Tubular Deficiency of Heterogeneous Nuclear Ribonucleoprotein F Elevates Systolic Blood Pressure and Induces Glycosuria in Mice

We reported previously that overexpression of heterogeneous nuclear ribonucleoprotein F (Hnrnpf) in renal proximal tubular cells (RPTCs) suppresses angiotensinogen (Agt) expression, and attenuates systemic hypertension and renal injury in diabetic Hnrnpf-transgenic (Tg) mice. We thus hypothesized that deletion of Hnrnpf in the renal proximal tubules (RPT) of mice would worsen systemic hypertension and kidney injury, perhaps revealing novel mechanism(s). Tubule-specific Hnrnpf knockout (KO) mice were generated by crossbreeding Pax8-Cre mice with floxed Hnrnpf mice on a C57BL/6 background. Both male and female KO mice exhibited elevated systolic blood pressure, increased urinary albumin/creatinine ratio, tubulo-interstitial fibrosis and glycosuria without changes in blood glucose or glomerular filtration rate compared with control littermates. However, glycosuria disappeared in male KO mice at the age of 12 weeks, while female KO mice had persistent glycosuria. Agt expression was elevated, whereas sodium-glucose co-transporter 2 (Sglt2) expression was down-regulated in RPTs of both male and female KO mice as compared to control littermates. In vitro, KO of HNRNPF in human RPTCs (HK-2) by CRISPR gRNA up-regulated AGT and down-regulated SGLT2 expression. The Sglt2 inhibitor canagliflozin treatment had no effect on Agt and Sglt2 expression in HK-2 and in RPTCs of wild-type mice but induced glycosuria. Our results demonstrate that Hnrnpf plays a role in the development of hypertension and glycosuria through modulation of renal Agt and Sglt2 expression in mice, respectively.

Physiological measurements in Hnrnpf KO mice. Deletion of renal tubular Hnrnpf did not influence body weight gain nor the non-fasting blood glucose levels in both male and female mice from the age of 6 to 24 weeks (Supplemental Fig. 1c-f, respectively). Longitudinal SBP measurements revealed consistently higher SBP in both male ( Fig. 2A) and female (Fig. 2B) KO mice aged week 6 to 24 compared to Ctrl. Significant increases of Agt mRNA and protein expression were detected in both male and female KO mice compared to Ctrl at 8 weeks (Supplemental Fig. 2a) and 24 weeks of age (Fig. 2C,D, respectively). No significant difference of Agt expression in RPTs was observed between male and female Ctrl as well as between male and female KO mice. These were confirmed with immunostaining (Fig. 2E).
Increased urinary Ang II and urinary albumin/creatinine ratio (ACR) were also observed in both male and female KO mice compared to Ctrl at 24 weeks of age with no significant difference between male and female Ctrl as well as between male and female KO mice. (Fig. 2F,G, respectively). In contrast, body weight (BW), kidney weight (KW)/BW, GFR and glomerular tuft volume did not differ significantly between KO mice and Ctrl at 24 weeks of age (Table 1). Twenty-four h urine volume were significantly increased but not food and water intake in both male and female KO mice as compared to Ctrl. No differences were detected in serum and urine levels of sodium, calcium and phosphorus between male and female Ctrl and KO mice. We detected no significant differences in serum Ang II among different groups of mice (Fig. 2H). Glycosuria and Sglt2 expression in Hnrnpf KO mice. Unexpectedly, increased glucose excretion was detected in the urine using dipsticks in both male and female KO mice at age 6 weeks (Fig. 4A). From 8 weeks of age, urinary glucose levels in male KO mice steadily decreased and returned to levels similar to Ctrl mice at 12 weeks of age (Fig. 4B). In contrast, urinary glucose excretion steadily increased from week 6 in female KO mice, reached an apparent plateau at the age of 12 weeks and did not abate (Fig. 4B). No changes in urinary glucose level were detected in male and female Ctrl.

Tubulo-interstitial fibrosis in
Having observed that both male and female Hnrnpf KO mice develop glycosuria, we performed intraperitoneal glucose tolerance test (IPGTT) at the age of 23 weeks in male and female mice (Fig. 4C). KO of Hnrnpf in RPTs did not influence the glucose tolerance in either male or female KO mice.
RT-qPCR revealed lower Sglt2 expression in RPTs isolated from both male and female KO mice at 8 weeks of age (Supplemental Fig. 2b) and 24 weeks of age (Fig. 4D) as compared with Ctrl. At both 8 and 24 weeks of age, Sglt2 expression decreased by ~40% in RPTs of male Hnrnpf KO mice as compared to Ctrl, whereas persistently lower Sglt2 expression (decreased by ~60% of baseline level) was observed in RPTs from female KO mice. However, no significant difference of Sglt2 expression in RPTs was observed between male and female Ctrl as well as between male and female KO mice. WB of isolated RPTs confirmed these changes at the protein level (Fig. 4E). Consistently, semi-quantitation of immunofluorescence staining with anti-Sglt2 antibodies and LTL-FITC confirmed reduced Sglt2 expression in RPTs of 24 week-old KO mice as compared to Ctrl (Fig. 4F,G, respectively).
www.nature.com/scientificreports www.nature.com/scientificreports/ No significant changes were detectable in Slc5a1 (Sglt1) mRNA expression in RPTs isolated from both male and female KO mice as compared to Ctrl (Supplemental Fig. 2c).

Effect of canagliflozin treatment on Agt and Sglt2 expression in RPTCs in vivo.
To investigate the role of Sglt2 on Agt and Sglt2 expression in RPTCs in vivo, wild-type mice were treated with the selective Sglt2 inhibitor canagliflozin (0.2 mg/ml in drinking water). Four weeks of canagliflozin treatment had no detectable AGT and SGLT2 expression in HK-2 with or without HNRNPF KO. To validate our in vivo observations, we generated HK-2 cells with HNRNPF KO by CRISPR gRNA technology. Consistent with our in vivo observation, immunoblots revealed that HK-2 cells with HNRNPF KO exhibited non-detectable HNRNPF (Fig. 6A,B), higher AGT ( Fig. 6A,C) and lower SGLT2 protein expression (Fig. 6A,D) as compared to control HK-2. These findings were confirmed by RT-qPCR of HNRNPF (Fig. 6E), AGT (Fig. 6F) and SGLT2 expression (Fig. 6G), respectively. Finally, in human cells expression of HNRNPF (Fig. 6H,I), AGT (Fig. 6H,J) and SGLT2 protein (Fig. 6H,K) and mRNA ( Fig. 6L-N, respectively) did not differ significantly in HK-2 cells treated with canagliflozin and untreated cells, indicating a lack of causality of inhibition of SGLT2 activity and AGT and SGLT2 expression in RPTCs.

Discussion
Our results identify a novel mechanism by which Hnrnpf affects the development of hypertension and glycosuria in mice through modulation of intrarenal Agt and Sglt2 expression, respectively.
Hnrnpf, a member of the 30 pre-mRNA-binding protein family, modulates gene expression at both transcriptional and post-transcriptional levels [11][12][13][14][18][19][20][21] . Hnrnpf engages in alternative splicing of various genes and associates with TATA-binding protein, RNA polymerase II, nuclear cap-binding protein complex and various transcription factors to modulate gene expression 22 . We have reported previously that Hnrnpf overexpression in RPTCs attenuates hypertension and kidney injury in both diabetic Akita 13 and db/db 14 mice via inhibition of intrarenal Agt expression, implying an important role for Hnrnpf in modulating the development of hypertension and nephropathy in diabetic mice.
Our present findings document that genetic deletion of Hnrnpf in tubules enhances renal Agt expression, hypertension development and kidney injury in both non-diabetic male and female mice. These observations are consistent with our hypothesis that Hnrnpf plays an important role in the development of hypertension and tubulo-interstitial fibrosis via modulation of Agt and pro-fibrotic genes expression in RPTs.
Initially, we generated global Hnrnpf KO mice by cross-breeding a general Cre-deleter mouse line (CMV-Cre; B6.C-Tg(CMV-cre)1Cgn/J) with our Hnrnpf fl/fl mice on a C57BL/6 background to explore the phenotype of global Hnrnpf KO mice and found that like the global Hnrnpk deletion 23 , global Hnrnpf KO also results in embryonic death (Supplemental Fig. 3). To circumvent this issue, we generated renal tubule-specific Hnrnpf KO mice by cross-breeding our Hnrnpf fl/fl mice with a renal tubule-specific Cre deleter (Pax8-Cre; B6.129P2(Cg)-Pax8 tm1.1(cre)Mbu /J) mouse line 17 . Several labs have also successfully employed Pax8-Cre mice to delete genes in renal tubules [24][25][26] . Our homozygous Pax8-Hnrnpf KO mice are viable and fertile without symptoms of body weight loss, physiological imbalance and altered hearing. However, they develop hypertension and elevated ACR with increased Agt expression in RPTCs by 8 weeks of age (Supplemental Fig. 4a,b). Since Pax8 is also expressed in the thyroid gland and hindbrain 17 , it is possible that Pax8-Hnrnpf KO mice might exhibit abnormality in thyroid gland and hindbrain thereby indirectly affecting cardiac and renal function. However, we did not detect significant changes in Hnrnpf mRNA levels or serum T 4 levels in Pax8-Hnrnpf KO mice (Supplemental Fig. 5) or histological changes in thyroid gland and hindbrain. Thus, our Pax8-Hnrnpf KO mouse appears to be a valid murine model with which to study the phenotype with tubule-specific Hnrnpf KO.
An unexpected finding of our present study was that Hnrnpf deletion led to glycosuria with reduced expression of Sglt2 in RPTs of Hnrnpf KO mice. Intriguingly, serum and urine levels of Na, Ca and P did not differ www.nature.com/scientificreports www.nature.com/scientificreports/ between Hnrnpf KO mice and Ctrl. The phenotype of glycosuria appears to be similar to that reported in Sglt2 deficient mice 27 and in patients with familial renal glycosuria (FRG) [28][29][30] but differs from that in Sweet Pee mice, which are characterized by elevated urinary excretion of calcium and magnesium and growth retardation 31 , as well as from that in patients with renal Fanconi syndrome 32 . Intriguingly, male Hnrnpf KO mice exhibited a transient glycosuria between the ages of 6 and 12 weeks and then returned to levels similar to Ctrl. Furthermore, glycosuria correlates with reduced Sglt2 expression in RPTs of male Hnrnpf KO mice. In contrast, female Hnrnpf KO mice displayed persistent glycosuria throughout 6 to 24 weeks of age with similar inhibition of Sglt2 expression www.nature.com/scientificreports www.nature.com/scientificreports/ To explore the impact of Sglt2 inhibition on Agt expression, we treated HK-2 cells and WT mice with canagliflozin. Canagliflozin had no detectable effects on SBP and blood glucose levels but enhanced the development of glycosuria in both male and female mice as compared to non-treated mice. Furthermore, canagliflozin treatment had no effect on the expression of Agt and Sglt2 expression in RPTs of mice. Thus, our data would argue against www.nature.com/scientificreports www.nature.com/scientificreports/ a causal relationship between Sglt2 inhibition and Agt expression in RPTCs; rather our data would indicate that inhibition of Hnrnpf expression modulates both Agt and Sglt2 expression in RPTCs.
Finally, to replicate our in vivo observations, we studied a human renal proximal tubular cell line (HK-2) 35 . By employing CRISPR gRNA technology, we obtained several clones of HK-2 cells with HNRNPF KO. Consistent with our findings in Hnrnpf KO mice, HK-2 with HNRNPF KO displayed significantly higher AGT and lower SGLT2 expression as compared to HK-2 controls. These data lend further support our previous observations that Hnrnpf down-regulates RPT Agt expression and RAS activation, leading to improve tubulo-interstitial fibrosis in the kidney. Moreover, consistent with our in vivo results, canagliflozin treatment had no effect on SGLT2 and AGT expression in HK-2 cells. www.nature.com/scientificreports www.nature.com/scientificreports/ At present, the underlying mechanism(s) by which genetic deletion of HNRNPF led to down-regulation of SGLT2 transcription in HK-2 cells are unclear. One possibility might be that HNRNPF affects SGLT2 transcription at the promoter activity level. This is unlikely since transfection of the HNRNPF cDNA did not affect the SGLT2 promoter activity (pGL4.2/SGLT2-N-1,986/+22 promoter) in HK-2 cells (Supplemental Fig. 6a). However, we could not rule out the possibility of putative HNRNPF-response element(s) upstream of 2 kb of the SGLT2 promoter. The second possibility is that HNRNPF deletion might alter the splicing of SGLT2 to yield mutant forms of SGLT2. This is also unlikely since only one species of SGLT2 was detectable in HK-2 cells with HNRNPF KO, which was similar to the size of SGLT2 in HK-2 (Supplemental Fig. 6b). The third possibility is that HNRNPF might affect SGLT2 mRNA stability. This notion is supported by the observations of Chu et al. 19 that HNRNPF regulates YAP expression via binding to the 3′UTR of YAP to affect its mRNA stability and of Decorsiere et al. 21 that Hnrnph/f interacts with a G-quadruplex in maintaining p53 pre-mRNA 3′-end processing during DNA damage. The fourth possibility is that deletion of HNRNPF might suppress other un-defined signaling pathway(s) or factor(s) that might have a greater impact (stimulation) on SGLT2 expression and activity. Clearly, further studies are needed to elucidate the mechanisms underlying HNRNPF down-regulation of SGLT2 expression.
The exact mechanism(s) of Hnrnpf regulation of Agt expression is unknown. One possibility is that Hnrnpf binds to the insulin-responsive element (IRE) in the Agt promoter 11,12 and functions as a negative trans-acting protein to inhibit the binding of other positive trans-acting factor(s) to TATA-binding protein (TBP) and RNA polymerase II, subsequently attenuating Agt transcription. This possibility is supported by the studies of Yoshida et al. 22 showing that Hnrnpf is associated with TBP, RNA polymerase II and nuclear cap-binding protein complex. A second possibility is that Hnrnpf is associated with Hnrnpk to form an Hnrnpf/k complex and that the Hnrnpf/k complex is more effective in inhibiting Agt transcription. Indeed, we have previously reported that Hnrnpf co-immunoprecipitated with Hnrnpk and that co-transfection of Hnrnpf with HnrnpK was more effective in inhibiting Agt transcription than either Hnrnpf or HnrnpK alone 35 . A third possibility is that the Hnrnpf/k complex might recruit unidentified repressor molecules and subsequently repress Agt transcription. This third possibility is supported by the studies of Denisenko et al. 36 demonstrating that Hnrnpk could bind the murine repressor Zik1. Clearly, more work is needed to elucidate the precise molecular mechanism of action of Hnrnpf on Agt transcription in RPTCs.
In summary, the present study reveals a novel role for Hnrnpf in the development of hypertension, tubule-interstitial fibrosis and glycosuria in mice via up-regulation of Agt and down-regulation of Sglt2 expression in RPTs, respectively. With the recent development of SGLT2 inhibitors as a novel treatment for diabetic patients [37][38][39][40][41] , it would be important to understand the regulation of SGLT2 expression at the molecular level. Our findings raise the possibility that Hnrnpf KO mice may be a useful animal model for advancing studies on SGLT2 regulation and familial renal glycosuria in human.  Table 1. Restriction and modifying enzymes were purchased from New England Biolabs (Whitby, ON, Canada). The sources of antibodies used are listed in Supplemental Table 2. HK-2 (an immortalized human renal proximal tubular cell line) (Cat. No. CRL-2190) was obtained from American Tissue Cell Collection (ATCC) (Manassas, VA) (http://www.atcc.org). Human SGLT2 gene promoter (N-1,986/+22) was amplified from HK-2 genomic DNA by PCR with specific primers (Supplemental Table 1) and then inserted into pGL4.20 reporter vector (Promega, Sunnyvale, CA) at Xho1and Bgl II restriction sites.

Chemical and reagents.
Generation of tubular Hnrnpf KO mice. Tubule-specific Hnrnpf KO mice were generated by cross-breeding male Hnrnpf-floxed mice with female Pax8-Cre mice 17 (Stock number: 028196; Jackson Laboratory, Bar Harbor, ME). Briefly, the mouse Hnrnpf gene (Gene ID: 98758) is localized on chromosome 6: 117,900,340-117,925,622. Four exons have been identified in Hnrnpf with the ATG start codon and TAG stop codon both located in exon 4. The lox-modified Hnrnpf targeting vector was created by including 5′ and 3′ homology arms as well as two loxP sites flanking the fourth exon region amplified from SV129 BAC genomic DNA and confirmed by sequencing. C57BL/6 ES cells were used for gene targeting (Cyagen Biosciences, Santa Clara, CA). These mice allow the excision of exon 4 of Hnrnpf gene and disruption of the protein expression in the presence of Cre recombinase. By cross-breeding male Hnrnpf-floxed mice with female Pax8-Cre mice, heterozygous Hnrnpf-floxed allele mice were generated (genotype: Hnrnpf fl/wt ,Cre). These mice were back-crossbred to generate homozygous Hnrnpf-floxed allele and carrying the Cre allele (genotype: Hnrnpf fl/fl ,Cre). The Pax8-Hnrnpf KO mice (genotype: Hnrnpf fl/fl ,Cre) and control littermates (Ctrl) (genotype: Hnrnpf fl/fl ) as well as heterozygous littermates (genotype: Hnrnpf fl/wt ,Cre and Hnrnpf fl/wt ) were used in the present studies. Experimental mice were generated from at least three different breeding couples. Offspring were genotyped by PCR to detect the Cre-recombinase as well as the presence or absence of the 5′ loxP site using specific primers (Supplemental Table 1). www.nature.com/scientificreports www.nature.com/scientificreports/ Weekly random blood glucose levels were measured in mice by Accu-Chek Performa (Roche Diagnostics, Laval, QC, Canada). SBP was measured with BP-2000 tail-cuff (Visitech Systems, Apex, NC) at least 2 to 3 times per week per animal in the morning without fasting as previously described 13,14 . Baseline SBP was measured daily over a 5-day period before initiation of actual measurement at 6 weeks of age.

Physiological studies.
At 24 weeks of age, twenty-four h prior to euthanasia, mice were housed individually in metabolic cages. Food, water consumption, and urine output were recorded. Mouse serum and urine samples were extracted with C18 Sep-Pak columns (Waters, Mississauga, ON) and assayed for Ang II by specific ELISA (Bachem America, Torrence, CA) according to the recommended number III protocol 13,14,20,43 . Urines were also assayed for levels of albumin and creatinine (ELISA, Albuwell and Creatinine Companion, Exocell, Inc., Philadelphia, PA) 13,14 and glucose (Glucose colorimetric kit, Cayman Chemical, Ann Arbor, MI).
For tissue studies, mice were euthanized at the age of 8 or 24 weeks. Blood samples were collected by cardiac puncture. The kidneys were isolated, decapsulated and weighed. The left kidneys were processed for histology and immunostaining, and the right kidneys were used for isolation of renal proximal tubules (RPTs) by Percoll gradient 13,14,20,42,43 . Aliquots of freshly-isolated RPTs from individual animals were used immediately for total RNA isolation and Western blotting.
Serum and urine biochemical measurements. Serum  Intraperitoneal glucose tolerance Test. An intraperitoneal glucose tolerance test (IPGTT) was performed after 6 h fasting in non-anesthetized mice at the age of 23 weeks, as described previously 44 .
Details of the sources of antibodies and working dilutions are listed in Supplemental Table 2.
Masson's Trichrome staining, Sirius Red staining and immunostaining for Fn1 were performed to assess tubule-interstitial fibrosis. Semi-quantitation of the relative staining was done by NIH Image J software (http:// rsb.info.nih.gov/ij/). Mean glomerular tuft and RPTC volumes, and the tubular luminal area were determined by the methods of Weibel 45 and Gundersen 46 , as described previously 47,48 . Immunofluorescence staining. Immunofluorescence (IF) staining was performed on 3-μm tissue sections from mouse kidney fixed in formalin and embedded in paraffin followed by staining with ALEXA FLUOR-594-labeled secondary antibody (Invitrogen). Proximal tubules were identified by fluorescein-labeled lotus tetragonolobus lectin (LTL, a marker of renal proximal tubule 49 ) (Vector Labs, Burlingame, CA). Image quantification and merge were assessed by ImageJ software (http://rsb.info.nih.gov/ij/). To quantify the amount of Sglt2 expression, the pixel intensity of Sglt2 was divided by LTL intensity. To calculate the average ratio, 6 sections per mouse, 6 mice per group were analyzed. A35509, CRISPR1099776_CR, Invitrogen):tracrRNA (Cat. No. A35506, Invitrogen)) were transfected to HK-2 and cultured for 2 days at 37 °C. Single cell clones were then isolated by using limiting dilution cloning in 96-well plates. The positive clones were identified for the absence of HNRNPF by WB of cellular extracts and confirmed by PCR of genomic sequence. The clones with HNRNPF expression were used as controls.

Human renal proximal tubular cells with or without
To test the pharmacologic effect of SGLT2 inhibition on SGLT2 and AGT expression, HK-2 cells were harvested after 24 hours of culture in serum-free normal glucose (5 mM) DMEM in the absence or presence of 0.5 mM canagliflozin as described by Pirklbauer et al. 51 . WB and RT-qPCR were used to quantify SGLT2 and AGT protein and mRNA expression, respectively.
Canagliflozin treatment in wild-type (WT) mice. To investigate the impact of Sglt2 inhibition and Agt expression in RPTCs in vivo, male and female WT mice were treated with or without canagliflozin (0.2 mg/ml in drinking water) at the age of 4 weeks as described previously 52 . Body weight, blood and urinary glucose and SBP