Introduction: Fibrosis is an important contributor to atrial fibrillation (AF) substrate in heart failure, with TGF-β signaling generating atrial fibrosis. Inflammation is also known to contribute to AF. However, it is not known: a) if inflammation underlies the development of atrial fibrosis, b) if TGF-β signaling is involved in the generation of atrial inflammation and c) what is the relative role of canonical SMAD2/3 versus non-canonical TGF-β-activated kinase 1 (TAK) signaling in the creation of atrial inflammation/fibrosis. Methods: 21 dogs underwent transfection in the posterior left atria of either a plasmid expressing a dominant negative TGF-β type II receptor (pUBC-TGFβ-DN-RII) (N=9) or control vector (pUBC-LacZ) (N=12), followed by 3-4 weeks of right ventricular tachypacing (VTP) (240 bpm). After 3-4 weeks of VTP, the following were assessed: AF by open-chest mapping. Atrial tissue was assessed for signaling (pSMAD2/3 and pTAK) by western blot, fibrosis by trichrome staining and iflammation by CD68 staining. Results: The interstitial fibrosis in LacZ atria was replacement in character (Fig 1A). TGFβ-RII-DN dogs had a significant decrease in replacement fibrosis leading to a decrease in duration of inducible AF (LacZ 63.64 ± 12.12 s vs. TGFβ 26.52 ± 3.79 s, p < 0.05). While both groups demonstrated evidence of inflammation (CD68 staining), there was no difference in inflammation between the groups (Fig 1B). TGFβ-RII-DN dogs had a significant decrease in pSMAD2/3 but not pTAK, compared to pUBC-LacZ dogs. Conclusion: The development of fibrotic substrate for AF in advanced HF is driven not by inflammation but primarily by SMAD2/3 mediated myocardial replacement fibrosis. These findings may have important implications for mechanism-guided therapies for AF, with canonical TGF-β signaling appearing to be an attractive therapeutic target in AF.