PU.1 inhibition attenuates atrial fibrosis and atrial fibrillation vulnerability induced by angiotensin‐II by reducing TGF‐β1/Smads pathway activation

Abstract Fibrosis serves a critical role in driving atrial remodelling‐mediated atrial fibrillation (AF). Abnormal levels of the transcription factor PU.1, a key regulator of fibrosis, are associated with cardiac injury and dysfunction following acute viral myocarditis. However, the role of PU.1 in atrial fibrosis and vulnerability to AF remain unclear. Here, an in vivo atrial fibrosis model was developed by the continuous infusion of C57 mice with subcutaneous Ang‐II, while the in vitro model comprised atrial fibroblasts that were isolated and cultured. The expression of PU.1 was significantly up‐regulated in the Ang‐II‐induced group compared with the sham/control group in vivo and in vitro. Moreover, protein expression along the TGF‐β1/Smads pathway and the proliferation and differentiation of atrial fibroblasts induced by Ang‐II were significantly higher in the Ang‐II‐induced group than in the sham/control group. These effects were attenuated by exposure to DB1976, a PU.1 inhibitor, both in vivo and in vitro. Importantly, in vitro treatment with small interfering RNA against Smad3 (key protein of TGF‐β1/Smads signalling pathway) diminished these Ang‐II‐mediated effects, and the si‐Smad3‐mediated effects were, in turn, antagonized by the addition of a PU.1‐overexpression adenoviral vector. Finally, PU.1 inhibition reduced the atrial fibrosis induced by Ang‐II and attenuated vulnerability to AF, at least in part through the TGF‐β1/Smads pathway. Overall, the study implicates PU.1 as a potential therapeutic target to inhibit Ang‐II‐induced atrial fibrosis and vulnerability to AF.


| INTRODUC TI ON
Atrial fibrillation (AF) can result in heart failure, stroke and increased cardiac mortality and morbidity. 1,2 Atrial remodelling serves an important role in the pathological mechanism of AF, and atrial fibrosis is the crucial substrate for atrial remodelling. [3][4][5] Therefore, attenuating atrial fibrosis is considered an effective strategy for preventing the occurrence and maintenance of AF.
Transforming growth factor-β1 (TGF-β1), an important fibrogenic factor, binds its type II receptor (TβRII) and type I receptor (TβRI) to activate a cascade of phosphorylation reactions that promote inactivation of the Smad proteins 2, 3 and 4, before finally translocating to the nuclei and regulating the expression of profibrotic proteins, including connective tissue growth factor (CTGF) and collagen. 1,6 There is substantial evidence that angiotensin-II (Ang-II) can up-regulate TGF-β1 and collagen expression in vitro and in vivo, and Ang-II plays an important role in atrial fibrosis and AF through its regulation of TGF-β1. [7][8][9][10][11] Therefore, inactivation of the TGF-β1/ Smads pathway attenuates the atrial fibrosis induced by Ang-II.
PU.1, a member of the ETS family of transcription factors, is a vital regulator of gene expression, particularly in macrophages, B cells, T cells and dendritic cells. 12,13 PU.1 binds its recognition motifs to facilitate DNA methyltransferase activity, chromatin opening and the binding of other transcription initiation factors. 14,15 PU.1 is also a reprogramming factor that converts fibroblasts or neural stem cells into monocytes, macrophages and dendritic cells, [16][17][18][19][20] and deregulation of PU. 1 is considered a crucial contributor to leukaemia pathogenesis. 21,22 Research has revealed that PU.1 is up-regulated in the fibroblasts in various fibrotic diseases. [23][24][25][26][27][28] A recent report has found that the majority of PU.1 + cells in fibrotic tissues are fibroblasts; in contrast, PU.1 + fibroblasts are not found in normal or inflamed tissues of the skin, lung, liver, kidney or joints, as TGF-β1 promotes PU.1 expression in fibrotic fibroblasts but not in resting fibroblasts or inflammatory fibroblasts; interestingly, Smad3 inhibition significantly reduces the PU.1 expression induced by TGF-β1 in fibrotic fibroblasts. 26 These findings support the hypothesis that PU.1 can control the fibroblast polarization and tissue fibrosis mediated by the TGF-β1/Smads pathway. However, whether PU.1 regulates Ang-II-induced atrial fibrosis via the TGF-β1/Smads pathway remains unclear. Based on these observations, we developed a mouse model of atrial fibrosis induced by continuous subcutaneous Ang-II infusion to reveal whether PU.1 inhibition attenuates Ang-IIinduced atrial fibrosis and vulnerability to AF via the TGF-β1/Smads pathway.

| Animals
All animal experiments were approved by the Animal Care and Use Committee of Renmin Hospital at Wuhan University and performed in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (8th Edition, NRC 2011). Male C57BL/6 mice were housed under standard conditions with a controlled temperature, humidity and 12-h light/ dark cycle. Food and water were provided ad libitum. An osmotic minipump (Model 1002; Alzet, Cupertino, CA, USA) was implanted in the mice for the subcutaneous infusion of Ang-II (750 ng/kg/min) or phosphate-buffered saline (sham) for 28 days. 8 The mice were divided into four groups: sham (no Ang-II induction), sham+DB1976 (a PU.1 inhibitor), Ang-II and Ang-II+DB1976. Mice in the sham and Ang-II groups were given saline, while mice in the sham+DB1976 group and Ang-II+DB1976 group were administered with DB1976 (5 mg/day/kg; Glpbio, Montclair, CA, USA) by intraperitoneal injection once a day for 28 days, as previously described. 26 Mice were anaesthetized with 50 mg/kg sodium pentobarbital (Merck KGaA, Darmstadt, Germany) prior to the assays.

| Western blot analysis
Proteins were extracted from the frozen atrium tissues, and the protein concentrations were determined using the Bicinchoninic Acid Protein Assay Kit (Abcam, Cambridge, UK) according to previously described. 29  Abcam) and anti-GAPDH (1:1000; Abcam). Finally, the membranes were incubated with the secondary horseradish peroxidase (HRP)-goat anti-rabbit (AS1107; 1:10 000; Aspen Biotechnology, Bedford, MA) or HRP-goat anti-mouse (AS1106; 1:10 000; Aspen Biotechnology) at room temperature for 1 hr and exposed to enhanced chemiluminescence (ECL) substrate (AS1027; Aspen Biotechnology). Western blot results were quantified by densitometry (Image Lab, Hercules, CA). The relative intensities of the proteins of interest were normalized to that of the sham/control group, which was set to a value of 1 (100%).

| Histological studies
The heart of each mouse was removed, fixed in 4% paraformaldehyde and embedded in paraffin. Atrium tissue sections (4 μm thick) were stained with Masson's trichrome, visualized with an Olympus BX51 microscope (Olympus Corporation, Tokyo, Japan) and analysed using Image-Pro Plus 6.0 software (Media Cybernetics, Inc, Rockville, MD, USA). The percentage of fibrosis was measured by calculating the ratio of fibrotic tissue area to that of the normal myocardial tissue.

| Isolation, culture and treatment of atrial fibroblasts
Atrial fibroblasts from the Ang-II-infused mice were isolated as previously described 8

| Atrial fibroblast proliferation assay
The proliferation of atrial fibroblasts isolated from Ang-II-infused mice was measured with the Cell Counting Kit-8 (CCK-8; Dojindo Molecular Technologies, Kumamoto, Japan) according to the manufacturer's instructions. 30 Atrial fibroblasts were plated in 96-well plates, serum-starved for 24 h and exposed to Ang-II, DB1976, siRNA-Smad3 or Ad-PU.1 according to our assays. Following incubation, CCK-8 solution was added for 4 h and the absorbance at 450 nm (OD 450 ) was read with a Sunrise microplate reader (Tecan Group, Ltd., Männedorf, Switzerland).

| Immunofluorescence
The atrium tissue samples were fixed and sectioned for analysis as previously described. 31 The atrium tissue sections were incubated

| Electrophysiological studies using isolated perfused hearts
Langendorff-perfused hearts were isolated as previously described. 32 Electrophysiological studies on the perfused hearts were CO 2 at 37℃ at a constant pressure of 60 mmHg to evaluate the induction of AF as previously described. 33 Each heart was perfused until stabilization for 20 min before conducting the programmed electrophysiology tests. Hearts that exhibited irreversible myocardial ischaemia or did not recover to a regular spontaneous rhythm were discarded. 34 Teflon-coated (except at the tips) silver bipolar electrodes were placed on the appendages of the right atrium and left atrium. AF inducibility was tested with the burst pacing method, as previously described. 8 AF was defined as a rapid irregular atrial rhythm with irregular RR intervals lasting for at least 5 s. The duration of AF was measured from the end of burst pacing to the first P wave detected after the rapid irregular atrial rhythm.

| Statistical analysis
All data are presented as the mean ±standard deviation (SD).
Student's t test was used to determine differences between two groups, and one-way analysis of variance (ANOVA) was used for multiple group comparisons. All data analysis was performed with SPSS Statistics software v. 20.0 (IBM Corp., Armonk, NY, USA). A P value of <0.05 was considered statistically significant.

| PU.1 expression in atrial tissue and atrial fibroblasts is induced by Ang-II in vivo and in vitro
To assess the effect of PU.1 on Ang-II-mediated fibrosis, PU.1 expression following induction by Ang-II was assessed in atrial tissue and atrial fibroblasts in vivo and in vitro, respectively. qRT-PCR, Western blotting and immunofluorescence staining analyses demonstrated that PU.1 expression was significantly higher (P < .01) in the Ang-II-induced group compared with the sham group in vivo ( Figure 1A-D). Next, we isolated atrial fibroblasts from mice subjected to the subcutaneous infusion of Ang-II and performed Western blot and immunofluorescence analyses of PU.1 expression. PU.1 expression was significantly greater (P < .01) in the Ang-II-stimulated group compared with the control group ( Figure 1E-F).
Immunofluorescence staining indicated that double-positive cells of PU.1 and α-SMA (specific marker of myofibroblasts) expression were also increased in the induced atrial fibroblasts ( Figure 1G).
These findings suggest that PU.1 is associated with Ang-IImediated atrial fibrosis.

| Inhibition of PU.1 down-regulated activation of the atrial TGF-β1/Smads pathway induced by Ang-II in vivo and in vitro
The Smads signalling pathway plays a key role in regulating tissue fibrosis, and Ang-II is considered to induce the synthesis and release of fibrosis factors via the TGF-β1/Smads pathway. 11 A recent report suggested that PU.1 promotes skin, lung, liver, joint and kidney fibrosis by regulating the activity of the TGF-β1/Smad3 pathway. 26 However, whether PU.1 participates in the atrial fibrosis induced by Ang-II via the TGF-β1/Smads pathway is unknown. Therefore, F I G U R E 1 PU.1 expression in mice atrial tissue or fibroblasts was induced by Ang-II in vivo and in vitro, respectively. The expression of PU.1 was evaluated by qRT-PCR, Western blotting and immunofluorescence assays of atrial tissue following induction by subcutaneous infusion with Ang-II for 28 days and in cultured atrial fibroblasts sourced from the Ang-II-induced mice with or without Ang-II for 24 hours.

| Inhibition of PU.1-attenuated Ang-II-induced atrial fibroblast proliferation and differentiation in vivo and in vitro
To determine the effect of PU.1 on atrial fibroblast proliferation and differentiation in atrial tissue isolated from mice induced with Ang-II, double immunofluorescence staining was performed for proliferating cell nuclear antigen (PCNA). As shown in Figure 3A, the number of PCNA-positive cells was lower in the Ang-II+DB1976 group compared with the Ang-II group by immunofluorescence staining in vivo ( Figure 3A). Moreover, Western blotting indicated that the expression of PCNA and a-SMA was significantly higher (P < .01) in the Ang-II-induced group than in the sham group, and these increases were attenuated by PU.1 inhibition in the Ang-II + DB1976 group ( Figure 3B-E, P < .05). Next, cultured atrial fibroblasts isolated from Ang-II-induced mice were subjected to Western blotting, immunofluorescence staining and CCK-8 assays to assess cell proliferation and differentiation. Especially, α-SMA, SMemb and ED-A Fn are main markers of myofibroblast 35 ; in addition, DDR2 determines a-SMA-dependent collagen gene expression in the Ang-II-induced cardiac fibroblasts. 36 Therefore, we measured the above indicators, and the results further validated the in vivo findings ( Figure 3F-N).
To further confirm the above effects, PU.1 was knocked down in the cultured atrial fibroblasts using siRNA. The results showed that the levels of PU.1, PCNA, α-SMA, SMemb, ED-A Fn and DDR2 were significantly decreased (P < .05) in Ang-II-induced atrial fibroblasts transfected with si-PU.1 ( Figure 3O-W). These findings suggest that PU.1 inhibition attenuates Ang-II-induced atrial fibroblast proliferation and differentiation.

| Inhibition of PU.1-attenuated Ang-II-induced atrial fibrosis in vivo and in vitro
To examine the effects of PU.1 on Ang-II-induced atrial fibrosis, Ang-II-induced mice and cultured atrial fibroblasts sourced from these mice were exposed to DB1976. Masson's trichome staining of the atrial tissue showed that DB1976 exposure significantly decreased (P < .05) the interstitial, sub-epicardial and perivascular fibrosis induced by Ang-II in the atrial tissue ( Figure 4A, B), and Western

| Inhibition of PU.1-blunted electrical remodelling of the atrium in Ang-II-induced mice
Based on the observation that PU.1 inhibition attenuated Ang-IIinduced atrial fibrosis, we hypothesized that PU.1 modulates cardiac electrophysiological properties. Therefore, we performed an in vivo electrophysiology study of Langendorff-perfused hearts using  3). Data represent the mean ± SD. *P < .01 vs. the sham/control group, # P < .05 vs. the Ang-II group/Ang-II + si-Con group. GAPDH was used as the internal control. Ang-II, Angiotensin-II; si-PU.1, small interfering RNA against PU.1 a method previously described. 32,33 The resulting data revealed that vulnerability to AF and AF duration were significantly higher (P < .01) in the Ang-II-induced group than in the sham group, whereas these parameters in the DB1976-exposed group were significantly lower (P < .05) than in the Ang-II group ( Figure 5A-E). These findings indicate that PU.1 inhibition in vivo decreases Ang-II-induced AF vulnerability and AF duration by at least in part through decreasing atrial fibrosis.

| D ISCUSS I ON
The present study explored the effect of PU.1 on atrial fibrosis and vulnerability to atrial fibrillation (AF). The study findings indicated In clinical AF, almost 70% of patients present with AF secondary to underlying coronary heart disease, hypertension, rheumatic heart disease, or valvular disease, and in these patients, atrial fibrosis increases AF inducibility and maintains AF. [39][40][41] We have known that the tissue renin-angiotensin system (RAS) hyperactivity and high expression are involved in the pathomechanism of cardiac fibrosis 42 ; moreover, Ang-II has also been reported to participate in the maintenance of AF by inducing atrial structural remodelling, which is characterized by interstitial fibrosis. 43,44 Therefore, blocking the atrial fibrosis induced by Ang-II may provide a strategy to inhibit the occurrence and maintenance of AF.
PU.1, a transcription factor, is an important regulator of tissue fibrosis. 45 Earlier studies have shown that PU.1 is expressed in myeloid cells, B cells and so on. 12,13 Recent research suggests that PU.1 is also highly expressed in fibrotic fibroblasts isolated from liver, lung, and kidney tissue and fibrotic joints; interestingly, resting fibroblasts and inflammatory fibroblasts have not been found to exhibit PU.1 expression. 26 However, there remains limited information on the expression of PU.1 in Ang-II-mediated atrial fibrosis. Therefore, in the present study, PU.1 expression was analysed both in vivo (in mice subcutaneously infused with Ang-II) and in vitro (in atrial fibroblasts isolated from Ang-II-induced mice). The PU.1 inhibitor, DB1976, was used to assess the effect of PU.1 inhibition on Ang-II-mediated atrial fibrosis. Treatment with DB1976 in the mice model of subcutaneously infused with Ang-II did not affect bodyweight and pain (data not shown); it is similar to previous literature reports. 26 Research has reported that DB1976 do not bind PU.1 but strongly inhibit the PU.1/DNA complex in TGF-β1 administration has been shown to exert fibrotic effects by enhancing collagen synthesis in vitro and in vivo, with higher levels of TGF-β1 being associated with more extensive atrial fibrosis. 47 Moreover, infusion with Ang-II is reported to result in atrial fibrosis, increased AF inducibility and increased AF duration. 48 Thus, Ang-II and TGF-β1 activation are deemed to play an important role in both atrial fibrosis and electric remodelling. 7,49 Additionally, Ang-II exerts its effects by increasing TGF-β1 expression, as Ang-II is unable to induce cardiac fibrosis in the absence of TGF-β1. 1 Therefore, reducing activation of the TGF-β1/Smads pathway is expected to abrogate the atrial fibrosis induced by Ang-II. PU.1, a vital transcription factor, participates in the TGF-β1/Smad3 pathway to promote collagen expression. 26 However, whether PU.1 impacts Ang-II-mediated TGF-β1/Smads pathway activation in models of atrial fibrosis was previously unclear. Consistent with an earlier study, 26 the present study showed that Ang-II significantly increases the protein expression of TGF-β1, p-Smad3 and p-Smad2/3, and these effects are at- Previous studies have demonstrated that atrial fibroblast proliferation and differentiation are vital triggers of cardiac fibrosis. Ang-II and TGF-β1 are the most potent stimulators of collagen synthesis in cardiac fibroblasts. 50,51 We have known that fibroblast proliferation and differentiation into the myofibroblast phenotype contributes to the accumulation of extracellular matrix (including α-SMA, chemokines, cytokines and adhesion complexes) in the perivascular space and the perimysium surrounding the cardiac muscle bundles. 50,[52][53][54] The recent report has showed that PU.1 can prompt the conversion of fibroblasts into myofibroblasts in tissues of the lung, liver, kidney and skin via the TGF-β1/Smads pathway. 26 However, whether PU.1 participates in atrial fibroblast proliferation and differentiation was previously unclear. Especially, evidence shows that cardiac myofibroblasts up-regulate protein expression of α-SMA, SMemb and ED-A Fn; especially, ED-A Fn is the major driver of myofibroblast phenoconversion, and SMemb and α-SMA are the two major contractile proteins, which generate tension by contracting to generate F I G U R E 7 Summary of the roles of PU.1 in atrial fibrosis and vulnerability to atrial fibrillation induced by angiotensin-II in mice. PU.1 inhibition attenuates atrial fibrosis and vulnerability to atrial fibrillation induced by angiotensin-II in mice by reducing TGF-β1/Smads signalling pathway activation. MI: myocardial infarction force. [55][56][57][58] Additionally, DDR2 predominantly expresses in fibroblasts and is a fibroblast-selective marker, 59 DDR2 has been reported to regulate cell proliferation, 60 differentiation 61,62 and extracellular matrix remodelling. 63 Our results showed that PU.1 inhibition with DB1976 or si-PU.1 effectively reduced the number of atrial fibroblasts and decreased the expression of PCNA, α-SMA, SMemb and ED-A fibronectin and DDR2 in vivo and in vitro. These findings further demonstrate the key role of PU.1 in Ang-II-induced atrial fibrosis.
TGF-β1-mediated aberrant collagen I is produced by myofibroblasts and is a major constituent of the cardiac interstitium; moreover, TGF-β1 itself is activated by Ang-II and TGF-β1 stimulates CTGF expression. 64 27 Our results also indicated that the percentage of atrial fibrosis area in vivo was significantly decreased in the DB1976-exposed group compared with the Ang-II group based on Masson trichrome staining. In addition, DB1976 was observed to markedly reduce the up-regulation of CTGF and collagen I induced by Ang-II. Consistent with these in vivo results, cultured atrial fibroblasts derived from Ang-II-induced mice showed significantly reduced CTGF and collagen I expression following treatment with DB1976 or si-PU.1. It may be deduced that PU.1 inhibition attenuates the atrial fibrosis induced by Ang-II, which may also lessen vulnerability to AF.
Interstitial fibrosis accelerates conduction, and the presence of interstitial collagen is thought to represent both persistent AF and more rapid longitudinal conduction in patients with AF. [66][67][68] Additionally, fibroblast proliferation, differentiation and fibrosis in the myocardial interstitium promote the occurrence and maintenance of AF via electrical conduction disturbances. 69 These disturbances result in anisotropy and re-entry, eventually leading to the proliferation of ectopic foci and the propagation of irregular wave fronts. 68,70,71 A recent report demonstrated that the relationship between fibrosis and PU.1-related effects on the TGF-β1/Smads pathway exists in the skin, lung, joints, liver and kidney. 26 Here, we explored whether PU.1 inhibition reduces Ang-II-induced atrial fibrosis to further decrease vulnerability to AF. The present study demonstrated that PU.1 inhibition did indeed attenuate AF vulnerability and AF duration in mice post-Ang-II infusion. A possible mechanism for this reduction may be that PU.1 inhibition suppresses atrial fibrosis at least in part through blocking the TGF-β1/Smads pathway.
Evidences show that Smad3 is a primary intracellular signalling molecule along the TGF-β1/Smads pathway for fibrosis, and downregulation of the phosphorylation of Smad3 has been shown to block the pathological process of fibrosis. 72 Similarly, our goal was also to observe whether PU.1 would affect the activity of TGF-β1/Smads pathway by mediating Smad3. It is generally accepted that TGF-β1 binding of ALK5 activates an intrinsic serine/threonine kinase that mainly phosphorylates Smad3, and disruption of Smad3 appears to impair myofibroblast differentiation and restrain the expression of extracellular matrix. 73 Further, PU.1 can induce Smad3 activation to promote tissue fibrosis, 26 suggesting a direct impact of PU.1 on atrial fibrosis via TGF-β1/Smads. However, the specific mechanisms underlying the effects of PU.1 on atrial fibrosis remained to be further explored. Therefore, the present study employed atrial fibroblasts transfected with siRNA to knock down Smad3. The results showed that PU.1 overexpression in Ang-II-induced atrial fibroblasts obviously increased proteins of p-Smad3; on the contrary, Smad3 knockdown in atrial fibroblasts decreased PU.1 expression. Additionally, si-Smad3 attenuated the Ang-II-induced atrial fibroblast proliferation (the CCK-8 absorbance values) and differentiation (α-SMA, SMemb and collagen I), whereas these effects were reversed by the PU.1overexpression adenoviral vector, Ad-PU.1. These findings indicate that PU.1 inhibition attenuates the vulnerability to AF and AF duration induced by Ang-II, and the mechanisms underlying these effects may be related to the reduced activation of TGF-β1/Smads pathway.
In conclusion, PU.1 inhibition can attenuate Ang-II-induced atrial fibrosis and consequently decrease vulnerability to AF by suppressing activation of the TGF-β1/Smads pathway. These findings suggest that PU.1 may be a promising therapeutic target for the treatment of atrial fibrosis and the reduction of AF vulnerability.

ACK N OWLED G EM ENTS
We are grateful to our colleague Zhang Jingjing (Wuhan University) for help with carrying out the animal experiments. This work was

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to declare.