animal Care (NIH publication no. 85-23, revised 1985: http:// grants1.nih.gov/grants/olaw/references/phspol.htm).
Kidneys were removed immediately after GFR measurement, decapsulated and weighed. The left kidneys were processed for histology and immunostaining, and right renal cortices were harvested for renal proximal tubules (RPTs) isolation by Percoll gradient centrifugation [23,24,28,29]. Aliquots of freshly isolated RPTs from individual mice were immediately processed for total RNA and protein isolation.
Cell culture Immortalised rat RPTCs (passages 12-18) [35] were cultured in 5 mmol/l D-glucose DMEM containing 5% FBS until they reached 60-70% confluence. The media were then changed to serum-free DMEM, ensuring that endogenously secreted TGF-β1 would not interfere in the assay. After 45 min preincubation, active human recombinant TGF-β1 [36] (0 to 10 ng/ml) was added (considered as time 0 h) and incubated for various time periods up to 24 h. In separate experiments, RPTCs were incubated for 24 h in serum-free medium in the presence or absence of TGF-β1± various concentrations of SB431542.
Effect of hnRNP F overexpression on AGT, ACE, ACE-2 and MasR expression in Akita Hnrnpf-Tg mouse kidneys Immunostaining revealed that HnRNP F (Fig. 1a) was overexpressed in RPTCs of Hnrnpf-Tg and Akita Hnrnpf-Tg mice compared with WT and Akita mice, respectively. ACE-2 ( Fig. 1b) and MasR (Fig. 1c) expression was decreased in Akita mice compared with WT controls and normalised in Akita Hnrnpf-Tg mice. RPTC ACE (Fig. 1d) expression did not differ between WT and Hnrnpf-Tg mice, whereas ACE expression was significantly higher in Akita mice than in WT controls and was not normalised in Akita Hnrnpf-Tg mice. WB and RT-qPCR for hnRNP F, ACE-2, MasR and ACE protein and their mRNA levels ( Fig. 1e-l, respectively) confirmed these observations.
HnRNP F overexpression suppresses renal fibrosis in Akita Hnrnpf-Tg mice Akita mice developed renal structural damage compared with WT and Hnrnpf-Tg mice (ESM Fig. 1a, PAS staining), including tubular luminal dilatation with accumulation of cell debris, increased extracellular matrix proteins in glomeruli and tubules, and proximal tubule cell atrophy. HnRNP F overexpression markedly reversed but never completely resolved these abnormalities in Akita mice. We detected significant increases in Masson's trichrome staining (Fig. 3a) and immunostaining for collagen type IV (Fig. 3b), fibronectin 1 expression (Fig. 3c) and collagen type I (Fig. 3d) in glomerulotubular areas in Akita mice compared with WT controls and Hnrnpf-Tg mice. These changes were reduced in Akita Hnrnpf-Tg mice. Quantification of Masson's trichrome-stained (ESM Fig. 1b), immunostaining of collagen IV (Fig. 3e), fibronectin 1 (Fig. 3f) and collagen I (Fig. 3g), and RT-qPCR quantification of mRNA levels (Fig. 3h-j) confirmed their expression.

Discussion
The present report identifies a novel mechanism by which hnRNP F prevents hypertension and kidney injury in diabetic Akita mice, i.e. hnRNP F stimulation of renal Ace-2 gene transcription and mitigation of the inhibitory effect of TGF-β1 on Ace-2 gene transcription.
We reported previously that overexpression of hnRNP F prevents systemic hypertension, and inhibits renal Agt gene expression and RPTC hypertrophy in diabetic Akita Hnrnpf-Tg mice [29]. The present paper provides new in vivo and in vitro evidence that hnRNP F stimulates Ace-2 gene transcription via binding to the DNA-RE of the Ace-2 gene promoter, which is critical for the formation of renal Ang 1-7 and subsequent expression of its antihypertensive and renoprotective actions in Akita mice [37].
HnRNP F, a member of the pre-mRNA-binding protein family [38] regulates gene expression at both the transcriptional and post-transcriptional levels. Indeed, hnRNP F engages in alternative splicing of various genes [39][40][41] and associates with TATA-binding protein, RNA polymerase II, nuclear cap-binding protein complex and various transcriptional factors. [42,43] The Akita mouse is an autosomal-dominant model of spontaneous type 1 diabetes in which the Ins2 gene is mutated. Akita mice develop hyperglycaemia and systemic hypertension, leading to cardiac hypertrophy, left ventricular diastolic dysfunction, glomerulosclerosis and enhanced oxidative stress in RPTs, closely resembling those observed in patients with type 1 diabetes [44,45].
Our study provides evidence for a novel mechanism for hnRNP F lowering of SBP: inhibition of intrarenal Agt gene expression and RAS activation, concomitant with upregulation of the ACE-2/Ang 1-7/MasR axis. Indeed, our results show that hnRNP F overexpression inhibited renal AGT and Agt mRNA expression (ESM Fig. 1 c-e), lowered urinary AGT and Ang II levels and normalised urinary Ang 1-7 levels.
The precise mechanism by which hnRNP F overexpression leads to upregulation of renal Ace-2 and MasR gene expression in diabetes remains unclear. One possibility is that hnRNP F binds to putative Hnrnpf-RE(s) in the Ace-2 and MasR gene promoters, subsequently enhancing Ace-2 and MasR gene transcription. This possibility is supported by our findings that hnRNP F considerably augments the activity of an Ace-2 gene promoter and that the Hnrnpf siRNA and deletion of the putative Hnrnpf-RE markedly reduced the rat Ace-2 gene promoter activity in RPTCs. Furthermore, the biotinylated-labelled Hnrnpf-RE specifically bound to RPTC nuclear proteins and the addition of anti-hnRNP F antibody yielded a supershift of biotinylated-labelled Hnrnpf-RE binding with nuclear proteins in EMSA. These data demonstrate that hnRNP F binds to the putative Hnrnpf-RE and stimulates Ace-2 gene transcription. Of note, hnRNP F is not specific for  ---+ + + + + + + + + + + -BSA + + + + + + + + + Ace-2 gene expression but also affects the expression of Agt [25] and other genes [51,52].
Currently, little is known about the mechanisms by which TGF-β1 downregulates renal Ace-2 gene expression in diabetes. Chou et al [53] reported that SB431542 inhibited high glucose and TGF-β1 inhibition of Ace-2 mRNA expression in cultured NRK-52 cells. Our findings confirm these observations. Our present studies also demonstrate that TGF-β1 inhibits the activity of pGL 4.20-rat Ace-2 promoter (N-1, 091/+83) and that deletion of putative SMAD-REs in the Ace-2 gene promoter mitigates the inhibitory effect of TGF-β1 on the Ace-2 gene promoter activity. Furthermore, biotinylated-labelled SMAD-REs bound to RPTC nuclear proteins and the addition of anti-Smad2/3 antibody yielded a supershift of labelled DNA with nuclear proteins. These data demonstrate that the inhibitory effect of TGF-β1 on Ace-2 gene transcription is mediated, at least in part, via the SMAD-REs in the Ace-2 gene promoter.
Intriguingly, hnRNP F overexpression prevented TGF-β1 signalling on Smad2/3 phosphorylation and on TGF-β1 inhibition of Ace-2 gene promoter activity in RPTCs. At present, the underlying molecular mechanism of how hnRNP F prevents TGF-β1 inhibition of Ace-2 gene transcription is not yet defined. One possibility might be that hnRNP F directly inhibits Tgf-β1rII gene expression as shown in our studies. The second possibility is that hnRNP F might interfere or prevent the interaction of Smad2/3 with other transcriptional factor(s) to inhibit Ace-2 gene transcription. Clearly, more studies are needed to define the molecular interaction of hnRNP F with Smad2/3 on Ace-2 gene transcription.
In summary, the present study suggests a major role for hnRNP F in attenuating systemic hypertension and renal fibrosis in experimental diabetes and possibly in diabetic human kidneys. Our observations raise the possibility that selective targeting of this antihypertensive and anti-fibrotic protein may represent a novel approach for preventing or reversing the pathological manifestations of DN, particularly tubular fibrosis.