Paracrine signalling by cardiac calcitonin controls atrial fibrogenesis and arrhythmia.

Abstract
Lucia M. Moreira, Abhijit Takawale, Mohit Hulsurkar, David A. Menassa, Agne Antanaviciute, Satadru K. Lahiri, Neelam Mehta, Neil Evans, Constantinos Psarros, Paul Robinson, Alexander J. Sparrow, Marc-Antoine Gillis, Neil Ashley, Patrice Naud, Javier Barallobre-Barreiro, Konstantinos Theofilatos, Angela Lee, Mary Norris, Michele V. Clarke, Patricia K. Russell, Barbara Casadei, Shoumo Bhattacharya, Jeffrey D. Zajac, Rachel A. Davey, Martin Sirois, Adam Mead, Alison Simmons, Manuel Mayr, Rana Sayeed, George Krasopoulos, Charles Redwood, Keith M. Channon, Jean-Claude Tardif, Xander H. T. Wehrens, Stanley Nattel & Svetlana Reilly1,18 ✉


Nature | www.nature.com | 1


Article
Paracrine signalling by cardiac calcitonin controls atrial fibrogenesis and arrhythmia Lucia M. Moreira1,17, Abhijit Takawale2,3,17, Mohit Hulsurkar4,5, David A. Menassa6,7, Agne Antanaviciute8, Satadru K. Lahiri4,5, Neelam Mehta1, Neil Evans1, Constantinos Psarros1, Paul Robinson1, Alexander J. Sparrow1, Marc-Antoine Gillis3,9, Neil Ashley10, Patrice Naud3,9, Javier Barallobre-Barreiro11, Konstantinos Theofilatos11, Angela Lee1, Mary Norris1, Michele V. Clarke12, Patricia K. Russell12, Barbara Casadei1, Shoumo Bhattacharya1, Jeffrey D. Zajac12, Rachel A. Davey12, Martin Sirois3,9, Adam Mead8, Alison Simmons8, Manuel Mayr11, Rana Sayeed13, George Krasopoulos13, Charles Redwood1, Keith M. Channon1, Jean-Claude Tardif3,9, Xander H. T. Wehrens4,5,14, Stanley Nattel2,3,9,15,16,18 & Svetlana Reilly1,18 ✉ Atrial fibrillation, the most common cardiac arrhythmia, is an important contributor to mortality and morbidity, and particularly to the risk of stroke in humans1. Atrial-tissue fibrosis is a central pathophysiological feature of atrial fibrillation that also hampers its treatment; the underlying molecular mechanisms are poorly understood and warrant investigation given the inadequacy of present therapies2. Here we show that calcitonin, a hormone product of the thyroid gland involved in bone metabolism3, is also produced by atrial cardiomyocytes in substantial quantities and acts as a paracrine signal that affects neighbouring collagen-producing fibroblasts to control their proliferation and secretion of extracellular matrix proteins. Global disruption of calcitonin receptor signalling in mice causes atrial fibrosis and increases susceptibility to atrial fibrillation. In mice in which liver kinase B1 is knocked down specifically in the atria, atrial-specific knockdown of calcitonin promotes atrial fibrosis and increases and prolongs spontaneous episodes of atrial fibrillation, whereas atrial-specific overexpression of calcitonin prevents both atrial fibrosis and fibrillation. Human patients with persistent atrial fibrillation show sixfold lower levels of myocardial calcitonin compared to control individuals with normal heart rhythm, with loss of calcitonin receptors in the fibroblast membrane. Although transcriptome analysis of human atrial fibroblasts reveals little change after exposure to calcitonin, proteomic analysis shows extensive alterations in extracellular matrix proteins and pathways related to fibrogenesis, infection and immune responses, and transcriptional regulation. Strategies to restore disrupted myocardial calcitonin signalling thus may offer therapeutic avenues for patients with atrial fibrillation. Atrial fibrillation (AF), the most prevalent cardiac arrhythmia, is associated with considerable mortality and morbidity, and its treatment is complicated by adverse remodelling of the atria2. Current pharmacological strategies for AF have limited effectiveness and can have undesirable effects. The identification of new targets related to the pathophysiological mechanisms that underlie AF might open new therapeutic avenues2. Structural remodelling in AF involves the accumulation of cross-linked collagen from atrial cardiofibroblasts (ACFs); however, the underlying mechanisms are incompletely understood. We sought to assess whether these might have a connection with calcitonin (CT), a hormone predominantly (but not exclusively) produced by thyroid parafollicular cells that is involved in bone resorption and collagen turnover3 and also affects other tissues, including skeletal muscle4. Levels of circulating CT decrease with age5, which is the main risk factor for AF1,2. Genome-wide association studies (GWAS) report links between single-nucleotide polymorphisms that affect the CT receptor (CTR) and body mass index6, which is another AF risk factor. In early studies, CT was reported to prevent calcium-induced ventricular arrhythmias7 and inhibit atrial chronotropic and inotropic function8. No information was available, however, about the involvement of CT in AF or about functional extrathyroid production of CT. Here we sought to (i) assess whether atrial myocardium produces CT and, if so, identify the cellular source(s), (ii) explore CTR-mediated paracrine effects on ACF proliferation and collagen processing and (iii) determine whether CT signalling regulates atrial fibrotic remodelling and susceptibility to AF. https://doi.org/10.1038/s41586-020-2890-8 Received: 18 April 2019 Accepted: 13 August 2020


Published online: xx xx xxxx
Check for updates A list of affiliations appears at the end of the paper. 2 | Nature | www.nature.com


Article


Atrial cardiomyocytes produce CT
Atrial myocardium secretes several hormones9. We investigated CT-related gene expression in human right atrial tissue, isolated atrial cardiomyocytes (ACMs) and ACFs, and epicardial adipose tissue (detailed in Extended Data Table 1). Human CT originates from the calcitonin-related polypeptide-alpha (CALCA) gene on chromosome 11 (ID:ENSG00000110680), from which it is co-transcribed with alpha-calcitonin-gene-related peptide (αCGRP). Using primers specific for the mRNA splice variants that encode CT and αCGRP, we detected the expression of both CT and αCGRP transcripts in atrial tissue and in isolated ACMs and ACFs, but not in epicardial fat (Fig. 1a–c). CT protein was detectable in the secretome of human ACMs, but not ACFs (Fig. 1d). Patients with persistent AF showed impaired production of mature CT and its precursor procalcitonin (pro-CT; Fig. 1e–g), mirrored by an increase in αCGRP in human right atrial tissue lysates and the ACM secretome (Extended Data Fig. 1a, b). The abundance of CT and αCGRP transcripts and the ratio of CT to αCGRP mRNA levels in ACMs were unchanged in AF (Extended Data Fig. 1c–e); however, ACM CT protein levels correlated negatively with age (Extended Data Fig. 1f). Comparing the CT transcript level and CT protein secretion between human ACMs and human medullary thyroid carcinoma TT cells, which constitutively produce copious CT, revealed that CT transcript levels in ACMs were approximately half those in TT cells (Fig. 1h) and calcium-induced CT secretion in ACMs was around 16-fold greater than in TT cells, implying a high rate of CT production and/or its secretion in ACMs (Fig. 1i). Thus, human ACMs are an active source of extrathyroid CT. Human ACFs express functional CTR CT exerts its biological actions via the CTR. Human atrial myocardium exclusively expresses CTR isoform 1a (Extended Data Fig. 1g, h). The CTR is expressed in human atrial tissue and ACFs, as indicated by quantitative PCR (qPCR), immunoblotting and immunostaining (Fig. 1j, k, Extended Data Fig. 1g–i). CTR activation by CT caused time-dependent translocation of CTR from the ACF surface to the cytoplasm (Fig. 1j) and concentration-dependent changes in ACF morphology, as detected by impedance monitoring (Extended Data Fig. 1j). CT concentration-dependently increased cyclic adenosine monophosphate (cAMP) in human ACFs (Fig. 1l), and this effect was blocked by inhibition of Gαs, but not Gαi, and prevented by administration of the CTR antagonist sCT8-32 (Fig. 1m). CT administration did not affect the phosphorylation of ERK1/2 (Extended Data Fig. 1k), suggesting that CT does not influence Gαq-mediated responses. Thus, human ACFs express a fully functional CTR primarily coupled to Gαs.


CT–CTR signalling regulates ACF function
Treating human ACFs with 100 nM CT for 72 h produced an approximately 46% reduction in collagen accumulation, with no changes in fibronectin (Fig. 2a). CT-treated ACFs showed approximately twofold decreases in proliferation, cell migration and accumulation of calcium-enriched deposits (Fig. 2b–d), but no changes in the abundance of α-smooth muscle actin protein or mRNA (α-SMA; Fig. 2e, f). CT-mediated effects on collagen production and ACF proliferation were reversed by silencing of CTR with locked nucleic acid antisense oligonucleotides (Fig. 2g, h and Extended Data Figs. 1l, m and 8c). In TGFβ1-stimulated cells, 500 nM CT decreased both cell migration (by about 42%) and proliferation and reduced secreted collagen levels (by about 40%) in ACFs (Extended Data Fig. 1o–q). These results indicate that CT–CTR signalling actively inhibits production of collagen 1, as well as fibroblast proliferation and migration. Collagen accumulation and proliferation were unaltered by 10 or 100 nM αCGRP or by 10 nM αCGRP, respectively (Fig. 2i–k and Extended Data Fig. 1n, r). We next investigated whether CT affects the synthesis, degradation and processing of collagen 1. CTR activation upon 72 of h exposure to 100 nM CT did not change the expression of the fibrillar collagen transcripts (Extended Data Fig. 2a, b) or the abundance of C-terminal telopeptide (ICTP), a marker of collagen 1 degradation (Extended Data Fig. 2c, d), but it CT stimulation on CTR localization in ACFs. Scale bar, 50 μm. k, CTR gene expression in ACFs. l, m, Effect of CT on ACF cAMP, blocked by the Gs inhibitor NF499 or CTR antagonist sCT8-32, but not by the Gi inhibitor PTX. Data are mean ± s.e.m., except for f–i, l (median and interquartile range). n denotes independent donors. Two-sided tests: one-way ANOVA with Sidak correction (d, m), unpaired t-test (e), Mann–Whitney U-test (f–i), Kruskal–Wallis test with Dunn’s correction (l). Gel source data in Supplementary Fig. 1; replication information in Supplementary Methods. Nature | www.nature.com | 3 inhibited the cleavage and processing of procollagen into mature collagen alpha-2 chain (Extended Data Fig. 2e). CT increased the accumulation of unprocessed forms of collagen 1 (procollagen, collagen C-terminal propeptide and pC collagen) and decreased the formation of mature collagen 1. Bone morphogenetic protein type-1 (BMP1) cleaves pC collagen10, and we noted concentration-dependent CT-induced reductions in BMP1 activity without changes in BMP1 gene expression or protein level (Extended Data Fig. 2f–h). Stimulation of human ACFs with 500 nM CT for 72 h did not affect collagen 1 mRNA transcript levels, but it increased the degradation of collagen 1 (Extended Data Fig. 2i, j). These results indicate that the inhibitory effects of CT on collagen accumulation are at least partly mediated by BMP1 inhibition and result from interference by CT with collagen processing and degradation, rather than collagen synthesis. Culturing human ACFs with 100 nM CT for 24 or 72 h did not alter their transcriptome (as assessed by single-cell RNA sequencing (RNA-seq) and microarrays) (Extended Data Fig. 2k–n and Supplementary Data), but did lead to significant modification of the proteome. Specifically, CT suppressed the accumulation of 143 out of 191 fibrogenesis-related extracellular matrix (ECM) proteins in the human ACF secretome (with collagens 1 and 3 being among those most affected; Fig. 3a, b and associated Source Data) and altered the abundance of 226 out of 3,253 cellular proteins (Fig. 3c and associated Source Data). ACF secretion of selected non-ECM proteins was unaffected by CT (Extended Data Fig. 3a–d). A Gene Ontology (GO) analysis (Fig. 3d and Extended Data Fig. 3e) of the data from CT-exposed ACFs revealed cellular enrichment of proteins involved in the ribosomal pathway and in biological processes and functions related to fibrogenesis (for example, collagen fibril organization, cadherin binding and cell adhesion), immune and infection responses, and transcriptional regulation.


Disrupted CT–CTR signalling in AF
We next investigated whether CT could rescue the profibrotic phenotype of ACFs from patients with AF and whether AF was accompanied by changes in CTR protein content, gene expression or distribution. The protein content and gene expression of CTR were unchanged in postoperative and paroxysmal AF (Extended Data Fig. 3f–i); CTR protein level, but not gene expression, was modestly decreased in patients with persistent AF (Extended Data Fig. 3j, k). Persistent AF is typically accompanied by fibrosis11, but ACFs from patients with persistent AF did not respond to CT, as CT did not affect collagen 1 production or cell proliferation (as determined by scar-in-a-jar assay, Fig. 3e, f), fibronectin production (scar-in-a-jar), ACF migration (scratch assay) or α-SMA protein (immunoblotting) (Extended Data Fig. 3l–n). Because the modest reduction of CTR protein does not explain non-responsiveness to exogenous administration of CT, we looked for evidence of AF-associated dysregulation of pre-existing endogenous downstream signalling or disease-related modification of ACF phenotypes. Single-cell RNA sequencing (scRNA-seq, specifically SMART-seq2) of freshly isolated human ACFs identified five transcriptional clusters (Extended Data Figs. 4a, b and 5). The largest population of cells (cluster 1) contained abundant ACTA2 and NOTCH3 transcripts, typical of actively proliferating fibroblasts12. A smaller ACTA2-positive population (cluster 2) enriched for myosins (including MYH2, MYH3 and MYH7) represented cells with increased contractility that appear during wound repair and contribute to ECM stiffness13. Cluster 3 cells were ACTA2 negative (possibly representing embryonic fibroblasts, stellate cells or an intermediate cell subset between fibroblasts and myofibroblasts)14. Cells expressing CD45 (a marker of immature and leukocyte blood cells) and proliferation (i) and 72-h collagen secretion (Sirius red) in conditioned medium without ( j) or with 10 ng ml−1 TGFβ1 (k); BIBN4096, CGRP receptor antagonist. Results are mean ± s.e.m., except in a, h (median and interquartile range), c (mean ± s.d.). Two-sided tests: Mann–Whitney U-tests (a), unpaired t-tests (e, f), two-way ANOVA with Sidak correction (b, d), one-way ANOVA with Sidak correction (c, g), Kruskal–Wallis test with Dunn’s correction (h) and Friedman test (i–k). n, individual donors; fc, fold change. Gel source data in Supplementary Fig. 1; replication information in Supplementary Methods. 4 | Nature | www.nature.com


Article
endothelial cells incompletely depleted during ACF isolation formed clusters 4 and 5, respectively. Although clusters 2–5 were similar between cells derived from patients with AF and control cells from patients with sinus rhythm, the cluster 1 ACFs had 23 differentially expressed genes (DEGs) associated with AF (Extended Data Fig. 4c–f) that are related to fundamental functions of ACFs, including cell migration and invasion (RHOB), regulation of fibrogenesis (FOXF1, SIK1, NRF4A1, BHLHE40 and PDK4), differentiation (NR4A1, NR4A2, CEBP and SPC24), circadian rhythm (SIK1 and BHLHE40), metabolism (PDK4), immune response and inflammation (IL6, ADAMTS1 and BHLHE40) and cell transcription (NR4A1, NR4A2, BHLHE40 and FOXF1). We next compared atrial protein content (using immunoblotting) of selected components of the CT–CTR signalling cascade that may remain unchanged at the transcriptomic level. AF was associated with increased expression of atrial cAMP (Extended Data Fig. 6a), a downstream mediator of CT–CTR (shown in Fig. 1l, m). In light of the limited changes in CTR expression in ACFs seen in AF, we verified the subcellular localization of CTRs and found that in ACFs from patients with persistent AF, CTRs were relocalized from cell surfaces to intracellular spaces (Fig. 3g, Extended Data Fig. 6b). Because CTR responses require interaction with extracellular ligand, loss of cell surface CTRs may be important in the non-responsiveness of ACFs to CT in AF.


CT–CTR system controls susceptibility to AF
To assess the consequences of depressed CTR function for AF, we assessed atrial fibrosis and susceptibility to AF in global CTR knockout mice and control mice heterozygous for loxP-based (floxed) CTR15. Global CTR knockout mice showed significant atrial fibrosis (Fig. 4a–c) without any changes in the expression of the genes that encode collagens 1 and 3, fibronectin and α-SMA (Extended Data Fig. 7a–d) or in cardiac morphology (Fig. 4a, top panels). In vivo electrophysiological testing showed longer duration and greater inducibility of AF episodes in mice with CTR knockout compared to control mice (Fig. 4d–g and Extended Data Fig. 7m, n), but no differences in atrial effective refractory periods (Fig. 4h), morphological parameters or cardiac function (Extended Data Fig. 7e–l and Extended Data Table 2a). We next assessed the effect of CT production by ACMs on susceptibility to AF. We modified an existing mouse model of spontaneous AF, the LKB1-deficient mouse16, to atrial-specific knockdown of LKB1 (Fig. 4i–k), in order to avoid potential confounding effects of LKB1 deficiency in the ventricle. Mice with atrial-specific knockdown of LKB1 (Lkb1fl/fl) were treated with a cardiotropic adeno-associated viral type 9 (AAV9) vector to overexpress Cre only in the atria using a promoter fragment of atrial natriuretic factor (ANF)17 (Extended Data Fig. 7o). These atrial-specific LKB1-knockout mice developed spontaneous AF (Fig. 4o, q). Next, we overexpressed or knocked down CT using shRNA using AAV9–ANF viral vectors (Fig. 4i–k and Extended Data Fig. 7o). The double-knockdown mice (Fig. 4l, m) had ~2.5-fold greater atrial fibrosis than those with LKB1 knockdown alone (Fig. 4l, m), although with preserved cardiac structure and function (Extended Data Table 2b). Double-knockdown mice developed spontaneous AF starting at 8 weeks of age, 2 weeks earlier than those with only LKB1 knockdown (Fig. 4n, o, q). At 12 weeks, 62.5% of double-knockdown mice showed spontaneous AF versus 23% of LKB1-knockdown mice (Fig. 4o, q), and the maximum length of AF episodes was about 16-fold greater in the double-knockdown group (Fig. 4p). By contrast, the LKB1-knockdown mice with CT overexpression in atria did not show spontaneous AF or augmented atrial fibrosis (Fig. 4l–q) and had a heart rate 19% lower than that of control Lkb1fl/fl mice (Extended Data Table 2b). These findings support the hypothesis that CT–CTR signalling is important in the arrhythmogenesis of AF and in remodelling during atrial fibrosis.


Discussion
Here, we identified a previously undescribed and important role for a CT–CTR signalling cascade in human atrial myocardium that fine tunes the function of ACFs to prevent excess accumulation of fibrous tissue (Extended Data Fig. 8a). When this system becomes dysregulated, as a result of either heart disease or suppression of the expression of CT or CTR, excess ACF activity leads to collagen accumulation and susceptibility to AF. Human ACMs represent a potent source of myocardial CT that exerts paracrine effects on ACFs by binding to ACF individuals with AF and 15 ACFs from 6 individuals with SR. Scale bar, 50 μm. Data in a, b are adjusted for multiple testing with Benjamini–Hochberg false discovery rate (FDR) calculated with the Limma package (v3.34.5) and empirical Bayes (Ebayes) algorithm. Except in c, P values were not corrected for multiple testing given that 3,253 proteins were quantified. Functional enrichment analysis used DAVID 6.8 with human proteome background. Data in e, f are averages with interquartile ranges analysed by two-sided Kruskal– Wallis test with Dunn’s correction. Gel source data in Supplementary Fig. 1; replication information in Supplementary Methods. Nature | www.nature.com | 5 CTRs, inhibiting cell proliferation and fibrotic responses, in part by suppressing BMP1-dependent cleavage of collagen. Atrial fibrosis, the most prominent feature of structural remodelling in AF, is commonly implicated in the arrhythmogenic substrate and is believed to be of great clinical pathophysiological and prognostic importance2,18. Although many pathophysiological aspects of atrial fibrosis are understood18, no clinically effective targets have yet been identified, and there is a need to improve our mechanistic understanding to pinpoint mechanisms that could lead toward therapeutic breakthroughs2. CT is secreted primarily by thyroid C-cells, yet thyroid agenesis or thyroidectomy do not consistently change circulating concentrations of CT5, pointing to the existence of substantial extrathyroid sources. Recent studies have uncovered extrathyroid secretion of CT, for example, in human placenta19 and sperm20. Although CT synthesis in the heart had not previously been described, atrial myocardium is well known to secrete other hormones, such as atrial and brain natriuretic peptides, endothelin-1 and adrenomedullin. We found that ACMs secreted CT at a level about 16-fold greater than that of the TT cells that we studied. Our findings raise the intriguing possibility that human atrial myocardium may be a prominent source of CT and of pro-CT, an important mediator and marker of inflammation that is widely used as a biomarker21. The physiological basis or function and regulation of pro-CT and CT production in the atrium requires further study. The atria are particularly prone to fibrosis, an effect linked to the hypersensitivity of ACFs to profibrotic stimuli22, and the atrial CT–CTR axis might act as a counter-regulatory system. When cardiac pathology leads to an atrial fibrotic response, as in persistent AF, reduction of CT production by ACMs and of CTR expression in ACF membranes might, by removing the CT–CTR ‘brake’ on the system, allow fibrosis to occur. We found that ACFs from control (sinus rhythm) patients expressed fully functional CTRs, coupled principally to Gαs, consistent with prior observations of preferential CTR Gαs coupling in other cell types23. The CT-mediated increase in intracellular cAMP and the effect on collagen 1 accumulation and ACF proliferation were CTR specific, as they were fully prevented by administration of the CTR antagonist sCT8-3224. CT-mediated actions were independent of αCGRP, another splice product of the CALCA gene secreted by human ACMs, as exogenous αCGRP did not alter the collagen production or proliferation of human ACFs. Discordant changes in CT and αCGRP levels in ACMs might indicate a skewing of CALCA splicing toward αCGRP in persistent AF; this possibility requires further investigation. The CT-induced decrease in ACF collagen secretion might be caused by altered collagen synthesis, processing and/or degradation. Our results show that low concentrations of CT primarily inhibit the maturation and cleavage of unprocessed collagen, partly owing to decreased activity of BMP1, which cleaves the C-terminal propeptide of collagens 1–325 to allow CT maturation and is inhibited by increased intracellular cAMP26. CT stimulates cAMP production by ACFs, which in turn plays a prominent part in regulating cardiac fibrosis via downstream mediators including PKA and EPAC1/227. Higher concentrations of CT accelerated collagen degradation in ACFs, suggesting that larger amounts of CT may influence multiple steps in collagen processing and be more effective in suppressing fibrosis. separating them that allows the CT protein to be cleaved from mCherry (k, C). l, m, Masson Trichrome images of hearts (l, top) or atria (l, bottom); atrial fibrosis is quantified in m. Scale bars, 1 mm (top), 20 μm (bottom). n–q, Recordings of spontaneous AF (n), AF-free survival (o) and longest duration of AF (p) for the indicated groups; the number of animals at risk is listed in q. n, individual animals. Data are mean ± s.e.m., except in c, f ( j-pro-CT/j-pro-αCGRP), m, p (median and interquartile range). Two-sided tests: unpaired t-test (b, h), Mann– Whitney U-test (c, f), one-way ANOVA/Holm–Sidak test ( j, αCGRP), Sidak correction ( j, LKB1, CT-log-transformed, and p, log-transformed); one-sided Fisher’s exact test (g), Kruskal–Wallis test with Dunn’s correction ( j, pro-CT/ pro-αCGRP; m), log rank (o). Gel source data in Supplementary Fig. 1; replication information in Supplementary Methods. 6 | Nature | www.nature.com


Article
Unbiased high-throughput proteome profiling of human ACFs revealed broad effects of CT on human ACFs, suggesting both direct inhibition of ECM proteome production and effects on signalling that controls proliferation and migration. In ACFs from patients with persistent AF, we found no evidence of antifibrotic actions by CT, as CTRs were primarily localized in the ACF intracellular compartments in patients with AF, whereas they showed extensive cell-surface localization in the ACFs of control patients, precluding activation of membrane CTRs by extracellular CT in AF. Whether the intracellular abundance of CTR in AF affects other functions of ACFs remains to be explored. Given that AF-derived ACFs had unchanged CTR gene expression and only modestly reduced levels of CTR protein, defective CTR processing and signalling (for example, lack of CTR chaperonage by an intracellular binding protein such as filamin-A28 or altered CTR trafficking by receptor-activity-modifying proteins, or RAMPs29) may underlie its disordered subcellular localization. CTRs bound to RAMPs may respond to amylin and αCGRP29. Persistent AF also downregulated the CT–CTR–cAMP axis effectors CREB and EPAC1 (which facilitates collagen secretion and left atrial fibrosis in heart failure30), possibly contributing to reduced antifibrotic effects in AF. To test whether AF is associated with pre-existing transcriptional changes that could account for altered responses to CT, we performed scRNA-seq on freshly isolated human ACFs. Non-cultured ACFs fell into five distinct transcriptional clusters, with AF-associated differential expression present only in ACTA2+NOTCH3+ cells. These cells show a profile associated with ACF migration and invasion, differentiation and transcription, regulation of fibrosis, circadian rhythm and cellular immunity. These results reveal an additional level of complexity of AF-associated changes in ACFs that may underlie altered cellular responses, including those to CT. We tested the in vivo consequences of the disrupted CT–CTR signalling in genetically modified mice. Global deletion of the CTR gene enhanced atrial fibrosis in the absence of left atrial dilatation or left-ventricular dysfunction. To examine spontaneous occurrence of AF, we turned to a mouse model of LKB1 suppression. Myocardial-selective knockout of LKB1 involves effects secondary to ventricular LKB1 deletion16. Thus, we generated a new atrial-specific LKB1 knockdown mouse model that developed spontaneous AF at 10 weeks of age without ventricular remodelling. CT downregulation in the atria of these mice significantly worsened both arrhythmic and profibrotic phenotypes, and was fully rescued by atrial-targeted overexpression of CT. Through this work, we identified a previously undescribed CT–CTR paracrine signalling system in the human atrium. Human ACMs represent a substantial source of CT that, through binding to the CTR on the ACF membrane, controls fibroblast proliferation and BMP1-related collagen processing. Disruption of the CT–CTR axis permits excessive atrial fibrogenesis and promotes arrhythmogenesis. Restoration of the CT–CTR functional cascade might thus help to control the development of the AF-related arrhythmogenic substrate in humans.


Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41586-020-2890-8. 1. Chugh, S. S. et al. Worldwide epidemiology of atrial fibrillation: a Global Burden of Disease 2010 Study. Circulation 129, 837–847 (2014). 2. Heijman, J., Guichard, J. B., Dobrev, D. & Nattel, S. Translational challenges in atrial fibrillation. Circ. Res. 122, 752–773 (2018). 3. Karsdal, M. A., Byrjalsen, I., Leeming, D. J., Delmas, P. D. & Christiansen, C. The effects of oral calcitonin on bone collagen maturation: implications for bone turnover and quality. Osteoporos. Int. 19, 1355–1361 (2008). 4. Yamaguchi, M. et al. Calcitonin receptor signaling inhibits muscle stem cells from escaping the quiescent state and the niche. Cell Rep. 13, 302–314 (2015). 5. Basuyau, J. P., Mallet, E., Leroy, M. & Brunelle, P. Reference intervals for serum calcitonin in men, women, and children. Clin. Chem. 50, 1828–1830 (2004). 6. Locke, A. E. et al. Genetic studies of body mass index yield new insights for obesity biology. Nature 518, 197–206 (2015). 7. Cuparencu, B., Ticsa, I., Barzu, T., Sandor, V. & Gozariu, L. [Effect of calcitonin on certain experimental models of arrhythmia]. Therapie 1975, 555–563 (1975). 8. Chiba, S. & Himori, N. Effects of salmon calcitonin on SA nodal pacemaker activity and contractility in isolated, blood-perfused atrial and papillary muscle preparations of dogs. Jpn. Heart J. 18, 214–220 (1977). 9. Ogawa, T. & de Bold, A. J. The heart as an endocrine organ. Endocr. Connect. 3, R31–R44 (2014). 10. Vadon-Le Goff, S., Hulmes, D. J. & Moali, C. BMP-1/tolloid-like proteinases synchronize matrix assembly with growth factor activation to promote morphogenesis and tissue remodeling. Matrix Biol. 44–46, 14–23 (2015). 11. Martins, R. P. et al. Dominant frequency increase rate predicts transition from paroxysmal to long-term persistent atrial fibrillation. Circulation 129, 1472–1482 (2014). 12. Grieskamp, T., Rudat, C., Lüdtke, T. H., Norden, J. & Kispert, A. Notch signaling regulates smooth muscle differentiation of epicardium-derived cells. Circ. Res. 108, 813–823 (2011). 13. Bond, J. E. et al. Temporal spatial expression and function of non-muscle myosin II isoforms IIA and IIB in scar remodeling. Lab. Invest. 91, 499–508 (2011). 14. Li, H. et al. Reference component analysis of single-cell transcriptomes elucidates cellular heterogeneity in human colorectal tumors. Nat. Genet. 49, 708–718 (2017). 15. Davey, R. A. et al. Calcitonin receptor plays a physiological role to protect against hypercalcemia in mice. J. Bone Miner. Res. 23, 1182–1193 (2008). 16. Ozcan, C., Battaglia, E., Young, R. & Suzuki, G. LKB1 knockout mouse develops spontaneous atrial fibrillation and provides mechanistic insights into human disease process. J. Am. Heart Assoc. 4, e001733 (2015). 17. Ni, L. et al. Atrial-specific gene delivery using an adeno-associated viral vector. Circ. Res. 124, 256–262 (2019). 18. Nattel, S. Molecular and cellular mechanisms of atrial fibrosis in atrial fibrillation. JACC Clin. Electrophysiol. 3, 425–435 (2017). 19. Balabanova, S., Kruse, B. & Wolf, A. S. Calcitonin secretion by human placental tissue. Acta Obstet. Gynecol. Scand. 66, 323–326 (1987). 20. Fraioli, F. et al. Beta-endorphin, Met-enkephalin, and calcitonin in human semen: evidence for a possible role in human sperm motility. Ann. NY Acad. Sci. 438, 365–370 (1984). 21. Schuetz, P., Daniels, L. B., Kulkarni, P., Anker, S. D. & Mueller, B. Procalcitonin: A new biomarker for the cardiologist. Int. J. Cardiol. 223, 390–397 (2016). 22. Burstein, B., Libby, E., Calderone, A. & Nattel, S. Differential behaviors of atrial versus ventricular fibroblasts: a potential role for platelet-derived growth factor in atrial-ventricular remodeling differences. Circulation 117, 1630–1641 (2008). 23. Andreassen, K. V. et al. Prolonged calcitonin receptor signaling by salmon, but not human calcitonin, reveals ligand bias. PLoS One 9, e92042 (2014). 24. Hay, D. L., Garelja, M. L., Poyner, D. R. & Walker, C. S. Update on the pharmacology of calcitonin/CGRP family of peptides: IUPHAR Review 25. Br. J. Pharmacol. 175, 3–17 (2018). 25. Prockop, D. J., Sieron, A. L. & Li, S. W. Procollagen N-proteinase and procollagen C-proteinase. Two unusual metalloproteinases that are essential for procollagen processing probably have important roles in development and cell signaling. Matrix Biol. 16, 399–408 (1998). 26. Zhang, S., Kaplan, F. S. & Shore, E. M. Different roles of GNAS and cAMP signaling during early and late stages of osteogenic differentiation. Horm. Metab. Res. 44, 724–731 (2012). 27. Delaunay, M., Osman, H., Kaiser, S. & Diviani, D. The role of cyclic AMP signaling in cardiac fibrosis. Cells 9, 69 (2019). 28. Nakamura, F., Song, M., Hartwig, J. H. & Stossel, T. P. Documentation and localization of forcemediated filamin A domain perturbations in moving cells. Nat. Commun. 5, 4656 (2014). 29. Hay, D. L., Christopoulos, G., Christopoulos, A., Poyner, D. R. & Sexton, P. M. Pharmacological discrimination of calcitonin receptor: receptor activity-modifying protein complexes. Mol. Pharmacol. 67, 1655–1665 (2005). 30. Surinkaew, S. et al. Exchange protein activated by cyclic-adenosine monophosphate (Epac) regulates atrial fibroblast function and controls cardiac remodelling. Cardiovasc. Res. 115, 94–106 (2019). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. © The Author(s), under exclusive licence to Springer Nature Limited 2020 1Division of Cardiovascular Medicine, Radcliffe Department of Medicine, British Heart Foundation Centre of Research Excellence, University of Oxford, John Radcliffe Hospital, Oxford, UK. 2Research Centre, Montreal Heart Institute and University of Montreal, Montreal, Quebec, Canada. 3Department of Pharmacology and Therapeutics, McGill University, Montreal, Quebec, Canada. 4Cardiovascular Research Institute, Baylor College of Medicine, Houston, TX, USA. 5Department of Molecular Physiology and Biophysics, Baylor College of Medicine, Houston, TX, USA. 6Clinical Neurology, Nuffield Department of Clinical Neurosciences, University of Oxford, John Radcliffe Hospital, Oxford, UK. 7Biological Sciences, Faculty of Life and Environmental Sciences, University of Southampton, Southampton, UK. 8Medical Research Council (MRC) Human Immunology Unit, MRC Weatherall Institute of Molecular Medicine, University of Oxford, Oxford, UK. 9Department of Pharmacology and Physiology, Faculty of Medicine, University of Montreal, Montreal, Quebec, Canada. 10Single-Cell Genomics Facility, Weatherall Institute of Molecular Medicine, University of Oxford, Oxford, UK. 11King’s British Heart Foundation Centre, King’s College London, London, UK. 12Department of Medicine, Austin Health, The University of Melbourne, Heidelberg, Victoria, Australia. 13Cardiothoracic Surgery, Oxford Heart Centre, John Radcliffe Hospital, Oxford, UK. 14Department of Medicine, Baylor College of Medicine, Houston, TX, USA. 15Institute of Pharmacology, West German Heart and Vascular Center, University Duisburg– Essen, Essen, Germany. 16IHU LIRYC, Fondation Bordeaux Université, Bordeaux, France. 17These authors contributed equally: Lucia M. Moreira, Abhijit Takawale. 18These authors jointly supervised this work: Stanley Nattel, Svetlana Reilly. ✉e-mail: svetlana.reilly@cardiov.ox.ac.uk


Methods
No statistical methods were used to predetermine sample size. The experiments were not randomized; investigators were blinded to allocation during experiments and outcome assessment.


Patient cohorts
Studies involving human participants were approved by the local Research Ethics Committee (South Central–Berkshire B Research Ethics Committee, UK; ref. nos. 18/SC/0404 and 07/Q1607/38). All patients gave informed written consent. A total of 156 patients were included in the study; all patients underwent cardiac surgery (coronary artery bypass grafting or valve repair/replacement) in the John Radcliffe hospital at Oxford. Detailed patient characteristics are shown in Extended Data Table 1. Right atrial biopsies were collected before cardiopulmonary bypass and immediately processed for cell isolation (described below) or snap-frozen until use in other experiments (for example, gene expression and immunoblotting).


Animal models
All animal breeding, handling and experimental work were carried out at three centres: the Montreal Heart Institute (Montreal, Canada), The Department of Medicine, Austin Health, University of Melbourne (Melbourne, Australia) and the Baylor College of Medicine (Houston, USA). Mice with global CTR knockout (CTR-KO mice) in a C57BL/6J background were generated as described previously15. Ten- or twelve-week-old age- and sex-matched mice (121 in total) were used in all animal experiments. CTR-KO mice were compared to their control littermates (heterozygous CTR-floxed); females and males were analysed separately for some experiments (depicted in Fig. 4a–c and Extended Data Fig. 7a–l). All animal work was performed in accordance with the local (Montreal Heart Institute and Austin Health) Animal Care and Ethics Committee guidance and in accordance with US National Institutes of Health, the Canadian Council on Animal Care and the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. The CTR-KO and control mice (Montreal cohort) were housed in Allentown XJ cages at 20–22 °C, 50% humidity and 60 air changes/ hour ventilation conditions. Diet consisted of the sterilized food (2019S, Envigo) and osmotic water. The CTR-KO and control mice (Melbourne cohort) were housed in a specified-pathogen-free facility at 22 °C, in a 12-h light/dark cycle, and were supplied with standard irradiated mouse chow (1.2% calcium and 0.96% phosphorus; Ridley Agriproducts) and water ad libitum. Breeding mice were housed in micro-isolator cages and offspring used for experiments were transferred to open-top cages at weaning (3–5 mice/ cage). Cages contained corn-cob bedding, and cardboard tubes and tissues were supplied for environmental enrichment. Studies in LKB1 knockdown (LKB1-KD), LKB1 CT double-knockdown (LKB1/CT-dKD), LKB1 knockdown/CT overexpression (LKB1-KD/+CT) and control mice (Houston cohort) were performed according to protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the Baylor College of Medicine. All mice were housed in standard mouse cages provided with bags of sizzle nest as cage enrichment and were fed standard feeder chow as approved by the IACUC and recommended by ‘the Guide’ (NIH Publication 85-23, revised 1996). The Montreal mouse studies were not formally randomized, as the groups were determined by breeding rather than by intervention. The mice were studied consecutively as soon as they reached maturity upon breeding. All experimental studies (for example, in vivo electrophysiology studies, qPCR, echocardiography and assessment of atrial fibrosis) were performed by researchers blinded to mouse genotype. The Houston mouse studies were performed in mice of the same genotype (Lkb1fl/fl). Pups from same litter and same cage were injected as early as day 5 with the same virus and the males and females were separated at the time of weaning. Subsequently, litters were randomly allocated to each group, keeping overall numbers of mice for all the groups similar. Animals of both sex were studied. For in vivo studies, genotype was not disclosed to the investigators performing data collection and those generating quantitative readouts. Experimenters were blinded to the group allocation during data collection and analysis (Montreal and Houston cohorts).


Generation of LKB1/CT-KD and LKB1-KD/+CT mice
The Lkb1fl/fl mice were purchased from Jackson Laboratory (strain no. 014143, Lkb1fl; Jackson Laboratory). The shRNA for mouse Calca (TRCN0000184797; Sigma-Aldrich) was embedded within an miR-30a scaffold in an AAV9 vector containing the Cre recombinase gene under the regulation of the ANF promoter to facilitate its transcription by polymerase II (Extended Data Fig. 7l). As described previously17, 5 × 1010 genome-containing units of AAV9 were diluted in lactated Ringer’s solution and administered subcutaneously to 5-day-old pups. Mice injected with equal volume of lactated Ringer’s solution were used as negative controls. For the ease of identification, all the pups from one litter, one cage were injected with the same AAV9 or Ringer’s solution and returned to the cages to be nursed. Mice were weaned at the age of 21 days and males and females were separated into designated cages. A total of 38 mice were used for the final experiments.


Collection of mouse tissue
Mice were weighed and anaesthetized using isoflurane and euthanized via cervical dislocation. Hearts were extracted quickly and dipped once in clean saline solution to remove excess blood. For immunoblot and qPCR experiments, atria were separated from the ventricles. Left and right atria as well as ventricles were weighted, stored in respective tubes and flash frozen in liquid nitrogen. For histology, whole hearts were dipped into another container with clean saline solution (for Houston cohort) or arrested in diastole with 1 M KCl (for Montreal cohort), fixed in 10% neutral buffered formalin (HT501128; Sigma-Aldrich) for at least 24 h and embedded in paraffin. Isolation and culture of primary human ACFs Human ACFs were isolated and cultured from right atrial biopsies obtained from patients undergoing cardiac surgery. Tissue biopsies were cut into small (2–3 mm3) pieces and repeatedly digested using 4 mg/ml collagenase II and 0.0625% trypsin). Cells were washed twice with sterile phosphate-buffered saline (PBS) and plated onto 6-well plates in FBM-3 medium (CC-3131, Lonza) containing 10% FBS and a supplement pack (CC-4525, Lonza) and were kept in a humidified atmosphere at 37 °C and 5% CO2. The medium was renewed every 2–3 days. At ~80–90% confluence, cells were passaged using a standard trypsinization method. For the experiments with TGFβ1 and CT stimulation, we used commercially available donors of human primary atrial fibroblasts (CC-2903, Lonza), which were maintained and cultured in the same medium as outlined above. Experiments were carried out at cell passages P3–P4 and cells were cultured in serum-free medium for ~16 h before intervention and treatment with 100 or 500 nM of human CT (H-2250, Bachem), 10 or 100 nM of human αCGRP (3012, Tocris Bioscience), 10 nM of BIBN4096 (4561,Tocris Bioscience), 100 nM of sCT8-32 (4037182, Bachem,) or 10 ng/ml of TGFβ1 (HZ-1011, Proteintech). Isolation and maintenance of primary human ACMs Right atrial cardiomyocytes (ACMs) were isolated using a standard enzymatic dispersion technique, as described previously31 and detailed in Supplementary Methods section 1.1. Sources of other human cells and cells lines are detailed in Supplementary Methods section 1.2. Authentication of cell lines was carried by the manufacturer as outlined below: ATCC CRL-1803, https://www.lgcstandards-atcc.org/products/all/ CRL-1803.aspx?geo_country=gb#specifications;


Article
ATCC CRL-1573, https://www.lgcstandards-atcc.org/products/all/ CRL-1573.aspx#specifications; Gibco C-0135C, https://www.thermofisher.com/order/catalog/product/C0135C#/C0135C; Innoprot P30136, https://innoprot.com/product/hitseekercalcitonin-receptor-cell-line/. In-house cell authentication was performed by monitoring cell morphology and phenotypical tests (for example, immunostaining and flow cytometry for CTR protein and ELISA for CT secretion). Cell lines were certified by the manufacturer. Mycoplasma contamination was tested negative using immunofluorescence staining with DAPI (for example, as shown in the manuscript, Figs. 2a and 3g, and Extended Data Figs. 6b and 8b–d). Transfection of primary human ACFs Silencing of the CTR was carried out in ACFs transfected with 50 nM of antisense LNA (locked nucleic acid) oligonucleotides targeting CTR (300600, Exiqon; design 1, C*T*G*G*G*T*G*C*G*C*T*A*A*A*T*A, and design 2, A*T*G*A*C*A*T*A*G*A*T*G*A*G*A*C; LNA is not shown, as this information is proprietary), or antisense LNA oligonucleotide negative control A (300610, Exiqon) using Lipofectamine RNAiMAX transfection reagent (13778075, ThermoFischer Scientific) in antibiotic-deprived FBM-3 medium containing 2% FBS (both from Lonza and detailed above). Efficient knockdown was confirmed by the real-time qPCR and western blot (Extended Data Fig. 1l, m).


Western blotting
Immunoblotting is described in Supplementary Methods section 1.3. The list of antibodies and validation of the anti-CTR/anti-pro-CT antibodies are shown in Supplementary Table 1 and Extended Data Fig. 8b–g, respectively.


Colorimetric assays
Quantification of total secreted collagen in the cell culture supernatant was performed using a Sirius Red collagen detection kit (9062, Chondrex) as previously described32. The levels of human CT in cell supernatant was quantified using ELISA (CEA472Hu, Cloud-Clone) with a detection range of 12.35–1,000 pg/ml and lowest detectable level of less than 4.74 pg/ml; experimental recovery of cellular secretome matrix was 98% on average. This kit did not show any cross-reactivity with αCGRP or pro-CT (Extended Data Fig. 8h, i). Concentration of human pro-CT was measured with an ELISA kit (ab221828, Abcam). Concentration of human αCGRP was measured with an enzyme immunoassay (EIA) kit (A05481.96, BioVendor, BertinPharma) with a detection limit <10 pg/ml. The amount of total collagen in human ACFs was quantified by colorimetric detection of hydroxyproline using a QuickZyme total collagen assay kit (QZBTOTCOL1, lot 0795, QuickZyme Biosciences). Quantification of the human collagen 1 C-terminal telopeptide (ICTP) was carried out using an ELISA kit (CSB-E10363h, Cusabio). Cyclic adenosine monophosphate (cAMP) was quantified using a HitHunter cAMP Assay for Small Molecules kit (90- 0075SM2, DiscoverX-Eurofins); cAMP was measured in the presence or absence of the selective inhibitor of Gαs protein NF499 (4,4′,4″,4‴-(carbonylbis(imino-5,1,3-benzenetriylbis(carbonylimino))) tetrakis-benzene-1,3-disulfonic acid33; 10 μM, N4784, Sigma-Aldrich), the Gαi inhibitor pertussis toxin (PTX, 20 ng/ml, CAS 70323-44-3, Calbiochem), human CT (100 nM, H-2250, Bachem), a CTR antagonist (salmon calcitonin sCT8-32, 100 nM, 4037182, Bachem) and the cAMP activator forskolin (FSK, 100 μM, 1099, Tocris Bioscience). Concentrations of the selected non-ECM proteins CTGF, CCL2, TNFα and IGF-II secreted by human ACFs treated with 100 nM CT for 72 h were assessed with ELISA kits DY9190-05, DY279-05, HSTA00E and DY292-05, respectively (all from R&D Systems). All colorimetric assays were performed according to the manufacturer protocols. Accumulation of calcium-rich deposits by fibroblasts was assessed by Alizarin Red S staining (A5533, Sigma-Aldrich) as detailed in Supplementary Methods section 1.4. Immunostaining and imaging of human ACFs Immunostaining for CTR was carried out in human ACFs. In brief, cells were fixed in precooled (−20 °C) acetone/methanol (1:1) solution, air-dried and rinsed three times in PBS, blocked with serum-free blocking reagent (X090930-2, DAKO, Agilent Technologies) and incubated with anti-CTR and anti-filamin A (detailed in the Supplementary Table 1) antibodies overnight at 4 °C. After multiple rinsing steps with PBS, secondary antibodies conjugated to Alexa Fluor (Invitrogen) were applied for 2 h at room temperature. Imaging was performed with a Zeiss LSM 710 or Leica DM 6000 CFS confocal imaging system. To assess cellular localization of the CTR, optical sections of fibroblasts were imaged with a frame size of 157 μm × 157 μm at a z-depth of 1 μm and pixel resolution of 0.09 μm × 0.09 μm. Channels were subsequently split and then merged in the Fiji open source software. BMP1 enzyme activity assay BMP1 enzyme activity was measured with a fluorescent assay using fluorogenic substrate as detailed in Supplementary Methods section 1.5. Scratch wound migration assay Human ACFs migration was determined using in vitro scratch wound assays on confluent monolayers of cells using chambers with 2 well silicone insert with a defined cell-free gap (80206, Ibidi). In brief, 5 × 103 cells were seeded into each chamber in 70 μl of complete medium (with 10% FBS, as described above). When cells attached and reached ~95% confluency, they were synchronized in serum-free medium for 16 h, and then the chamber insert was removed; cells were then subjected to 24-h treatment with 100 nM CT and/or 10 ng/ml TGFβ1. Changes in the wound area were imaged at 0 and 24 h and quantified using ImageJ software.


Scar-in-a-jar assay
Collagen 1 accumulation in fibroblasts was assessed using a scar-in-a-jar assay detailed in Supplementary Methods section 1.6.


Assessment of cell proliferation
Cell proliferation at a single time point was assessed by ELISA using BrDU (5-bromo-2′-deoxyuridine) DNA-binding probe (QIA58, Calbiochem, Millipore) according to the manufacturer’s instructions. In brief, human ACFs were plated in a sterile 96-well plate in a medium (FBM-3 33-3131, Lonza) containing 10% FBS and supplement pack (CC-4525, Lonza). Cells were incubated overnight with BrdU (kit component JA1595) and fixed the next morning with the Fixative/Denaturing Solution (kit component JA1598). Anti-BrdU antibody (kit component JA1599) diluted 1:100 in antibody diluent (kit component JA1604) was added in each well and incubated for 1 h at room temperature; this was followed by three washes with a wash buffer (kit component JA1617) before a 30-min incubation of cells with peroxidase goat anti-mouse IgG (kit component JA1618) and reconstitution with conjugate diluent (kit component JA1615) followed by three more washes with wash buffer and deionized water. Cells were then incubated for 15 min in the dark at room temperature with the substrate solution and then with the stop solution. Spectrophotometric detection was performed at a wavelength of 450 nm. Real-time proliferation was measured using an xCELLigence real-time cell analysis (RTCA) DP system (ACEA Biosciences) to monitor cell response in real-time mode, as previously described33. The latter setup was also used to record impedance to monitor CTR response to the ligand binding, as previously described. The data were analysed using the manufacturer’s software, RTCA DA v1.0. Real-time quantitative or non-quantitative polymerase chain reaction (PCR) Total RNA isolation, reverse transcription and quantitative and nonquantitative PCR are detailed in Supplementary Methods section 1.7. Primer sequences and TaqMan assay IDs are listed in Supplementary Table 3. Histological assessment of atrial fibrosis in mice Collagen content in mouse hearts was assessed by staining with Masson’s trichrome as detailed in Supplementary Methods section 1.8.


Echocardiography of the mouse heart
Echocardiographic studies were performed as described previously34 and detailed in in Supplementary Methods section 1.9. In vivo assessment of AF inducibility using trans-jugular electrostimulation in mice Assessment of susceptibility to AF was carried out in control (heterozygous CTR-floxed) and CTR-KO mice using iox2 software (v.2.8.0.13, EMKA technologies, FR). Mice were anaesthetized with an isoflurane and oxygen mixture and positioned on a temperature-regulated operating table. In brief, platinum electrodes were inserted into the limbs for ECG measurement, and a 1.9 French Octapolar (Transonic) catheter was inserted into the right jugular vein and positioned in the right atrium. After a baseline stable ECG recording, rectangular stimulus pulses of twice the pacing threshold were applied with a multiprogrammable stimulator (ID). Atrial effective refractory period (ERP) was measured by delivering 7 (or 8) stimuli (S1) at a fixed cycle length of 100 ms followed by one short coupled extra stimulation (S2) at a coupling interval ranging from 70 ms to 20 ms, with 2-ms decrements for precise atrial ERP estimation. AF inducibility was determined with 50-Hz burst pacing for 3 s, with six bursts separated by 2-s intervals; the cycle was repeated three times. AF was defied as a rapid, irregular atrial rhythm. Once AF was induced, pacing was immediately stopped to avoid interfering with the induced arrhythmias. AF duration was calculated as a mean duration of all induced AF episodes in each mouse. Surface ECG & catheter signals were recorded and analysed using iox2 software (v.2.8.0.13, EMKA Technologies). The experimenter was blinded to the genotype throughout the protocol and analysis.


Surface ECG recording in mice
Mice of 3–4 weeks of age (after gaining sufficient body size) were anaesthetized with isoflurane and placed on the Rodent Surgical Monitor with two sets of Noninvasive ECG Electrodes (Indus Instruments), with animal limbs being taped to the electrodes. Isoflurane was provided constantly through the nose cone to ensure that the mouse remained asleep throughout the recording. The temperature of the ECG board was adjusted in order to maintain a constant the body temperature (monitored by a rectal temperature probe) in a range between 36.5 °C and 37.5 °C. The ECG tracing and recordings were acquired for 20 min per mouse, a minimum of once a week, with the IOX2.9.5.28 software (Emka Technologies). AF was defined by the absence of p-waves and the presence of irregularly irregular R–R intervals for a period of more than 10 s.


Proteome profiling
Processing of conditioned medium and deglycosylation were conducted as previously described35 as detailed in Supplementary Methods section 1.10. In-solution protein digestion and peptide clean-up are described in detail in Supplementary Methods section 1.10. For liquid chromatography and tandem mass spectrometry (LC-MS/ MS), cleaned peptides were separated on a nanoflow LC system (Thermo Scientific Dionex UltiMate 3000 RSLCnano) as described in Supplementary Methods section 1.10. In the database search of LC-MS/MS data and data filtering, Proteome Discoverer software (ThermoFisher Scientific, version 2.3.0.523) was used to search raw data files against a hybrid human–bovine database (UniProtKB/Swiss-Prot version of January 2019) using Mascot (Matrix Science, version 2.6.0) as described in Supplementary Methods sections 1.10 and 1.16. Correlation between technical duplicates is shown in Extended Data Fig. 8j.


Flow cytometry
Human cultured or freshly isolated ACFs were sorted on a Becton Dickinson (BD) FACS Aria Fusion III sorter using a 100 μm nozzle and FACSDiva software v.8 (detailed in in Supplementary Methods section 1.11).


Single-cell RNA-sequencing (scRNA-seq) of human ACFs
Freshly isolated cells were used in SMART-seq2 assays, while cultured ACFs were processed by a droplet-based 10× scRNA-seq procedure. SMART-seq2 work flow. Freshly isolated cells were resuspended in ice-cold PBS containing 3% BSA and stained with DAPI, and viable singlets were sorted on a BD FACS Aria Fusion-III sorter (using FACSDiva v.8.0 software) into 96-well plates containing 4 μl SMART-seq2 lysis buffer and then frozen at –80 °C until needed for further processing. The released RNA was converted to cDNA and then sequence ready libraries as described previously36, with minor modifications. ThermoFisher Superscript II reverse transcriptase and Roche Kapa PCR enzyme were substituted for Takara Smartscribe reverse transcriptase and SeqAmp PCR enzyme, respectively. Twenty PCR cycles were used to amplify cDNA, and an Illumina Nextera XT kit was used to generate the sequence-ready libraries. The 384 single cells were sequenced as a single pool on the Illumina Nextseq 500 system using a high-output 75-bp kit. SMART-seq2 scRNA-seq data analysis. Raw SMART-seq2 sequencing data were demultiplexed using Illumina bcl2fastq software (v.2.20.0.422) as described in detail in Supplementary Methods sections 1.12 and 1.16. Droplet-based 10× scRNA-seq workflow. Cultured human ACFs were quickly and gently trypsinized and resuspended in ice-cold PBS containing 3% BSA and stained with DAPI, and only viable singlets were sorted on a BD FACS Aria Fusion III sorter into individual Lo-bind tubes. Cells were resuspended in 100 μl of staining buffer (3% BSA, 0.01% Tween and PBS), incubated with a serum-free blocking reagent (DAKO) for 10 min at 4 °C and labelled (20 min at 4 °C) with unique Biolegend Total-seq A hashing antibodies (1 μg/ml, detailed in Supplementary Table 2) diluted in a staining buffer. After three washing steps with a staining buffer, cells were centrifuged at 4 °C for 5 min at 350g and all samples were merged at equal ratios in 1 ml of a staining buffer, centrifuged for 5 min at 350g at 4 °C, resuspended in ice-cold PBS at ~1,000 cells/μl and immediately processed with a 10X Genomics Chromium B chip; cells were kept on ice through the whole procedure. The sample exome library was processed to a sequence ready library using the V3 3′ Prime Gene Expression kit as per manufacturer’s protocol. The hashing library was processed as per hashing method version 2019- 02-13, New York Genome Center Technology Innovation Lab (www. CITE-SEQ.com). Both libraries were pooled before sequencing on an Illumina Novaseq 6000. Analysis of the droplet-based 10× scRNA-seq data. Raw sequence reads were quality-checked using FastQC software (v.0.11.8)37 using the human hg38 reference genome analysis set obtained from the University of California Santa Cruz (UCSC)38 ftp site). Further details are described in Supplementary Methods section 1.13.


Article
Data analysis and sample demultiplexing of the droplet-based 10× scRNA-seq data. Hashed samples were demultiplexed as described in detail in Supplementary Methods sections 1.14 and 1.16.


Gene expression microarrays
Microarrays were performed on human ACFs treated with 100 nM CT or vehicle for 72 h as described in Supplementary Methods sections 1.15 and 1.16.


Statistical analysis
Student’s t-test was used in two-group comparisons of normally distributed data; normal distribution was assessed by Kolmogorov–Smirnov test. Multiple groups of normally distributed data of similar variance were compared by one-way or two-way ordinary or repeated-measures ANOVA; for multiple comparisons, P values corrected with the Sidak or Holm–Sidak test are shown as appropriate. The Kruskal–Wallis or Mann–Whitney U-tests were used when the normality assumption was not met. Categorical variables were compared by one-sided Fisher’s exact test. Age–CT relationship was analysed by Pearson correlation test. Analysis of AF-free survival was performed using log-rank (Mantel–Cox and Gehan–Breslow– Wilcoxon) tests applied to Kaplan–Meier survival curves. A value of P < 0.05 was considered statistically significant. Statistical analysis was performed using GraphPad Prism v7.05, v6.04, v8.02 or v8.04 software. Proteomic results were analysed as follows: each data set was filtered to keep only the consistently quantified proteins defined as the ones with less than 30% missing values. All remaining missing values were imputed with the KNNImpute method using the default k value (k = 3). The relative quantities of the proteins were scaled using log2 transformation. The Limma package v3.34.5 has been used to compare between different phenotypes using the Ebayes algorithm and performing paired analysis when paired samples were available. The initial P values were corrected for multiple testing using Benjamini–Hochberg false discovery rate (FDR) correction method. Functional enrichment analysis was conducted in a DAVID 6.8 web tool with the human proteome as background. The scRNA-seq data sets were analysed using R package software, as outlined in Supplementary Methods sections 1.12–1.14. A detailed list of the software packages used is provided in Supplementary Methods section 1.16.


Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this paper.


Data availability
All data generated or analysed during this study are included in the Article and its Supplementary Information. The scRNA-seq data are deposited in the GEO database (accession codes GSE148506, GSE148507 and GSE148504). Source data are provided with this paper. 31. Reilly, S. N. et al. Up-regulation of miR-31 in human atrial fibrillation begets the arrhythmia by depleting dystrophin and neuronal nitric oxide synthase. Sci. Transl. Med. 8, 340ra74 (2016). 32. Schafer, S. et al. IL-11 is a crucial determinant of cardiovascular fibrosis. Nature 552, 110– 115 (2017). 33. Hohenegger, M. et al. Gsalpha-selective G protein antagonists. Proc. Natl Acad. Sci. USA 95, 346–351 (1998). 34. Alsina, K. M. et al. Loss of protein phosphatase 1 regulatory subunit PPP1R3A promotes atrial fibrillation. Circulation 140, 681–693 (2019). 35. Barallobre-Barreiro, J. et al. Extracellular matrix remodelling in response to venous hypertension: proteomics of human varicose veins. Cardiovasc. Res. 110, 419–430 (2016). 36. Picelli, S. et al. Full-length RNA-seq from single cells using Smart-seq2. Nat. Protoc. 9, 171–181 (2014). 37. Andrews, S. FastQC: a quality control tool for high throughput sequence data. http:// www.bioinformatics.babraham.ac.uk/projects/fastqc (2010). 38. Kuhn, R. M., Haussler, D. & Kent, W. J. The UCSC genome browser and associated tools. Brief. Bioinform. 14, 144–161 (2013). Acknowledgements We thank the Oxford Genomics Centre at the Wellcome Centre for Human Genetics (funded by Wellcome Trust grant reference 203141/Z/16/Z) for the generation and initial processing of the ACF microarrays data; M. Farrall for the help with statistics; K. Clark in the WIMM Flow Cytometry Facility for his help; J. Digby for assistance with analysis and detection of CT by ELISA in human ACFs and ACMs; C. St-Cyr for managing, handling and genotyping mouse colonies at the Montreal site; R. Hiram for initial help with EP analysis in mice; J. Dewing for creating an artistic sketch summary of the main findings; L. E. Schmidt and X. Yin for help with the proteomic experiments; S. Farid and V. Srivastava for help with collection of some human atrial specimens during revision; P. Wookey for advice on CTR protein detection; R. Wijesurendra and P. Gajendragadkar for initial help in obtaining patient consent for the study; and A. Recalde and M. C. Carena for initial help with optimizing the fibroblast isolation protocol. Funded by the British Heart Foundation (BHF) Intermediate Fellowship in Basic Science, the Oxford BHF Centre of Research Excellence (CRE; RG/13/1/30181) Transitional Fellowship, a BHF CRE Overseas Collaboration Travel award, the Medical Science Division Internal Fund, the Wellcome Trust Institutional Strategic Support Fund, the Oxfordshire Health Services Research Committee, the National Institute for Health Research (NIHR) Oxford Biomedical Research Centre and LAB282 grants (to S.R.); BHF Chair award CH/16/1/32013 (to K.M.C.); the Canadian Institutes of Health Research (CIHR) and Heart and Stroke Foundation of Canada (to S.N.); and Fonds de Recherche en Santé de Québec (FRQS) and CIHR postdoctoral fellowships to A.T. Author contributions S.R. and S.N. conceived the study, designed the experiments and wrote and edited the manuscript. L.M.M. and A.T. wrote parts of the manuscript and carried out and analysed most of the experimental work. K.M.C. provided intellectual input on experiments using clinical samples. N.E., P.R., A.J.S. and C.R. performed some PCR, ELISAs and HPA measurements in human samples. Imaging and analysis of the CTR cellular localization in human ACFs and HEK293 cells were performed by D.A.M. and L.M.M. All in vivo work in mice was carried out by A.T., M.H. and S.K.L., supervised by X.H., T.W. and S.N. Staining, imaging and analysis of fibrosis in mouse heart sections were done by A.T., M.H., S.K.L. and C.P.; M.S. provided full access and supervision of the histological and imaging experiments at the Montreal site. Primer design and gene expression assays in mice were carried out by P.N. and M.H.; primer design for human transcripts was performed by N.E., L.M.M. and C.P. Functional electrophysiological studies in CTR-KO mice were carried out by M.-A.G. and A.T., supervised by S.N.; echocardiography in CTR-KO mice was supported by J.C.T. The CTR-KO mice were generated and provided by J.D.Z. and R.A.D., who also supervised mouse tissue collection at the Melbourne site, genotyping and analysis of the selected morphologic parameters carried out by M.V.C. and P.K.R. in Melbourne. All experimental work and data analysis in the LKB1-based mice was carried out by M.H. and S.K.L., supervised by X.H. and T.W. Experiments in human cells were done by L.M.M., N.M. and N.E. The proteomic study was designed, executed and analysed by J.B.B. and K.T. under the supervision of M.M. Transcriptome profiling by scRNA-seq, performed by N.A., L.M.M. and N.M., was designed and supervised by A.M. and S.R. Bioinformatics analysis of scRNA-seq results was carried out by A.A., supervised by A.J.S. Scar-in-a-jar assay and analysis was carried out by A.L., supervised by S.B. Patient consent was obtained by N.M., L.M.M. and M.N.; human atrial biopsies were collected by cardiac surgeons R.S. and G.K. under ethical approval granted to B.C. and S.R. All authors discussed the results and had the opportunity to comment on the manuscript. Competing interests The authors declare no competing interests. Additional information Supplementary information is available for this paper at https://doi.org/10.1038/s41586-0202890-8. Correspondence and requests for materials should be addressed to S.R. Peer review information Nature thanks Igor R. Efimov, Jeffery D. Molkentin, Andrew F. Russo and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Reprints and permissions information is available at http://www.nature.com/reprints. Extended Data Fig. 1 | See next page for caption.


Article
Extended Data Fig. 1 | Effects of human αCGRP on human ACF function. a, Secretion of αCGRP (ELISA) by human ACMs vs TT cells. b, αCGRP protein (immunoblot) in human right atrial tissue lysates obtained from patients with SR or AF. c–e, mRNA of human ACM CT, αCGRP or CT/αCGRP ratio between SR and AF groups. f, Correlation between donors’ age and ACM-CT secretion (ELISA over 4–6 h); 95% CI = –0.7912 to 0.01258, R = –0.4862, R2 = 0.236, P = 0.056 by Pearson correlation test. g–i, Human atrial myocardium (g) expresses CTR isoform 1a, but not 1b (PCR using specific isoform primers) and CTR protein (h; TT cells, positive control; see Extended Data Fig. 8a); CTR protein content in ACFs shown in i. j, Representative traces (real-time impedance assay) showing CT-induced concentration-dependent increase of the baseline-normalized cell index (CI). k, Total and phosphorylated ERK were not altered by CT (immunoblotting). l, m, CTR mRNA (qRT–PCR) and protein content (immunoblotting) were decreased in human ACFs with CTR knockdown due to LNA antisense oligonucleotides (designs LNA-aCTR1 and LNA-aCTR2); fc, fold change versus the CTR-NC control. n, Effect of 10 and 100 nM αCGRP on 72-h collagen accumulation (by Sirius red) in ACF secretomes. o–q, Effect of 500 nM CT in human ACFs stimulated with TGFβ1 (10 ng/ml) on cell migration (o; fc, fold change versus vehicle at 0 h), collagen content in conditioned medium (p) and cell proliferation (q). r, Effect of 10 nM αCGRP on 72-h collagen 1 accumulation in human ACF cell lysates and secretomes; representative blots (left) and quantification (right); n, individual participants; fc, fold change versus control. Data are presented as mean ± s.e.m., except in a, b (pro-αCGRP), c–e, k, m, panels 2, 3 of r (medians and interquartile ranges), n (mean with paired scattered dots), o (mean ± s.d.). P values were determined by two-sided tests: unpaired t-test (b, αCGRP, p–q), Mann–Whitney U-test (b, Pro-αCGRP, c, e), Friedman test (n, panels 2, 3 of r), Kruskal–Wallis test with Dunn’s correction (a, k) and repeated-measures one-way ANOVA with Sidak correction (l, o, panels 1, 4 of r). Data are pooled from individual donors assessed in single replicates (a, b, f, g–k, m, o–r) or duplicates (c–e, l, n); all results were reproduced independently twice. For gel source data, see Supplementary Fig. 1. Extended Data Fig. 2 | See next page for caption.


Article
Extended Data Fig. 2 | Effect of CT on collagen 1 processing and single-cell transcriptome (10× scRNA-seq) of cultured human ACFs. a–d, Effect of CT on synthesis of collagen 1 (a) and 3 (b) (by qRT–PCR) and on extracellular (c) and intracellular (d) content of collagen 1 C-terminal telopeptide (ICTP). e, Representative blots (left) and quantification (right) of unprocessed (pro-Col, pro-collagen; pc-Col, pC collagen) and processed collagen 1 (Col 1) in human ACFs treated with 100 nM CT (fc, fold change versus vehicle). f–h, Effect of exogenous CT on bone morphogenetic protein 1 (f; BMP1, immunoblotting), BPM1 gene expression (g; qRT–PCR) and BMP1 activity (h) in the presence or absence of BMP1 inhibitor (BMP1 inh; RFU, relative fluorescence units). i, j, Effect of 24-h 500 nM CT on collagen-1 (Col1A1) mRNA (qRT–PCR) and C-terminal telopeptide (ICTP by ELISA). Data are mean ± s.e.m., except in b, g, j (medians with interquartile ranges); n, individual participants. Two-sided tests: unpaired t-test (a, c–f, i), Mann–Whitney U-test (b, g, j) and one-way ANOVA with Sidak correction (h). Data are pooled from individual donor cells assessed in single replicates, or duplicates in a, b, g, i, on the same day in one batch. Results were reproduced twice (a–c, f–h) in different donors. For gel source data, see Supplementary Fig. 1. k–n, Unbiased transcriptional clustering of scRNA-seq data from human ACFs cultured with 100-nM CT for 24 h or vehicle; demultiplexed by final cell count per hash-tag in (k), transcriptional clusters in (l), pharmacological intervention in (m) and by each donor in (n); D1–D6 indicate individual donors. Active cycling cells are pointed by arrow. All data are colour-coded within the figure. Data are pooled from 6 individual donors in SR assessed in 14,742 cells (after quality control after filtering the initial 18,466 total cellular barcodes) on the same day in one batch. tSNE, t-distributed stochastic neighbour embedding; UMAP, Uniform Manifold Approximation and Projection. Source data for k–n have been deposited in the GEO database. Extended Data Fig. 3 | See next page for caption.


Article
Extended Data Fig. 3 | CTR expression and CT-mediated changes in ACF. a–d, Effects of 72-h treatment with 100 nM CT on IGF-II, CCL2, CTGF and TNFα in human ACF conditioned medium. Data are pooled from individual donor cells assessed on the same day in technical duplicates, repeated twice; n, individual donors. P values were calculated by two-sided tests: paired t-test (a–c) and Wilcoxon test (d). e, GO enrichment analysis (David 6.8 web-tool) of the differentially expressed ACF proteins under the above GO terms stratified by adjusted P values. The bold number next to each GO term represents the number of genes under each term. The original data used for this analysis were pooled from 6 individual donors treated with vehicle or 72-h 100 nM CT assessed in single replicates on the same day in one batch. f–k, Representative blots of the CTR protein (f, h, j) and gene expression (qPCR, g, i, k) in human ACFs from patients with AF or SR. l–n, Effects of CT treatment of ACFs with persistent AF on fibronectin (l), α-SMA protein (m) and cell migration (n) by scratch wound assay (fc, fold change). Data are mean ± s.e.m., except in l, n (medians with interquartile ranges), a–d (means and linked paired samples); n, individual donors. P values were determined by two-sided tests: paired t-test (a–c), unpaired t-test (f–k, m), Wilcoxon test (d), Mann–Whitney U-test (n), and Kruskal–Wallis test with Dunn’s correction (l). Data were pooled from individual donors (l) or separate days (m, n) and are assessed in single replicates on the same day in one batch apart from n (single replicates on two different days), or in duplicates in assessed on the same day (g, i, k). Findings in a–d, j were validated by another method (Fig. 3a, b, g, Extended Data Fig. 6b). All except e were reproduced twice in cells from different donors. For gel source data, see Supplementary Fig. 1. Extended Data Fig. 4 | Single-cell transcriptome of freshly isolated human ACFs (scRNA-seq SMART-seq2). a, b, Transcriptional clustering (a) of freshly isolated human ACFs stratified by donors in b labelled on the graph as SR1–SR4 or AF1–AF4. c–f, Differentially expressed genes (DEGs) in transcriptional cluster 1 (c, d, f) and volcano plots for clusters 2–5 (e; also see Source Data for c–f). P values for DEGs were calculated by a log-likelihood ratio test on a hurdle model (MAST framework tool) and corrected for multiple testing using Benjamini–Hochberg correction (see Supplementary Methods sections 1.12 and 1.16). Data are pooled from 268 single cells isolated from 8 individual donors; scRNA-seq workflow was performed on the same day in one batch.


Article
Extended Data Fig. 5 | Cluster comparison of single-cell transcriptomes (SMART-seq2) of freshly isolated human ACFs. a, Transcriptional clustering of human ACFs (after quality control) pooled from 4 individual donors in SR and 4 individual donors in AF; figure shows the top 10 most abundant genes in each cluster. b, Gene Ontology (GO) functional enrichment analysis for human ACF transcriptional clusters. The number of significantly enriched genes is shown within the figure. The P values for GO panels are generated from a hypergeometric distribution with a Benjamini–Hochberg correction. The original data were pooled from 268 single cells isolated from 8 individual donors; scRNA-seq workflow was carried out on the same day in one batch. Extended Data Fig. 6 | See next page for caption.


Article
Extended Data Fig. 6 | Protein profiling of the selected CT–CTR downstream targets. a, Representative blots showing atrial protein content of BMP1, PKA subunit C (PKAC), PKA subunit R2 (PKAR2), EPAC2, EPAC1, CREB and cAMP in AF (4 individual donors) versus 5 individual control donors in the SR group. All proteins but CREB were assessed in the same membrane after protein stripping; all proteins are normalized to GAPDH and expressed as fold of SR control (fc); the red dotted line indicates y axis value of 1; n, individual donors. Data are presented as medians with interquartile ranges. P values were determined by two-sided Mann–Whitney U-test between the SR and AF groups for each protein. Data are pooled from individual donors assessed in single replicate on the same day; results were reproduced in the same donors twice. For gel source data, see Supplementary Fig. 1. b, Immunofluorescence staining shows predominantly intracellular localization of the CTRs (green) in ACFs obtained from patients with persistent AF. By contrast, in SR-ACFs, the CTR is localized to the cell surface. Cells were counterstained with filamin A (red) and nuclei (DAPI). Data are pooled from the individual donors (a few cells in each field as shown in the figure) collected over 2-year period, assessed on separate days and validated by 3 independent experimenters. For source data, see Supplementary Fig. 1. Extended Data Fig. 7 | See next page for caption.


Article
Extended Data Fig. 7 | Atrial gene expression, morphological parameters and AF-duration in mice. a–d, Global CTR gene deletion does not alter atrial expression of the genes for collagen 1 (Col1A1), collagen 3 (Col3A1), fibronectin (Fn) and alpha-smooth muscle actin (ACTA2) in male and female mice. e–l, Selected morphological parameters in the CTR-KO males and females. m, n, Mean duration of AF in CTR-KO and control mice expressed as ‘mean of all AF episodes in mice that experienced AF’ (m) or ‘mean of all AF episodes in all mice’ (n). o, Schematic representation of the constructs used to generate atrialspecific LKB1-KD, LKB1/CT-dKD and LKB1-KD+CT mice. The Lkb1fl/fl mice were injected with AAV9-ANF-CRE. Because the ANF promoter drives expression of CRE exclusively in the atria, LKB1 was downregulated only in the atria of these LKB1-KD mice. The LKB1-KD+CT cDNA mice received AAV9-ANF-CRE + AAV9ANF-CT cDNA injections. Under the control of the ANF promoter, CT was overexpressed exclusively in the atria of these mice. The LKB1/CT dKD mice received AAV9-ANF-CRE + AAV9-loxP-STOP-loxP-shCT injections. Both LKB1 and LoxP-STOP-LoxP were deleted by atrial-specifically expressed Cre enzyme, allowing the expression of CT shRNA, which selectively targets the CT/pro-CT but not the αCGRP sequence and hence resulted in the downregulation of both LKB1 and CT. Data are presented as mean values ± s.e.m., except in d (females), g (males), medians with interquartile ranges. P values were determined by twosided tests: unpaired t-test in all apart from d (females), g (males), j, m, n, which were analysed by Mann–Whitney U-test; n denotes individual animals. Data are pooled from individual animals assessed in single replicates on the same day and reproduced in two centres in a, c, d. Results in m, n were obtained from individual animals over ~2.5 years. Extended Data Fig. 8 | See next page for caption.


Article
Extended Data Fig. 8 | Validation of anti-CTR antibody and study summary. a, Data summary: under physiological conditions in SR (left), human atrial cardiomyocytes produce and excrete endogenous CT, which binds to the CTR of atrial cardiofibroblasts (ACFs). Increased Gs-mediated cAMP inhibits multiple steps of fibrogenesis including, but not limited to, BMP1 activity and collagen processing by ACFs, thus keeping atrial fibrosis in check. In persistent AF (right), atrial cardiomyocytes secrete less CT, and ACFs show abnormal intracellular CTR localization; the consequent reduced CT–CTR activation enables unchecked structural remodelling and fibrosis in the atria that promotes AF maintenance and inducibility. b, Immunostaining with anti-CTR antibody shows barely detectable signal for CTR (green) in human kidney embryonic cell line (HEK293) and adult human dermal fibroblasts (HDF), and prominent positive CTR staining in human medullary carcinoma (TT) cells; red, filamin A; blue (DAPI), nuclei. Negative control for secondary antibodies (with primary antibodies omitted) in human ACFs is shown. c, Detection of positive immunofluorescence staining (green) with anti-CTR antibody in control ACFs, but not in CTR-KD ACFs, using anti-CTR LNA antisense oligonucleotides. d–f, The same antibody was used to detect CTR in HEK293 cells stably overexpressing (confirmed by qRT–PCR) human CTR protein (+hCTR; d) by flow cytometry (e; control cells negative for CTR (left plot) were used to determine the position of the P2 gate and the CTR+ cells (right plot) were sorted based on this gate and an antibody for CTR bound to AF647) and by immunofluorescence (f; CTR+ cells are stained in green and nuclei with DAPI, in blue). Gating strategy shown (bottom 3 panels): cells were first gated by general size and granularity (left plot), and then doublets were excluded using a standard plot of forward scatter height versus area (middle plot), eliminating cells with a large area for any given signal height, and then plotted on a log scale for mean fluorescent intensity of AF647 (right plot, gate P2) for CTR+ cells. The P2 gate was set based on unstained cells and shows events from the sample with a mean fluorescent intensity higher than the control in the P2 gate. g, Validation of the antibody for human pro-CT in human atrial tissue by immunoblotting. Representative example of the blot performed on 4 individual donors assessed on one day; this antibody was also tested in another 4 individual donors on a different day with the same result; recombinant human pro-CT was used as a positive control. h, i, CT ELISA kit confirms detection of human recombinant (in black) or synthetic CT (in green) in concentration-dependent manner with no cross-reactivity with recombinant human αCGRP or recombinant human pro-CT (in magenta) at serial dilutions. j, Cellular pellets in proteomic experiments were processed in duplicates to validate reproducibility. Data in e are presented as medians with interquartile ranges, as analysed by two-sided unpaired t-test after log transformation. FSC-A, forward scatter area. Data in b–d are representative images of cells stained on the same day and reproduced three times on three separate days. Data were pooled from individual cultures assessed in duplicates (e) or from technical triplicates (h, i) and technical duplicates ( j) analysed on the same day. For gel source data, see Supplementary Fig. 1. Extended Data Table 1 | Clinical characteristics of the study participants ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; AVR, aortic valve replacement; CABG, coronary artery bypass surgery; COPD, chronic obstructive pulmonary disease; MI, myocardial infarction; MVR, mitral valve replacement. The one-sided Fisher’s exact test was use to compare gender, surgical procedures, smoking status and medical history between groups. The two-sided unpaired t-test was used to compare age. Percentage in parenthesis (%) indicates percentage within the same group (for example, SR or AF). Characteristics of the participants used in scRNA-seq SMART-seq2 experiment are shown in columns 10–12.


Article
Extended Data Table 2 | In vivo echocardiographic and haemodynamic parameters in mice A, transmitral flow atrial filling; a′, mitral annulus moving velocity during atrial filling; CO, cardiac output; E, transmitral flow early filling; e′, mitral annulus moving velocity during early filling; EF, ejection fraction + (LVVd – LVVs)/LVVd × 100; FS, fractional shortening = (LADs – LADd)/LADs × 100; FS, fractional shortening = (LVDd – LVDs)/LVDd × 100; HR, heart rate; LV, left ventricle; LADd, left atrial dimension at end cardiac diastole; LADs, left atrial dimension at end cardiac systole; LV, left ventricle; LVDd, LV dimension at end cardiac diastole; LVAWd, LV anterior wall thickness at end cardiac diastole; LVIDd, LV internal diameter at diastole; LVIDd, LV internal; dLVAWs, LV anterior wall thickness at end of cardiac systole; diameter at systole; LVPWd, LV posterior wall thickness at end cardiac diastole; LVPWs, LV posterior wall thickness at end of cardiac systole; LVDd, LV diameter systole; LVDs, LV dimension at end cardiac systole; LVVd, LV volume at end cardiac diastole; LVVs, LV volume at end cardiac systole; SV, stroke volume. Data in a were analysed by two-sided unpaired t-test or Mann–Whitney U-test as appropriate; data in b were analysed by two-sided tests: one-way ANOVA with Holm-Sidak correction, except for LVAWd, which was analysed by Kruskal–Wallis test with Dunn’s correction. 1 nature research | reporting sum m ary April 2020 Corresponding author(s): Svetlana Reilly Last updated by author(s): 17-07-2020 Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Research policies, see our Editorial Policies and the Editorial Policy Checklist. Statistics For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section. n/a Confirmed The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one- or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section. A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable. For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated Our web collection on statistics for biologists contains articles on many of the points above. Software and code Policy information about availability of computer code Data collection Fibrosis in murine hearts was assessed in Masson Trichrome stained heart sections that were imaged at 20x, 10x and 4x magnification using Image-Pro 7.0. Electrocardiography in mice was carried out using Vivid 7 Dimension ultrasound system (GE Healthcare Ultrasound, Horten, Norway). Assessment of AF inducibility was performed by iox2 software (version 2.8.0.13, EMKA technologies, FR). Surface ECG & catheter signals were recorded and analysed using IOX2 software (version 2.8.0.13 and 2.9.5.28, EMKA technologies, FR). Cell sorting was carried out on a Becton Dickinson (BD) FACS Aria Fusion III sorter using a 100 μm nozzle and FACSDiva software v8. Cleaned peptides were separated on a nanoflow LC system (Thermo Scientific Dionex UltiMate 3000 RSLCnano); Fluorescence signal in BMP1 activity assay was measured by the fluorescence plate reader (Molecular Devices SpectraMax M2) at the excitation and emission wavelengths of 320 nm and 405 nm respectively. Scar-in-a-jar data were collected using operetta High Content Imaging System v4.9 (PerkinElmer). Imaging of primary cells and cell lines was performed with a Zeiss LSM 710 or Leica DM 6000 CFS confocal imaging system. To assess cellular localisation of the CTR, optical sections of fibroblasts were imaged with a frame size of 157 μm x 157 μm at a z-depth of 1 μm and pixel resolution of 0.09 μm x 0.09 μm. Channels were subsequently split and then merged in Fiji open source software (version 2.0.0-rc-69/1.52i). Data analysis bcl2fastq version 2.20.0.422 - Fastq file conversion from raw bcl format FastQC Version 0.11.8 Software (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) - Raw sequence data quality control software. STAR Version 2.4.2 - Transcriptome alignment. Cutadapt Version 1.16 - Sequence data trimming Picard Subread Version 1.6.2 - feature count summary Cellranger pipeline version 3.1.0 - 10x raw data processing. R Version 3.5.3 and 3.4.2 - Language for statistical analysis. limma 3.34.5 The following packages were used in addition to base installation: Seurat Bioconductor R Package, Version 3.1.5.9900 - single cell data QC, normalisation, clustering and differential gene expression analysis. 2 nature research | reporting sum m ary April 2020 irlba, R Package, Version 2.3.2 - Principal component analysis. clusterProfiler, Bioconductor R Package, Version 3.6.0 - Gene Ontology enrichment analysis. ggplot2, R Package, Version 2.2.1 - Plotting package. DESeq2, Bioconductor R Package, Version 1.20.0 - Differential expression analysis of bulk /pseudobulk RNA sequencing data. uwot R package Version 0.1.8 - UMAP visualisations MAST R package, Version 1.8.2 - Single Cell differential gene expression DropletUtils R package, Version 1.2.2 - data cleaning and cell identification sctransform R package, version 0.2.1 - data normalisation FLUOstar Omega software version 5.11, firmware version 1.43 (BMG LABTECH, Germany) Affymetrix Genechip Command Console software (1.4.1.46) SoftMax ProSoftware v5.0 ImageJ software (version 1.52a) Fiji open source software (version 2.0.0-rc-69/1.52i) DAVID 6.8 web tool GraphPad Prism (versions v7.05, v6.04, v8.02, or v8.04) Image-Pro v7.0 Harmony high-content analysis software v3.5 (Perkin Elmer, Revision 105159) QuantStudio 7 Flex Real-Time PCR System and ABI 7900HT Detection 480 System (both are from Applied Biosystems) Proteome Discoverer software (ThermoFisher Scientific, v2.3.0.523) hybrid human-bovine database (UniProtKB/Swiss-Prot version of January 2019) Mascot (Matrix Science, version 2.6.0) Cleaned peptides were separated on a nanoflow LC system (Thermo Scientific Dionex UltiMate 3000 RSLCnano) IOX2 software (version 2.8.0.13 and 2.9.5.28, EMKA technologies, FR) FACSDiva software v8 RTCA DA v1.0 Gpower v3.1.9.4. For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information. Data Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: - Accession codes, unique identifiers, or web links for publicly available datasets - A list of figures that have associated raw data - A description of any restrictions on data availability All data generated or analysed during this study are included in this published article. The scRNA-seq data are deposited on GEO at https://www.ncbi.nlm.nih.gov/ geo/query/acc.cgi?acc=GSE148507 (ref: GSE148506, GSE148507 and GSE148504). source data for all graphically expressed figures are provided with the paper. Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf Life sciences study design All studies must disclose on these points even when the disclosure is negative. Sample size We designed the in vivo studies to ensure that minimum number of mice are used to obtain biologically significant results. For the mouse studies in Houston cohort, the sample size was estimated based on our previous study (https://pubmed.ncbi.nlm.nih.gov/24398018/) and we found that n= 8-10/group would be sufficient to give at least 40% biologically meaningful difference for the AF onset (the main readout) between groups for measured parameters with the power set at 80% and the level of significance at 5%. For the Montreal cohort of mice, sample size was estimated based on the results of a pilot studies performed prior to the main experiment; we found that a minimum of 15 (for AF duration) and 4-6 (for fibrosis) animals would be needed to achieve 8)% power with significance at 5%. Data exclusions No data was excluded from analysis. Replication Except for the RNAseq, microarrays and proteomic high-throughput profiling, the in vitro findings were successfully replicated in at least two independent experiments, or reproduced with additional methods throughout the manuscript: e.g., protein content - with immunoblotting, ELISA, scar-in-a-jar, colourimetric assays and immunostaining in numerous biological replicates; gene expression - with semi-quantitative PCR and real time quantitative PCR; cellular assays : cell proliferation - with BrDU and impedance-based assays; collagen content - with HPA, scarin-a-jar and Sirius Red; in vitro studies were supported with complementary in vivo studies. Atrial fibrosis results in Montreal cohort were replicated in both sexes males and females; results on atrial fibrosis in Houston cohort were reproduced with different methods, i.e., sirius red and trichrome Masson's staining. More details are provided below. Fig.1. Data are pooled from individual donors assessed on the same day in one batch in single replicates (a, d-g), or (b-c, h-i) - in duplicates; all results are reproduced independently twice, or three times (a-c, j, k) in different subjects. Fig. 2. Data are pooled from individual donor cells assessed on the same day in single replicates except duplicates in (f, i-k). Data in (c) and (l) 3 nature research | reporting sum m ary April 2020 were pooled from individual donor cells treated on 3 different days assessed on the same day in single replicates. All results were reproduced twice in different donors on different days, or for (a) and (b) - by different method in (g) and (h) respectively. Fig. 3. Data are pooled from individual donor cells (a-f) treated on the same day and assessed in single replicates on the same day in one batch apart from (g) – assessed over 2-year period. The (e-f) were reproduced twice, except (g) - 3 times in different donors by different experimenters; data in (a) were partially reproduced by other methods (Fig.2a, 2e; Extended Data Fig.2a, 2e). Proteomic analysis was performed in single replicates following validation of reproducibility for (c) in duplicates (see Extended Data Fig.8j). Fig.4. Data are pooled from individual animals assessed in single replicates on the same day in one batch. Data in (a-c) and (l-m) are pooled from 6-9 sections per animal stained on different days and assessed in one batch. Results in (a-c) were reproduced in another batch of animals by different lab using different method, or by another experimenter using different method (l-m); results in (6i) were repeated on two different days in the same animals. Results in (d-h) and (i-j, l-r) were obtained from individual animals over 2.5-years and 4-month period respectively. Extended Data Fig.1. Data are pooled from individual donors assessed in single replicates (a, b, f, g-k, m, o-s) or duplicates (c-e, l, n); all results were reproduced independently twice. Extended Data Fig.2. Data are pooled from individual donor cells assessed in single replicates, except duplicates (a-b, g, i), on the same day in one batch. Results were reproduced twice (a-c, f-h) in different donors. Data in (l-o) are pooled from 6 individual donors in sinus rhythm assessed in 14742 cells (post QC after filtering the initial 18466 total cellular barcodes) on the same day in one batch. Extended Data Fig.3. The original data in (e) used for the GO analysis were pooled from 6 individual donors assessed in single replicates on the same day in one batch. Data are pooled from individual donors (l), or separate days (m-n) and assessed in single replicates on the same day in one batch apart from (n, single replicates on two different days), or in duplicates in (g, i, k) assessed on the same day. Findings in (a-d, j) were validated by another method (Fig.3a-b, g, Extended Data Fig.6b). All (except e) were reproduced twice in different donors. Extended Data Fig.4. Data are pooled from 268 single cells isolated from 8 individual donors; scRNA-seq workflow was performed on the same day in one batch. Extended Data Fig.5. The original data used for the analysis are pooled from 268 single cells isolated from 8 individual donors; the scRNA-seq workflow is carried out on the same day in one batch. Extended Data Fig.6. Data in (a) are pooled from individual donors assessed in single replicate on the same day; results were reproduced in the same donors twice. Data in (b) are pooled from individual donors (a few cells in each field as shown in the figure) collected over two-year period, assessed on separate days and validated by 3 independent experimenters. Extended Data Fig.7. Data are pooled from individual animals assessed in single replicates on the same day and reproduced in two centres in (a, c-d). Results in (m-n) were obtained from individual animals over ~2.5 years. Extended Data Fig.8. Data in (a-c) are representative images of cells stained on the same day and reproduced three times on three separate days. Data are pooled from individual cultures assessed in duplicates (d), or from technical triplicates (g-h) and technical duplicates (i) analysed on the same day. Blots in (f) were reproduced twice in different donors. Randomization The Montreal mouse studies were not formally randomized, since the groups were determined by breeding rather than by intervention. The mice were studied consecutively as soon as they reached maturity upon breeding. All experimental studies (e.g., in vivo electrophysiology studies, qPCR, echocardiography and assessment of atrial fibrosis) were blinded to mouse genotype. The Houston mouse studies were performed in mice of the same genotype - LKB1 FL/FL. Pups from same litter and same cage were injected as early as day-5 with the same virus and the males/females were separated at the time of weaning. Subsequently, litters were randomly allocated to each group, keeping overall numbers of mice for all the groups similar. Blinding For in vivo studies, genotype was not disclosed to the investigators performing data collection and those generating quantitative readouts. Experimenters were blinded to the group allocation during data collection and analysis (Montreal and Houston cohorts). Reporting for specific materials, systems and methods We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response. Materials & experimental systems n/a Involved in the study Antibodies Eukaryotic cell lines Palaeontology and archaeology Animals and other organisms Human research participants Clinical data Dual use research of concern Methods n/a Involved in the study ChIP-seq Flow cytometry MRI-based neuroimaging Antibodies Antibodies used Calcitonin Receptor AHP635, Bio-Rad, UK, 1/500 BMP1 ab38953, Abcam, UK, 1/1000 COL1A2 (C-19) sc-8786, Santa Cruz Biotechnologies, USA, 1/1000 α-Smooth Muscle Actin A5228, Sigma, USA, 1/1000 p44/42 MAPK (Erk1/2) 9102S, Cell Signaling, USA, 1/1000 Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) 9101S, Cell Signaling, USA, 1/1000 Filamin A MAB1680-C, Merck, USA, 1/1000 Pro-calcitonin 130-00025-500, RayBiotech, USA, 1/1000 GAPDH-HRP-conjugated G9295, Sigma, USA, 1/1000 4 nature research | reporting sum m ary April 2020 GAPDH MAB374, Millipore , USA, 1/1000 Calcitonin ab8553, Abcam, UK, 1/1000 αCGRP ab189786, Abcam, UK, 1/1000 CREB (E306) ab32515, Abcam, UK, 1/1000 PKA R2 (phospho S96) ab226754, Abcam, UK, 1/1000 PKA C alpha AF4268-SP, R&D Systems, UK, 1/1000 Cyclic-AMP MAB2146-SP, R&D Systems, UK, 1/1000 EPAC1 (5D3) 4155S, Cell Signaling, USA, 1/1000 EPAC2 (D3P3J) 43239S, Cell Signaling, USA, 1/1000 Collagen-1 C2456, Sigma, USA, 1/100 Fibronectin F3648, Sigma, USA, 1/1000 TotalSeq™-A0251 anti-human Hashtag 1 394601 BioLegend 1 μg/ml TotalSeq™-A0252 anti-human Hashtag 2 394603 BioLegend 1 μg/ml TotalSeq™-A0253 anti-human Hashtag 3 394605 BioLegend 1 μg/ml TotalSeq™-A0254 anti-human Hashtag 4 394607 BioLegend 1 μg/ml TotalSeq™-A0257 anti-human Hashtag 7 394613 BioLegend 1 μg/ml TotalSeq™-A0258 anti-human Hashtag 8 394615 BioLegend 1 μg/ml TotalSeq™-A0259 anti-human Hashtag 9 394617 BioLegend 1 μg/ml TotalSeq™-A0263 anti-human Hashtag 13 394625 BioLegend 1 μg/ml TotalSeq™-A0264 anti-human Hashtag 14 394627 BioLegend 1 μg/ml TotalSeq™-A0260 anti-human Hashtag 10 394619 BioLegend 1 μg/ml TotalSeq™-A0255 anti-human Hashtag 5 394609 BioLegend 1 μg/ml TotalSeq™-A0256 anti-human Hashtag 6 394611 BioLegend 1 μg/ml. Validation Calcitonin Receptor AHP635 https://www.bio-rad-antibodies.com/polyclonal/rat-calcitonin-receptor-antibody-ahp635.html? f=purified BMP1 ab38953 https://www.abcam.com/bmp1pcp-antibody-ab38953.pdf COL1A2 (C-19) sc-87 86 https://www.scbt.com/p/col1a2-antibody-c-19 α-Smooth Muscle Actin A5228 https://www.sigmaaldrich.com/catalog/product/sigma/a5228 p44/42 MAPK (Erk1/2) 9102S https://www.cellsignal.com/products/primary-antibodies/p44-42-mapk-erk1-2-antibody/9102 Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) 9101S https://www.cellsignal.com/products/primary-antibodies/phospho-p44-42mapk-erk1-2-thr202-tyr204-antibody/9101 Filamin A MAB1680-C https://www.sigmaaldrich.com/catalog/product/mm/mab1680c?lang=en®ion=GB Pro-calcitonin 130-00025-500 https://www.raybiotech.com/rabbit-anti-human-calcitonin-en-2 GAPDH-HRP-conjugated G9295 https://www.sigmaaldrich.com/catalog/product/sigma/g9295 GAPDH MAB374 https://www.merckmillipore.com/GB/en/product/Anti-Glyceraldehyde-3-Phosphate-Dehydrogenase-Antibodyclone-6C5,MM_NF-MAB374 Calcitonin ab8553 https://www.abcam.com/calcitonin-antibody-ab8553.html αCGRP ab189786 https://www.abcam.com/cgrp-antibody-ab189786.html CREB (E306) ab32515 https://www.abcam.com/creb-antibody-e306-ab32515.html PKA R2 (phospho S96) ab226754 https://www.abcam.com/pka-r2pkr2-phospho-s96-antibody-ab226754.html PKA C alpha AF4268-SP https://www.rndsystems.com/products/human-mouse-rat-pka-calpha-antibody_af4268 Cyclic-AMP MAB2146-SP https://www.rndsystems.com/products/camp-antibody-250532_mab2146 EPAC1 (5D3) 4155S https://www.cellsignal.com/products/primary-antibodies/epac1-5d3-mouse-mab/4155 EPAC2 (D3P3J) 43239S https://www.cellsignal.com/products/primary-antibodies/epac2-d3p3j-rabbit-mab/43239 Collagen 1 C2456 https://www.sigmaaldrich.com/catalog/product/sigma/c2456 Fibronectin F3648 https://www.sigmaaldrich.com/catalog/product/sigma/f3648 TotalSeq™-A0251 anti-human Hashtag 1 394601 https://www.biolegend.com/en-us/products/totalseq-a0251-anti-humanhashtag-1-16080 TotalSeq™-A0252 anti-human Hashtag 2 394603 https://www.biolegend.com/en-us/products/totalseq-a0252-anti-human-hashtag-2antibody-16081 TotalSeq™-A0253 anti-human Hashtag 3 394605 https://www.biolegend.com/en-us/products/totalseq-a0253-anti-human-hashtag-3antibody-16084 TotalSeq™-A0254 anti-human Hashtag 4 394607 https://www.biolegend.com/en-us/products/totalseq-a0254-anti-human-hashtag-4antibody-16086 TotalSeq™-A0257 anti-human Hashtag 7 394613 https://www.biolegend.com/en-us/products/totalseq-a0257-anti-human-hashtag-7antibody-16090 TotalSeq™-A0258 anti-human Hashtag 8 394615 https://www.biolegend.com/en-us/products/totalseq-a0258-anti-human-hashtag-8antibody-16092 TotalSeq™-A0259 anti-human Hashtag 9 394617 https://www.biolegend.com/en-us/products/totalseq-a0259-anti-human-hashtag-9antibody-16093 TotalSeq™-A0263 anti-human Hashtag 13 394625 https://www.biolegend.com/en-us/products/totalseq-a0263-anti-humanhashtag-13-antibody-16096 TotalSeq™-A0264 anti-human Hashtag 14 394627 https://www.biolegend.com/en-us/products/totalseq-a0264-anti-humanhashtag-14-antibody-16097 TotalSeq™-A0260 anti-human Hashtag 10 394619 https://www.biolegend.com/en-us/products/totalseq-a0260-anti-humanhashtag-10-antibody-16094 TotalSeq™-A0255 anti-human Hashtag 5 394609 https://www.biolegend.com/en-us/products/totalseq-a0255-anti-human-hashtag-5antibody-16088 TotalSeq™-A0256 anti-human Hashtag 6 394611 https://www.biolegend.com/en-us/products/totalseq-a0256-anti-human-hashtag-6antibody-16089. In addition we have also performed in-house validation of the antibodies against pro-CT and CTR (as shown in the manuscript in Extended Data Fig.8b-d, g) using appropriate negative and positive controls, or HEK293cells stably overexpressing full length nontagged human CTR (Extended Data Fig.8e-f), or using Locked Nucleic Acid-antisence oligonucleotides to knockdown the CTRs (Extended Data Fig.8c, Extended Data Fig.1l-m). 5 nature research | reporting sum m ary April 2020 Eukaryotic cell lines Policy information about cell lines Cell line source(s) All functional experiments were carried out in primary cells. The following cell lines were used as positive or negative controls during internal antibody validation steps with immunoblotting or real time qPCR: TT cell line ((#CRL-1803, ATCC) obtained from a 77 year old female patient suffering with medullary thyroid carcinoma; HEK293 cells (#CRL-1573, ATCC), human adult dermal fibroblasts (HDFa; #C-0135C, Gibco, Invitrogen cell culture, USA) and HEK293 stably expressing full length non-tagged human CTR (#P30136, Innoprot, USA). Authentication Authentication of cell lines was carried by the manufacturer as outlined below: ATCC CRL-1803 - https://www.lgcstandards-atcc.org/products/all/CRL-1803.aspx?geo_country=gb#specifications; ATCC CRL-1573 - https://www.lgcstandards-atcc.org/products/all/CRL-1573.aspx#specifications; Gibco C-0135C - https://www.thermofisher.com/order/catalog/product/C0135C#/C0135C; Innoprot P30136 - https://innoprot.com/product/hitseeker-calcitonin-receptor-cell-line/ In-house cell authentication was performed by monitoring cell morphology and phenotypical tests (e.g., immunostaining and flow cytometry - for CTR protein and ELISA - for CT secretion). Cell lines were certified by the manufacturer. Mycoplasma contamination Tested negative using immunofluorescence staining with DAPI (as shown in the manuscript, Fig. 2a, Fig.3g, Extended Data Fig.6b, 8b-d). Commonly misidentified lines (See ICLAC register) None. Animals and other organisms Policy information about studies involving animals; ARRIVE guidelines recommended for reporting animal research Laboratory animals Animal studies were carried out in mice; age: 10-12 weeks old; gender: males and females; genetic background: CMV CRE https:// www.jax.org/strain/006054, LKB1 FL/FL https://www.jax.org/strain/014143, CTR flox , B6.129-Calcrtm1Rda/Apb (generated by Prof RA Davey, University of Melbourne). Housing conditions: The CTR-KO and control mice (Montreal cohort) were housed in Allentown XJ cages at 20-22°C, 50% humidity and 60 air changes/hour ventilation conditions. Diet consisted of the sterilized food (#2019S, Envigo) and osmotic water. The CTR-KO and control mice (Melbourne cohort) were housed in a specified pathogen-free facility at 22°C, in a 12-hour light/dark cycle and were supplied with standard irradiated mouse chow (1.2% calcium and 0.96% phosphorus; Ridley Agriproducts, Western Australia) and water ad libitum. Breeding mice were housed in micro-isolator cages and offspring used for experiments were transferred to open-top cages at weaning (3–5 mice/cage). Cages contained corn-cob bedding, and cardboard tubes and tissues were supplied for environmental enrichment. Studies in LKB1-KD, LKB1/CT-dKD, LKB1-KD+CT and controls (Houston cohort) were housed in standard mice cages provided with the bags of sizzle net as cage enrichment and were fed standard feeder chow as approved by the IACUC and recommended by ‘the Guide’ (NIH Publication #85-23, revised 1996). Wild animals We did not use wild animals. Field-collected samples We did not use field-collected samples. Ethics oversight All animal work was performed in accordance with the local Montreal Heart Institute and Austin Health Animal Care and Ethics Committee Ethics oversight guidance and in accordance with NIH guidelines. Studies in LKB1-KD, LKB1/CT-dKD, LKB1-KD+CT and controls (Houston cohort) were performed according to protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the Baylor College of Medicine. Note that full information on the approval of the study protocol must also be provided in the manuscript. Human research participants Policy information about studies involving human research participants Population characteristics Patients (n=156 in total) undergoing coronary artery bypass grafting (CABG) and/or valve surgery in the John Radcliffe hospital, Oxford (UK). Detailed patients' characteristics are summarized in the Extended Data Table 1 of the manuscript. All patients gave informed written consent. Right atrial biopsies were collected before cardiopulmonary bypass, rinsed of any blood contamination and immediately processed for cell isolation or snap-frozen until use. Recruitment Patients were recruited according to the approved study protocol, i.e., via face-to-face communication with a trained researcher; every recruited participant signed a consent form which is safely stored securely. All patients admitted to the John Radcliffe hospital for cardiac surgery were eligible for this study, unless they: (1) underwent previous cardiac surgery, (2) younger 18 or older 85 years of age, and (3) are not able/not willing to consent. Clinical characteristics of the participants were derived from medical notes. There were no self-selection bias or other biases during patients recruitment. Ethics oversight Studies involving human participants were approved by the local Research Ethics Committee (South Central - Berkshire B Research Ethics Committee, UK; ref: 18/SC/0404 and 07/Q1607/38). Note that full information on the approval of the study protocol must also be provided in the manuscript. 6 nature research | reporting sum m ary April 2020 Flow Cytometry Plots Confirm that: The axis labels state the marker and fluorochrome used (e.g. CD4-FITC). The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers). All plots are contour plots with outliers or pseudocolor plots. A numerical value for number of cells or percentage (with statistics) is provided. Methodology Sample preparation For flow cytometry was used to validate anti-CTR antibody used in the study. We used HEK293 cells stably expressing full length non-tagged human CTR (#P30136, Innoprot, USA). The cells were sub-cultured according to the manufacturer instruction, gently trypsinised and stained with the anti-CTR antibody or IgG control using standard staining protocol. Cells were kept on ice through the preparation process and until flow cytometry. Instrument Becton Dickinson (BD) FACS Aria Fusion III sorter. Software BD FACSDiva v8.0 Cell population abundance Highly abundant HEK293 cells positive for human CTRs. Gating strategy Cells were gated for viable (based on DAPI_negative signal) singlets (standard gating strategy) as shown in the Extended Data Fig. 8ef Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.


Metadata
Authors
Lucia M Moreira, Abhijit Takawale, Mohit Hulsurkar, David A Menassa, Agne Antanaviciute, Satadru K Lahiri, Neelam Mehta, Neil Evans, Constantinos Psarros, Paul Robinson, Alexander J Sparrow, Marc-Antoine Gillis, Neil Ashley, Patrice Naud, Javier Barallobre-Barreiro, Konstantinos Theofilatos, Angela Lee, Mary Norris, Michele V Clarke, Patricia K Russell, Barbara Casadei, Shoumo Bhattacharya, Jeffrey D Zajac, Rachel A Davey, Martin Sirois, Adam Mead, Alison Simmons, Manuel Mayr, Rana Sayeed, George Krasopoulos, Charles Redwood, Keith M Channon, Jean-Claude Tardif, Xander H T Wehrens, Stanley Nattel, Svetlana Reilly
Journal
Nature
Publisher
Date
pm33149301
PM Id
33149301
PMC Id
Images