Defective Desmosomal Adhesion Causes Arrhythmogenic Cardiomyopathy by Involving an Integrin-αVβ6/TGF-β Signaling Cascade

Background: Arrhythmogenic cardiomyopathy (ACM) is characterized by progressive loss of cardiomyocytes with fibrofatty tissue replacement, systolic dysfunction, and life-threatening arrhythmias. A substantial proportion of ACM is caused by mutations in genes of the desmosomal cell–cell adhesion complex, but the underlying mechanisms are not well understood. In the current study, we investigated the relevance of defective desmosomal adhesion for ACM development and progression. Methods: We mutated the binding site of DSG2 (desmoglein-2), a crucial desmosomal adhesion molecule in cardiomyocytes. This DSG2-W2A mutation abrogates the tryptophan swap, a central interaction mechanism of DSG2 on the basis of structural data. Impaired adhesive function of DSG2-W2A was confirmed by cell–cell dissociation assays and force spectroscopy measurements by atomic force microscopy. The DSG2-W2A knock-in mouse model was analyzed by echocardiography, ECG, and histologic and biomolecular techniques including RNA sequencing and transmission electron and superresolution microscopy. The results were compared with ACM patient samples, and their relevance was confirmed in vivo and in cardiac slice cultures by inhibitor studies applying the small molecule EMD527040 or an inhibitory integrin-αVβ6 antibody. Results: The DSG2-W2A mutation impaired binding on molecular level and compromised intercellular adhesive function. Mice bearing this mutation develop a severe cardiac phenotype recalling the characteristics of ACM, including cardiac fibrosis, impaired systolic function, and arrhythmia. A comparison of the transcriptome of mutant mice with ACM patient data suggested deregulated integrin-αVβ6 and subsequent transforming growth factor–β signaling as driver of cardiac fibrosis. Blocking integrin-αVβ6 led to reduced expression of profibrotic markers and reduced fibrosis formation in mutant animals in vivo. Conclusions: We show that disruption of desmosomal adhesion is sufficient to induce a phenotype that fulfils the clinical criteria to establish the diagnosis of ACM, confirming the dysfunctional adhesion hypothesis. Deregulation of integrin-αVβ6 and transforming growth factor–β signaling was identified as a central step toward fibrosis. A pilot in vivo drug test revealed this pathway as a promising target to ameliorate fibrosis. This highlights the value of this model to discern mechanisms of cardiac fibrosis and to identify and test novel treatment options for ACM.


DSG2-W2A mouse model
All mouse experiments were carried out according to the protocol approved by the Cantonal Veterinary Office of Basel-Stadt (License number 2973_32878 and 3070_32419). All mice were housed under specific pathogen-free conditions with standard chow and bedding with 12 hours day/night cycle according to institutional guidelines. Animals of both sexes were applied without bias. For inhibitor treatments, all mice were age-and sex-matched and randomly allocated to treatment or control group.
The Dsg2-W>A allele was obtained by Cas9/CRISPR embryo electroporation. The Cas9/CRISPR target sequence tggttcgtcaaaagagggcc(tgg) (PAM sequence in brackets is also the TGG-Trp codon) spanning the mutation site was selected with the help of CRISPOR software (http://crispor.tefor.net/) 42 . ssDNA oligonucleotide 5'gtgataactcaaggtaattgtattaacaggtcttcagccc aagaaatgaaggcaaaccgttccctaagcacactcacttggttcgtcaaaagagggcagctatcactgcccctgtgg ctctgcgggagggcgaagacctgtccagaaagaacccgattgccaaggtagcagctacagaagaatgtggcgag ggtgttggc3' (GCT -Ala codon in bold underlined) was designed to insert the W>A mutation into the Cas9-generated DSB by homologous recombination and at the same time mutate the TGG PAM sequence to GCT. C57BL/6J female mice underwent ovulation induction by i.p. injection of 5 IU equine chorionic gonadotrophin (PMSG; Folligon, InterVet, Vienna, Austria), followed by i.p. injection of 5 IU human chorionic gonadotropin (Pregnyl, Essex Chemie, Lucerne, Switzerland) 48 hours later. For the recovery of embryos, C57BL/6J females were mated with males of the same strain immediately after the administration of human chorionic gonadotropin. Embryos were collected from oviducts 24 hours after the human chorionic gonadotropin injection, and were then freed from any remaining cumulus cells by a 1-2 min treatment of 0.1 % hyaluronidase (Sigma-Aldrich, St. Louis, MO, ISA) dissolved in M2 medium (Sigma-Aldrich). Prior to electroporation, the zona pellucida was partially removed by brief treatment with acid Tyrode's solution and the embryos were washed and briefly cultured in M16 medium (Sigma-Aldrich) at 37 °C and 5 % CO2. Electroporation with a mixture of ssDNA oligonucleotide targeting template, 16 µmol/l cr:trcrRNA hybrid targeting Dsg2 and 16 µmol/l Cas9 protein (all reagents from IDT, Coralville, IA, USA) was carried out using 1 mm gap electroporation cuvette and the ECM830 electroporator (BTX Harvard Apparatus, Holliston, MA, USA). Two square 3 ms pulses of 30 V with 100 ms interval were applied. Surviving embryos were washed with M16 medium and transferred immediately into the oviducts of 8-16-weeks-old pseudopregnant Crl:CD1(ICR) females that had been mated with sterile genetically vasectomized males the day before embryo transfer (0.5 dpc). Pregnant females were allowed to deliver and raise their pups until weaning age. In total 150 embryos were electroporated and 147 surviving embryos were transferred into 7 foster mothers. All foster mothers produced live litters with a total of 20 viable F0 pups. One F0 pup carried the desired mutation as confirmed by sequencing. This founder animal was bred to C57BL/6J partner. The mut/wt offspring from this mating was bred to C57BL/6J partners for 2 generations to establish the Dsg2-W2A mouse line.
For genotyping of the DSG2-W2A line, DNA was extracted from biopsies in 25 mmol/l NaOH and 0.2 mmol/l EDTA at 98 °C for 1 hour and neutralized with 40 mmol/l Tris pH 5.5. PCR was performed using GoTaq G2 (M7845, Promega, Madison, WI, USA) according to manufacturer's instructions with the primers Dsg2-W2A for: electrophoresis in a 3 % (w/v) agarose gel containing Midori Green Advanced (Nippon Genetics, Düren, Germany) for fluorescence DNA visualization, presence of a Dsg2-WT allele was detectable as 197 bp fragment, while Dsg2-W2A mutant allele was cut into a 109 bp and 72 bp fragment ( Figure S1C).

Murine sample collection
For heart dissection, mice were euthanized via i.p. pentobarbital overdose according to guidelines of the Cantonal Veterinary Office of Basel-Stadt and the University of Basel. Hearts were removed by lateral thoracotomy and directly immersed in ice-cold HBSS supplemented with 20 mmol/l 2,3-Butanedione monoxime (BDM, Sigma-Aldrich) unless stated otherwise. Morphology of the hearts was analyzed using a binocular stereo microscope (SZX2, Olympus, Shinjuku, Japan) and documented with a SLR camera (EOS 800D, Canon, Tokyo, Japan). Tissue was processed as described in the respective section.
For dissection of embryos, timed matings were performed and pregnant mice euthanized via i.p. pentobarbital overdose after the respective days. Embryos were dissected from the uterus and placed in HBSS. After image acquisition as described above, a tissue sample for genotyping was collected from the tail and embryos were processed as described in the Histological staining section.

In vivo experiments
At begin of the experiments mut/wt mice from the same litter were sex-matched and randomly allocated by stratified randomization to control or treatment group. Within the limits of a pilot experiment, number of animals were estimated with a power calculation.

Plasmid generation and cloning
For lentiviral overexpression plasmids, DNA for full length Dsg2-WT and full length Dsg2-W2A mutation, respectively, were amplified from cDNA originating from liver tissue of DSG2-WT and DSG2-W2A mice using AscI-msDsg2-N forward and NotI-GT- Darmstadt, Germany) and 2 mmol/l L-glutamine (Sigma-Aldrich) at 37 °C, 5 % CO2 and full humidity. For experiments, cells were seeded on TC-treated plastic cell culture plates, grown to confluency and differentiated for seven days. All cells were quarterly checked for mycoplasma contaminations using PCR and were proven negative.
CaCo2 cells were routinely authenticated by Short Tandem Repeat profiling.

Lentivirus generation and transduction
Lentiviral particles were generated according to standard procedures. In brief, HEK293T cells were co-transfected with the packaging vector psPAX2 (#12259, Addgene, Watertown, MA, USA), the envelope vector pMD2.G (#12260, Addgene) and the respective construct plasmid using TurboFect (Thermo Fisher Scientific, Waltham, MA, USA). After 48 hours, virus particle containing supernatant was collected and enriched using LentiConcentrator (OriGene). Cells were transduced with the respective concentrated virus particles using 10 µg/mL polybrene (Sigma-Aldrich) according to the manufacturer's instructions. After 24 hours, medium was changed and cells cultivated for one week before starting with the respective experiments.
Expression of the respective construct was confirmed via Western blot analysis.

Generation and cultivation of murine keratinocytes
For isolation of murine keratinocytes from DSG2-W2A mut/mut and wt/wt mice, the epidermis of neonatal mice was separated from the dermis via incubation in dispase II Aidenbach, Germany) supplemented with 10 % calcium-free foetal bovine serum (S0615, Merck), 2 mmol/l stable glutamine (BioConcept, Allschwil, Switzerland), 50μg/ml penicillin/streptomycin (Applichem), 10 ng/ml murine epidermal GF (Invitrogen, Carlsbad, CA), 1 mmol/l sodium pyruvate, 0.18 mmol/l adenine, 120 pmol/l cholera toxin, 5 µg/ml insulin, and 500 ng/ml hydrocortisone (all Sigma-Aldrich). Cells were kept in an incubator at 35 °C with 5 % CO2 and 100 % humidity, the medium was changed every third day. When reaching confluency, cells were transferred into a new coated culture dish. After around six passages, cells were immortalized and could be expanded and seeded for experiments. 48 hours before experiments were conducted, 1.8 mmol/l calcium was added to the medium to induce cell differentiation. Cells were quarterly checked for mycoplasma infections using PCR and were proven negative.

Dissociation assay
Cells were treated as indicated and grown to confluency in 24-well plates. Cell monolayers were washed with HBSS and incubated with dissociation buffer (dispase II 2.5 U/mL, Sigma-Aldrich, D4693 in HBSS) at 37 °C till detachment of the cell monolayer from well bottom. After detachment, monolayers were mechanically stressed by defined pipetting using an electrical pipette (Eppendorf, Hamburg, Germany). The total number of resulting fragments per well was determined using a binocular stereo microscope (SZX2, Olympus). Fragments were counted if they were clearly visible at 1.25-fold magnification. The number of fragments is an indirect measure for intercellular cohesion. Images were acquired with a SLR camera (EOS 800D, Canon).
Median band density was quantified applying ImageStudio (both Li-Cor) according to manufacturer's instructions and normalized to the respective loading control.
The  . For analysis of staining intensity following masks and regions were defined: (i) "cardiomyocytes" -outline of the total cardiomyocyte area by thresholding of the f-acting signal, (ii) "nucleus" -detection of nuclei via DAPI staining applying the QuPath cell detection tool, (iii) "ICD" -outlines of ICD area were created with threshold of a respective counterstain or QuPath brush tool, (iv) "cytosol" -area of ICD was subtracted from cardiomyocyte area. Respective masks were applied to the channel of interest to measure mean nuclear intensity in selected areas. The corresponding mean signal background was subtracted for ICD intensity.

Histological staining
Tissue was embedded and cut as described in the section Immunostaining.
Haematoxylin/ Eosin (HE) staining was performed according to standard procedures. In

Echocardiography and Electrocardiogram (ECG)
Transthoracic echocardiography was performed using the Vevo 2100   ECG data were recorded and analysed using the LabChart Pro 8 software (ADInstruments) equipped with the ECG Analysis Module. Peak amplitudes and intervals were determined from a curve averaged from 50 subsequent QRS complexes as mean of three time points. Definition of peaks is shown in Figure 2I.
After final measurements, mice were euthanized via cervical dislocation under anethesia and hearts were dissected. Organs were embedded and stained as described in the section Histological staining for further analysis to determine the amount of fibrosis. For wet/dry ratio, weight of the isolated lungs was determined before and after drying the tissue for 24 hours at 37 °C.

Transmission Electron Microscopy (TEM)
Ventricular cardiac tissue was dissected, cut and fixed in 2 % paraformaldehyde and Gapdh and Tubg2 of the respective sample was used.

Mouse RNA-sequencing (RNA-Seq)
For transcriptomic analysis before and after onset of fibrosis, hearts of 5-days-and 9weeks-old mice were dissected. Wt/wt and mut/mut mice were matched for age and sex. For 5-days-old mice, atria were removed and both ventricles lysed. For 9-weeksold animals, similar sized tissue samples were taken from the right and left ventricle via a 3 mm diameter biopsy punch (Viollier, Allschwil, Switzerland). RNA was isolated as described in the section RNA isolation. RNA samples were quality-checked on the

Mouse RNA-Seq data analysis
Reads were aligned to the mouse mm10 genome using the aligner STAR (version 2.7.3a) 45 with extra options "--outFilterMultimapNmax 10 --outSAMmultNmax 1" for handling multimapping reads. Aligned reads were assigned to ensembl genes (

Re-analysis of available ACM data sets and comparison with murine data sets
Raw data of GEO data sets GSE107157 and GSE107480 were downloaded from the European Nucleotide Archive and mapped to the human hg38AnalysisSet genome using STAR. Gene expression was quantified with featureCounts and relied on the human ensembl gene annotation (version 96). Gene counts of both data sets were imported into R and analysed using the edgeR workflow. Specifically, genes were filtered by expression which retained 25293 genes, followed by a differential expression analysis contrasting the 4 sample groups which arise from tissue source (left and right ventricle) and disease (ARVC and healthy). In order to compare mouse specific gene lists to the human data, mouse genes were mapped to human orthologs using BioMart. The gene set of human orthologs was tested for differential enrichment in the mouse gene expression contrasts defined above using the function cameraPR.
For analysis of the GEO data set GSE29819, the GEO2R tool was used (ACM samples versus normal). Results with adjusted p-values<0.05 were considered as significantly changed. The Venn diagram for overlap between two groups was determined with the

Structured illumination microscopy (SIM)
For structured illumination microscopy (SIM) of cryopreserved mouse heart tissue, sectioning and immunostaining were performed as described. Image stacks spanning full ICDs were acquired using a DeltaVision OMX-Blaze (Version 4; Applied Precision) equipped with a 60x PL APO NA = 1.42 objective (Olympus) yielding a voxel size of 0.04 x 0.04 x 0.125 µm.
For analysis, the observer was generally blinded for the genotype. For evaluation of signal size and frequency within ICDs, the image stacks (typically containing one clearly defined ICD) were loaded into Imaris 9.6 (Bitplane AG, Schlieren, Switzerland).
The ICD was annotated by manually outlining the DSP signal in every third image plane. The outlines were converted into a volume using the Manual Surface module.  Presented data correspond to echocardiography and ECG measurements in Figure 2 and include the same animals.      Lower row shows an overview image of a fibrotic area in mut/mut hearts. Dotted orange line highlights border of fibrotic tissue. Scale bar: 50 µm. Images representative for 5 mice per genotype.