Circulation Journal
Online ISSN : 1347-4820
Print ISSN : 1346-9843
ISSN-L : 1346-9843
Arrhythmia/Electrophysiology
Functional Role of the L396R Mutation of Tks5 Identified by an Exome-Wide Association Study in Atrial Fibrillation
Xiaoxi YangTetsuo SasanoYusuke EbanaJun K. TakeuchiKensuke IharaMasahiro YamazoeTetsushi Furukawa
Author information
JOURNAL OPEN ACCESS FULL-TEXT HTML
Supplementary material

2020 Volume 84 Issue 12 Pages 2148-2157

Details
Abstract

Background: Atrial fibrillation (AF) is the most common cardiac arrhythmia; however, the current treatment strategies for AF have limited efficacy. Thus, a better understanding of the mechanisms underlying AF is important for future therapeutic strategy. A previous study (Exome-Wide Association Study (ExWAS)) identified a rare variant, rs202011870 (MAF=0.00036, GenomAD), which is highly associated with AF (OR=3.617, P<0.0001). rs202011870 results in the replacement of Leu at 396 with Arg (L396R) in a molecule, Tks5; however, the mechanism of how rs202011870 links to AF is completely unknown.

Methods and Results: The association of rs202011870 with AF was examined in 3,378 participants (641 control and 2,737 AF cases) from 4 independent cohorts by using an Invader assay. Consequences of rs202011870 in migration ability, podosome formation, and expression of inflammation-related molecules in macrophages were examined using RAW264.7 cells with a trans-well assay, immunocytochemistry, and qPCR assay. Validation of the association of rs202011870 with AF was successful. In vitro studies showed that RAW264.7 cells with L396R-Tks5 increased trans-well migration ability, and enhanced podosome formation. RAW264.7 cells with L396R-Tks5 also increased the expression of several inflammatory cytokines and inflammation-related molecules.

Conclusions: L396R mutation in Tks5 associated with AF enhances migration of macrophages and their inflammatory features, resulting in enhanced susceptibility to AF.

Atrial fibrillation (AF) is the most common cardiac arrhythmia, which is currently estimated to affect over 33 million individuals worldwide.1,2 AF has a strong effect on both mortality and morbidity; however, the current treatment strategies for AF have limited efficacy. Thus, the better understanding of the mechanisms underlying AF is important for future therapeutic strategy.

The Genome-Wide Association Study (GWAS) was introduced to identify single nucleotide polymorphisms (SNPs) associated with pathogenesis of common diseases. In addition to its main purpose of risk stratification of diseases, GWAS is also expected to provide insights into unknown pathogenesis pathways and novel drug development targets. Although nearly 20 years have passed since the first introduction of GWAS in 2002,3 new drugs have not yet been developed based on GWAS findings. It is, at least in part, due to the fact that most SNPs identified in the GWAS reside in regions outside of genes and that the relation with diseases is not strong, with the odds ratio (OR) <1.5 for most SNPs. In contrast, the Exome-Wide Association Study (ExWAS) identifies SNPs, which are localized in exons and potentially have strong influence on disease development. The ExWAS, therefore, could have a better chance to unravel novel pathogenesis pathways and drug-development targets for many common diseases.

Our group had previously identified a rare variant, rs202011870 (0.00036, GenomAD), which is highly associated with AF (OR=3.617, P<0.0001). This variant is localized in the protein coding region of the TKS5 gene. The association of rs202011870 with AF was also identified in a recently reported large scale meta-analysis of AF-associated common and rare variants.4 However, the replication in an independent dataset is still necessary to validate its association with AF. More importantly, it is totally unknown whether this mutation in Tks5 causes any biological effects related to high AF prevalence. The purposes of the present study were, therefore, two-fold: first, to validate the association of rs202011870 with AF in samples obtained from 4 health-care cohorts in Japan; and second, to investigate the biological evidence underlining this epidemical association.

Methods

Human Whole Blood Samples, Genome DNA Extraction, and Invader Assay

Overall, 3,378 participants (641 controls and 2,737 AF cases) were recruited from 4 independent cohorts in Japan: Saitama Red-Cross Hospital, National Disaster Medical Center, Tsuchiura Kyodo Hospital, and Kashiwa City Hospital. The study protocol was approved by the ethical committee of each hospital and followed the Declaration of Helsinki and the ethical standards of the responsible committee on human experimentation, and written informed consent was obtained from all participants. DNAs were extracted from whole blood samples by using an extraction Kit (Wizard® SV Genomic DNA Purification System by Promega). DNAs were diluted into 5 ng/mL at a ratio of 1:3, with Tris-EDTA buffer before use, and Ex-Taq (Takara) PCR mixture was added to the solution. PCR assay was performed to amplify the target region according to the protocol recommended by the manufacturer. An invader assay was performed, as previously described.5 Because of the low frequency of the variant, we included one positive control (hetero) as a reference to discriminate the possible minor allele from the bulk (Supplementary Figure 1A). Oligonucleotide sequences for the invader probe are shown in Supplementary Table 1.

Immunohistochemistry Staining

All animal experiments were approved and performed under the regulation of the Institutional Animal Care and Use Committee of Tokyo Medical and Dental University. The immunohistochemistry procedure was carried out by using a previously described protocol.6 In brief, hearts were harvest from wild-type C57BL/6 mice followed by perfusion with saline. Cryosections from fixed hearts were incubated with an anti-Tks5 antibody (CUSABIO TECHNOLOGY LLC CSB-PA080044, 1:100) and an anti-F4/80 antibody (Abcam ab6640, 1:100), followed by second antibodies conjugated with Alexa Flour 488 and 568 (Life Technology, Thermo Fisher Scientific). The images were captured on a Leica DM5000B microscope by using Leica Suite software.

Flow Cytometry

Cells in the atrium and the ventricle were dissociated by incubation with 0.1% collagenase in phosphate buffered saline (PBS) for 15 min, and the cells were stained with an anti-F4/80-APC antibody (Biolegend BL123115, 1:100) and an anti-Tks5 antibody (Cusabio, CSB-PA080044, 1:200 dilution). An anti-Tks5 antibody (Cusabio, CSB-PA080044, 1:200 dilution) was directly labeled with Alexa Flour 488 using Zenon® Labeling Complex & Direct (Thermo Fisher Scientific Z-25302, 1:100) in 2% fetal bovine saline (FBS)/PBS. The stained cells were analyzed using a FACS-ARIAIII (BD Bioscience).6

Lentivirus Vector (pLenti6.3 WT-Tks5 and pLenti6.3 L396R Tks5) Construction

Mouse cDNA was obtained by using a SuperscriptIII reverse transcription kit (Invitrogen, Waltham, USA) and using total RNA isolated from mouse heart tissue with an RNeasy Mini Kit (QIAGEN, Venlo, Netherland ) as templates. Three fragments of Tks5 overlapping each other were amplified by high fidelity PCR reaction (KOD-plus neo, TOYOBO, Osaka, Japan). After confirmation of the sequences, 3× Tks5 fragments were connected to obtain the full length of the open-reading frame of murine Tks5. The Tks5 plasmid with L396R point mutation (NM_008018.4:1544, homologous mutation of rs202011870, Supplementary Table 2) was generated by PCR reaction by using the 4 primer method.7 Both WT-Tks5 and L396R Tks5 were sub-cloned into the pEGFP-C1 expression vector to test expression efficacy in HEK293A cells. They were sub-cloned into the pENTR vector (Invitrogen), and then into the lentivirus vector (pLenti6.3/DESTTM GetewayTM vector kit). The sequences of 3 primers used for obtaining WT Tks5 and L396R mutant Tks5 are shown in Supplementary Table 3.

Cell Culture

RAW264.7 cells (RIKEN BioResource Center, Tsukuba, Japan) were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) (Nacalai Tesque, Kyoto, Japan), supplemented with 10% inactivated FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. HL-1, a mouse atrial myocyte cell line, was kindly gifted from Dr. Claycomb (University of Louisiana, New Orleans, LA, USA).8,9 HL-1 cells were cultured on dishes coated with gelatin/fibronectin. Cells were cultured in Claycomb medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, 0.1 mmol/L norepinephrine, and 2 mmol/L L-glutamine. HEK293A (Invitrogen Thermo Fisher) were maintained in DMEM containing 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. HEK293FT cells were maintained in DMEM containing 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 500 μg/mL Geneticin.

Creation of Tks5 Over-Expression Cell Lines

Both pLenti6.3-EmGFP WT Tks5 and pLenti6.3-EmGFP L396R Tks5 were transfected into HEK293FT cells by polyethylenimine PEIMax (Polysciences, Warrington, Pennsylvania, USA) simultaneously with the plasmids containing other packaging components (pMDLg-pRRE, pRSV-Rev, and pVSV-G). Virus-containing supernatants were harvested 48–72 h post-transfection, followed by centrifugation (1,000 g, 5 min, 4℃) to pellet cell debris, and by filtration with a 0.22-μm PolyVinylidene DiFluoride (PVDF) filter unit. The prepared WT Tks5 and L396R Tks5 virus lysates (1 mL) were applied to RAW264.7 in 6-well plates (1×105 per well). Seventy-two h post-transduction, Blasticidin (6ug/mL) was applied to cells to select cells stably expressing EmGFP-Tks5. After 10–12 days of Blasticidin application, almost all cells became GFP positive. All experimental procedures were performed in parallel for both WT-Tks5 and L396R Tks5.

Rhodamine-Phalloidin Staining

Raw264.7 cells were seeded on 3.5 cm glass-bottom dishes (Matsunami Glass, Japan). Cells were fixed with 4% paraformaldehyde (PFA), permeabilized with 0.5% Triton-X100, and followed by incubation with 0.1 μmol/L Rhodamine-Phalloidin (Cytoskeleton, Inc. USA) and 4’,6-diamidino-2-phenylindole (DAPI) (0.1%) sequentially. Cells were observed under a confocal microscope (Zeiss, LSM710) with an objective lens of ×63 magnification.

Application of Cyclic Stretch to Hl-1 Cells

HL-1 cells were seeded on silicone chambers (32 mm×32 mm, STREX, pre-coated with 0.1% Galatin) at a cell density of 1×106 per 1 chamber the day before stretch. After cells were confirmed to adhere to the silicone bottom the next day, cells were subjected to periodical mechanical stretch (Cell Strain ST-140, STREX) at a frequency of 1.0 Hz and at a stretching rate of 20% in serum-free culture medium (Claycomb with supplement of 100 U/mL penicillin, 100 μg/mL streptomycin, 0.1 mmol/L norepinephrine, and 2 mmol/L L-glutamine) for 24 h. Conditioned medium was collected, centrifuged (1,100 g, 15 min, 4℃), and filtered with 0.22 µm MILLEX-GV PVDF filters to remove cell debris.912

Trans-Well Assay

RAW264.7 cells were first seeded on the PET membrane with 8 μm pores (ThinCertTM, Cat#665638, Greiner Bio-One, Australia). The PET membrane insert was placed on a 12-well dish with 1 mL DMEM culture medium per well. Cells were incubated for 4 h to consolidate adherence to the membrane, followed by replacing DMEM culture medium with HL-1 conditioned medium in the wells. First, we performed titration of the duration of time exposed to HL-1-conditioned medium. An exposure for >4 h resulted in noticeable cell dispersion on the bottom of the dish, whereas exposure for <3 h resulted in a negligible number of cells dropping to the bottom of the dish. Thus, in subsequent experiments, we incubated the PET membrane in HL-1-conditioned medium for 3 h. Then, the PET membranes were removed, fixed with 4% PFA, permeabilized with Triton-X100, and stained with DAPI and Rhodamine-Phalloidin. The membranes were observed under a confocal microscope (ZEISS LSM710, Germany). Cell migration through the membrane was evaluated as a ratio of DAPI fluorescence beneath the membrane relative to DAPI fluorescence above the membrane. The co-localization of EmGFP-WT (or L396R) Tks5 and Rhodamine are quantified in Pearson’s coefficient and Mender’s co-localization coefficient (MCC) of EmGFP in each cell. The Pearson’s coefficient R value reflects the correlation of EmGFP and Rhodamine by the intensity of each pixel; the MCC of EmGFP reflects the fraction of EmGFP intensity located in pixels where there is a Rhodamine signal above threshold.

Statistical Analysis

Continuous data were expressed as mean±SEM. All image quantifications were performed with Image J software. Statistical analysis was conducted with Graphpad Prism5.0. Fisher’s exact test was used to examine the significance (P value) of the association of AF with the variant. The column statistic of expression level, migration velocity, podosome percentage and trans-well effect was performed with a 2-tailed Student’s t-test, and 1-way ANOVA and Bonferroni post-test for comparison of the designated 2 groups.

Results

Replication of an AF-Sensitive Rare Variant by Genotyping

ExWAS identified 1 rare variant, rs202011870 (MAF=0.06%) (Table 1). rs202011870 is highly associated with AF (OR=3.617, P<0.0001). This variant is localized in the protein-coding region of the TKS5 gene and is associated with the amino acid replacement of Leu (L) with Arg (R). To replicate the association of rs202011870 with AF, 3,378 genomic DNA samples (2,737 AF cases and 641 controls) were extracted from whole blood and genotyped with an Invader assay. The possible heterozygous (n=27) or minor allele homozygous (n=1) detected with the Invader assay were confirmed by DNA sequencing (Supplementary Figure 1B). We identified 29 validated minor alleles in the AF group (allele, n=5,474), with a frequency of 0.53%. The frequency of this rare variant in the AF population was ∼10-fold of that in the general population. When we cited genotyping data from the East Asia group in the 1,000 genomic project on NCBI (Assay ID: ss488941307), the frequency of this variant in AF turned out to be statistically significantly different from that in the non-AF control (Table 2).

Table 1. Association of AF and rs202011870 Calculated With ExWAS Data
Data analyzed Minor Allele_C Wild-type Allele_A Total
AF 66 16,294 16,360
Control 64 57,154 57,218
Total 130 73,448 73,578

P=4.843E-15 (Fisher’s exact test); Odds ratio=3.617. AF, atrial fibrillation; ExWAS, Exome-Wide Association Study.

Table 2. Association of AF and rs202011870 Calculated With Genotyping Data Plus Eastern Asia Data From NCBI
Data analyzed Minor Allele_C Wild-type Allele_A Total
AF 29 5,445 5,474
Control 4 2,286 2,290
Total 33 7,731 7,764

P=0.033 (Fisher’s exact test); Odds ratio=3.044. AF, atrial fibrillation.

Table 3 shows clinical characteristics of AF patients with and without rs202011870. The onset age of AF was younger by about 15 years for those with rs202011870 than those without. In addition, there were significantly more males that had rs202011870 compared with those without, and the PR interval was shorter in those with rs202011870 than those without.

Table 3. Clinical Characteristics of AF Patients in a Genotyping Assay
  AF_Major Allele (AA)
(n=2,709)
AF_Minor Allele (AC/CC)
(n=28)
P value
Clinical type (%)
 Paroxysmal 64.55 64.29 0.896
 Persistent 27.57 28.57 0.910
 Chronic 7.34 7.14 0.752
 Unknown 0.55 0  
Gender (male, %) 26.34 50 0.009
Age at onset (years) 58.94±0.2849 44.56±3.694 0.0001
Hypertension (%) 53.47 47.37 0.648
Diabetes mellitus (%) 11.97 0 0.154
BMI (kg/m2) 24.05±0.07189 23.47±0.6788 0.413
Congestive heart failure (%) 12.86 10.71 0.737
UCG
 Ejection fraction (%) 65.11±0.2306 64.82±1.871 0.905
 Left atrial diameter (mm) 37.85±0.1498 36.78±1.640 0.501
 Left ventricular diameter (mm) 46.04±0.1483 43.92±1.398 0.196
PR interval (when sinus rhythm)
 PR interval (ms) 170.7±0.8900 157.1±5.167 0.038

AF, atrial fibrillation.

Co-existence of Tks5 and Atrial Macrophages Determined by Immunohistochemistry Staining

Accumulating studies have shown the existence of macrophages that have infiltrated through the endothelium in AF patients’ atria13,14 and the role of Tks5 in invasive cells, such as tumor cells and osteoclasts.1417 In order to determine whether Tks5+ cells represent F4/80+ macrophages in the heart, we stained postnatal adult atrial and ventricular sections (at 4 weeks) by using immunohistochemistry. Tks5+:F4/80+ double-positive cells were present both in the atrium (Figure 1A-a) and in the ventricle (Figure 1A-b). Tks5+:F4/80+ double-positive cells were mainly localized near the endothelium of the atrium and the ventricle.

Figure 1.

Immunohistochemistry staining and flow cytometry analysis of Tks5 and F4/80 in mouse atrium and ventricle. (A) Atrium (panel a) and ventricle (panel b) immunostained for Tks5 and F4/80. Arrowheads indicate Tks5+:F4/80+ double-positive cells. (B) Expression of F4/80- and Tks5-positive populations in the atrium (panel a) or the ventricle (panel b) of the heart (postnatal 10 days). To determine the threshold for expression of Tks5-488 and F4/80-APC, cells un-stained with Tks5-488 or F4/80-APC were also analyzed (Left panel).

Next, to investigate the quantitative numbers of Tks5+ and F4/80+ cells, we used flow cytometric analysis of the cells isolated from the hearts of mice at the 10-day postnatal mark. Results showed that dual-positive cells for Tks5 and F4/80 were present both in the atrium (3.29%) and in the ventricle (3.89%) (Figure 1B).

Enhancement of Trans-Well Migration by L396R Tks5

Next, we examined whether L396R Tks5 affected the migration of macrophages. RAW264.7 cells overexpressing murine wild-type Tks5 (WT-ove) or Tks5 with homogenous mutation of human L396R (L396R-ove) were generated (Figure 2A). We carried out a macrophage migration assay in a trans-well model (Figure 2B). At the basal condition, macrophages with L396R Tks5 overexpression exhibited increased trans-well migration ability compared to control macrophages and macrophages with WT Tks5 overexpression (Figure 2C). There was no significant difference in trans-well migration ability between control macrophages and macrophages with WT Tks5 overexpression, suggesting that L396R mutant Tks5, but not WT Tks5, enhances macrophage migration in the basal condition.

Figure 2.

Promoted migration ability in macrophages with L396R-Tks5 overexpression. (A) rs202011870 homologous site in murine Tks5 causes the same amino acid transformation as in human Tks5. (B) Illustration of trans-well assay. (C) Comparison of cells migrated through the trans-well determined by the ratio of DAPI staining between below the membrane and above the membrane. (D) Experimental design for trans-well assay in the presence of a chemoattractant. (E) Comparison of trans-well migration of cells with L396R Tks5 overexpression and cells with WT Tks5 overexpression cultured in vehicle medium, medium obtained from non-stretched HL-1 cells, or medium obtained from stretched HL-1 cells. Nonst, medium obtained from non-stretched HL-1 cells, St, medium obtained from stretched HL-1 cells, WT-ove, cells with WT Tks5 overexpression. *P<0.05; **P<0.01; ***P<0.001; NS, not significant.

Left atrial dilatation is one of the strongest factors to provoke AF. We wanted to know whether the stretch of atrial myocytes enhanced macrophage migration. For these purposes, we carried out experiments on a chemotaxis assay in a trans-well model (Figure 2D). HL-1 cells were subjected to periodic mechanical stretch for 24 h, and conditioned medium was collected. RAW 264.7 cells were pre-seeded in a cell culture insert with pores on its underside. The trans-well ability of RAW264.7 cells was evaluated after applying HL-1 conditioned medium. In both control macrophages and those with WT Tks5 overexpression, the addition of condition medium increased trans-well migration. For macrophages with L396R Tks5 over-expression, the addition of conditioned medium further enhanced the trans-well migration ability, which was significantly higher than that for control macrophages and for those with WT Tks5 overexpression (Figure 2E).

Podosome Formation and Enhanced Trans-Well Migration in Cells With L396R Tks5

Tks5 is implicated in podosome formation and podosome functions in cell migration,1519 and thus we next examined whether L396R Tks5 affected podosome formation. We performed Rhodamine-Phalloidin staining of RAW264.7 cells. Podosome is defined as F-actin puncta with a diameter >0.4 μm on the periphery of cells17,20,21 (white arrows in Supplementary Figure 2A,B). In cells with WT-Tks5 overexpression, actin was distributed diffusely throughout the cytoplasm, sometimes accumulating as a ring-like assembly or puncta cluster in the cell body (white arrow heads in Figure 3A), but podosomes were rarely observed. In contrast, in cells with L396R Tks5 overexpression, actin was distributed mainly at the cell periphery and this formed podosomes (white arrows in Figure 3B,3C; Supplementary Figure 2C,2D). Cells with L396R Tks5 overexpression appeared to have a more stretched cell shape and protrusions with podosome assemblies. These protrusions were similar to sharply differentiated protrusions identified in WT Tks5 overexpression cells only after application of PMA (phorbol 12-myristate 13-acetate), a PKC activator that can induce the formation of podosomes in macrophages (Supplementary Figure 2A,2B).20

Figure 3.

Podosomes and cell morphology by Rhodamine-Phalloidin staining. (A,B) Rhodamine-Phalloidin staining of RAW264.7 cells with WT Tks5 over-expression (WT_Tks5 ove) (A) and those with L396R Tks5 over-expression (L396R_Tks5 ove) (B). Red staining indicates F-actin, stained with Rhodamine-Phalloidin; Blue indicates nuclei stained with DAPI. White arrows in (B) indicate podosomes of cells with L396R Tks5 over-expression. Scale bars: 10 μm. (C) Comparison of the fraction of cells with podosomes. *P<0.05, **P<0.01.

To further assess the role of Tks5 on macrophage migration, we examined co-localization of exogenous Tks5 and podosome formation, detected by EmGFP expression and Rhodamine-Phalloidin staining. We set 2 layers in the confocal z-stack images taken from the same field: 1 layer is just beneath the membrane (layer_m), and the other is close to the leading edge of the cells crossing through the membrane pores (layer_l). Layer_l is approximately 6 μm away from the layer_m toward the cell leading edge (Figure 4A). DAPI staining observed in layer_m but not in layer_l was more frequent for L396R Tks5 (Figure 4C) than for WT Tks5 (Figure 4B), suggesting that macrophages with L396R Tks5 overexpression migrated longer distances compared to those with WT Tks5 overexpression. The integrated area of co-assembly of Tks5 with F-actin at the cell leading edge, namely in layer_l (Figure 4D), and the number of cells with co-assembly of Tks5 with F-actin (Figure 4E), were significantly more for L396R Tks5 than for WT Tks5. In contrast, a Pearson’s R value for the EmGFP and Rhodamine fluorescence at layer_l was not significantly different between WT Tks5 (0.484±0.174) and L396R Tks5 (0.470±0.127) (Figure 4F); Mender’s co-localization co-efficient was not significantly different either between WT Tks5 (0.569±0.169) and L396R Tks5 (0.584±0.164) (Figure 4G).22 These findings indicate that exogenous Tks5 and F-actin assembly participate in the trans-well activity in cell-leading parts, regardless of WT Tks5 or L396R mutant Tks5, but that the incidence of co-assembly with F-actin in cell-leading parts is significantly higher in L396R mutant Tks5 than in WT Tks5.

Figure 4.

Enhanced cell protrusions of the trans-well cells with L396R Tks5. (A) Cell images that transverse the membrane were selected from 2 layers from Z-stack images of one field. One layer is just beneath the membrane (layer_m) and the other is near the leading edge of the cells (layer_l), which is 6 μm away from layer_m. (B,C) Images of L396R Tks5 overexpression cells (B) and WT Tks5 overexpression cells (C) emerging beneath the membrane. White arrows indicate cells with a nucleus just emerging at layer_m, which cannot be detected at layer_l. Layer_l is approximately 6 μm away from layer_m toward the cell leading edge. Scale bar: 50 μm. (D) Comparison of an integrated area stained with Rhodamine-Phalloidin in cells beneath the membrane. (E) Comparison of percentage of cells with podosome protrusion per field beneath the membrane. (F,G) Quantification of co-localized EmGFP and Rhodamine in cells captured at level_l. Pearson’s R value (F) reflects the correlation of EmGFP and Rhodamine by intensity of each pixel; Mender’s co-localization coefficient of EmGFP (G) reflects the fraction of EmGFP intensity located in pixels where there is a Rhodamine signal. *P<0.05, **P<0.01.

Enhanced Expression of Inflammatory Molecules by L396R Tks5

It is generally believed that macrophages infiltrated into the atrium provoke inflammation, and thus make atria substrate more susceptible to AF.2325 Then, we examined the expression levels of several classical inflammation cytokines and inflammation-related molecules.24 Expressions of CCR-2, TNF-α, TGF-β, ICAM, MYD88, and MMP9 were significantly elevated in L396R-ove cells than cells expressed with WT-Tks5 (Figure 5A–F). Expression of NOS1 or NOS2 were not changed (Figure 5G,5H), and MMP2 expression was not detected.

Figure 5.

Elevated expression level of pro-inflammatory cytokines and signal molecules. Expression level of CCR2 (A), TNFα (B), TGFβ (C), ICAM1 (D), MyD88 (E), MMP9 (F), NOS1 (G), and NOS3 (H). *P<0.05; **P<0.01; ***P<0.001; ns, not significant.

Discussion

The major findings of the present study are 2-fold: (1) we validated the association of rs202011870 with AF in an independent Japanese cohort; and (2) rs202011870 enhanced trans-well migration ability and increased expression of some inflammatory cytokines in macrophages.

In this study, we confirmed the association of rs202011870 with AF in Japanese cohorts. The OR for rs202011870 is more than 3 in our study, which is consistent with the ExWAS in originally identifying of the association of rs202011870 with AF. The rs202011870 frequency was 0.2% in the control group and 0.5% in the AF group in our study, which is also quite consistent with the original ExWAS study; the rs202011870 was 0.1% in the control group and 0.4% in the AF group. It is of worth noting that the gnomAD browser reports ethnic differences in the prevalence of rs202011870; it is most prevalent in the East Asian population (0.456%), followed by the South Asian population (0.03593%), and has not been detected in African, Latino, Ashkenazi Jewish or the European population (https://gnomad.broadinstitute.org/variant/10-105371375-A-C?dataset=gnomad_r2_1). Thus, rs202011870 could be important, especially in the East Asian population, including Japan. The onset age for AF was younger by ∼15 years for AF patients with rs202011870 than those without, indicating that rs202011870 has special importance for AF in the relatively young population. Tks5 is a scaffold protein containing 1 PX domain and 5 SH3 domains. Leu396 is localized between the second and third SH3 domain, which is a highly preserved region across various species. This region contains 3 possible phosphorylation sites (Ser393, Ser392, Ser378) just prior to Leu396. Polyphen2, which is software that is free online, indicates that a L396R mutation is possibly damaging, with a score of 0.843 (sensitivity 0.83. specificity 0.93).

We found that Tks5 with L396R mutation enhanced trans-well migration of macrophages, especially those that are induced by chemoattractant released from stretched HL-1 cells at a constant rhythm. It would of interest to observe if chemoattractant released from HL-1 cells stretched at an irregular rhythm; this should be examined in a future study. We also found that Tks5 with L396R mutation enhanced podosome formation in macrophages. These findings are in line with the fact that podosome is a structure implicated in cell migration,16,17,19 and that Tks5 plays some role in podosome formation. The role of Tks5 in the formation of podosomes is not fully understood yet. Tks5 is proposed to have a role in podosome formation in a contactin-dependent manner. Tks5 is shown to form an assembly with key actin regulators, NCK1, NCK2, N-WASP, and growth factor receptor-bound protein 2 (GRB2), and the assembly interacts with the small GTPase, CDC42, and dynamin depending on cortactin.18 How the Tks5 mutation of L396R mutation affects assembly formation with CNK1, NCK2, N-WASP, and GRB2, and how L396R mutation of Tks5 affects the interaction of this assembly with CDC42 and dynamin will be an important topic for future study.

Macrophages with overexpression of L396R Tks5 exhibited expression of inflammatory cytokines and molecules related to inflammation, CCR-2, TNF-α, TGF-β, ICAM, and MYD88. Thus, macrophages with L396R Tks5 polymorphism appear to have enhanced an inflammatory property. Our data also showed that MMP9 was upregulated in macrophages with L396R Tks5 overexpression. It is not known how L396R Tks5 causes expression changes in these molecules, and this should be examined in a future study. Tks5 is also implicated in reactive oxygen species (ROS) production via NADPH oxidase (NOX).24,25 Tks5 is composed of a component of NOXA1 activator for NOX1 and NOX3, but not NOX2 or NOX4. In macrophages, although NOS2, a phagocytic NOX, is most abundantly expressed in phagosomes, NOX1 and NOX3 are also expressed to a lesser extent in plasma membrane. Tks5 is shown to enhance ROS production through NOX1 and NOX3. ROS induces MAPK, STAT1, STAT6 and NFκB activation,12,26 resulting in polarization of macrophages into more inflammatory macrophages. ROS is known to induce expression of several MMPs via NFκB activation.27 Thus, it could be possible that alteration of ROS production by L396R occurs, resulting in transcriptional enhancement of molecules related to inflammation and macrophage invasion. Further studies are certainly needed to test this possibility.

It is possible to assume that macrophages with Tks5 expression also invade the ventricle and promote development of ventricular arrhythmias. Our immunohistochemistry analysis and flow cytometry analysis showed few cells positive for both Tks5 and F4/80 from ventricular sample (Figure 1A-b and 1B-b). However, in our studies, no patients exhibited ventricular tachycardia or fibrillation. Thus, currently, we do not have data detailing the role of Tks5 and its genetic polymorphism in ventricular arrhythmias. We cannot eliminate the possibility that rs202011870 is a very rare variant and thus we failed to detect the role of Tks5 and its genetic polymorphism in ventricular arrhythmias. Further studies are needed to clarify the role, if any, of Tks5 in ventricular arrhythmias.

In conclusion, L396R mutation in Tks5 associated with AF enhances migration of macrophages and their inflammatory features, resulting in enhanced susceptibility to AF. Whether Tks5 and its related pathways could be a novel target for drug development is unknown and is an intriguing topic to pursue.

Acknowledgments

We express our gratitude to the patients and their families for their contribution to this study. We thank Mrs. Akemi Oshikiri for her secretarial work and all lab members for supporting our study.

Disclosures

This study was supported by the Joint Usage/Research Program of the Medical Research Institute, Tokyo Medical and Dental University, a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology, Japan (18H02803), and Grants-in-Aid from the Japan Agency for Medical Research and Development (AMED; 17ek0210082 h0001).

IRB Information

All the experiments protocols with human samples and animal experiments protocols were approved and performed under the regulation of the Research Ethical Committee (2013-016C17) and Institutional Animal Care (0170013A) and Use Committee (2009-33-5) of the Tokyo Medical and Dental University.

Data Availability

The deidentified participant data will not be shared.

Supplementary Files

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-20-0101

References
 
© 2020 THE JAPANESE CIRCULATION SOCIETY

This article is licensed under a Creative Commons [Attribution-NonCommercial-NoDerivatives 4.0 International] license.
https://creativecommons.org/licenses/by-nc-nd/4.0/
feedback
Top