Transthyretin amyloid fibrils alter primary fibroblast structure, function, and inflammatory gene expression

Abstract

Age-related wild-type transthyretin amyloidosis (wtATTR) is characterized by systemic deposition of amyloidogenic fibrils of misfolded transthyretin (TTR) in the connective tissue of many organs. In the heart, this leads to cardiac dysfunction, which is a significant cause of age-related heart failure. The hypothesis tested is that TTR affects cardiac fibroblasts in ways that may contribute to fibrosis. When primary cardiac fibroblasts were cultured on TTR-deposited substrates, the F-actin cytoskeleton was disorganized, focal adhesion formation was decreased, and nuclear shape was flattened. Fibroblasts had faster collective and single-cell migration velocities on TTR-deposited substrates. In addition, fibroblasts cultured on microposts with TTR deposition had reduced attachment and increased proliferation above untreated. Transcriptomic and proteomic analyses of fibroblasts grown on glass covered with TTR showed significant upregulation of inflammatory genes after 48 h, indicative of progression in TTR-based diseases. Together, results suggest that TTR deposited in tissue extracellular matrix may affect the structure, function, and gene expression of cardiac fibroblasts. As therapies for wtATTR are cost-prohibitive and only slow disease progression, better understanding of cellular maladaptation may elucidate novel therapeutic targets.

NEW & NOTEWORTHY Transthyretin (TTR) cardiac amyloidosis involves deposition of fibrils of misfolded TTR in the aging human heart, leading to cardiac dysfunction and heart failure. Our novel in vitro studies show that TTR fibrils alter primary cardiac fibroblast cytoskeletal and nuclear structure and focal adhesion formation. Furthermore, both fibrillar and tetrameric TTR significantly increased cellular migration velocity and caused upregulation of inflammatory genes determined by transcriptomic RNA and protein analysis. These findings may suggest new therapeutic approaches.

INTRODUCTION

Amyloidosis is a condition characterized by deposition of misfolded proteins in the extracellular matrix (ECM) of tissues and organs, which can occur with human aging. The best-known amyloid is the extracellular amyloid beta (Aβ) deposition in the brains of patients with Alzheimer’s disease. In the heart, another common form of protein amyloidosis is due to deposited misfolded transthyretin (TTR), which is a carrier of thyroxine and retinol binding protein. TTR is a 127 amino acid, 55-kDa transport protein primarily synthesized in the liver, after which it is secreted and circulates systemically through the vasculature. In a healthy state, TTR exists in a tetrameric complex of four single-chain TTR monomers. With aging, TTR can become thermodynamically unstable and unfold into a monomeric state, which deposits systemically to form amyloidogenic fibrils in the connective tissues of various organs such as peripheral nerves causing neuropathy (1). In the heart, age-related wild-type transthyretin amyloidosis (wtATTR) leads to fibrosis and stiffening with resulting cardiac diastolic dysfunction, arrhythmia, and progression to heart failure (2–4). Although wtATTR has been previously underdiagnosed due to the slow progression of the disease and lack of diagnostics, recent nuclear imaging techniques have led to a better understanding of wtATTR prevalence in the elderly population (5). Up to 13% of all elderly patients having heart failure with preserved ejection fraction (HFpEF) may be due to TTR amyloidosis (2), and this population is likely to increase as the number of persons worldwide aged over 80 is expected to increase in the future. As of now, the sole FDA-approved therapy for wtATTR is tafamidis, which stabilizes the TTR tetramer and prevents unfolding and subsequent amyloid deposition to prevent disease progression (4, 6). However, tafamidis is cost-prohibitive as it is the most expensive cardiovascular therapeutic to date (7); furthermore, there is currently no way to remove the TTR fibrils once they are deposited in the cardiac interstitium, or to treat cells negatively affected by the TTR amyloid deposits.

Clinically, it is understood that amyloidogenic TTR deposition can trigger production of proinflammatory cytokines in patients that present with TTR-related familial amyloid polyneuropathy (FAP) (8) and that inflammation is associated with disease progression (1). Furthermore, upregulation of matrix remodeling associated genes such as matrix metalloproteinase-9 (MMP-9) have been observed in patients with FAP (8). Blood proteomic analysis of patients with cardiac symptoms due to wtATTR has also shown involvement of immune response in this disease pathology (9).

There is little information about how TTR affects individual cells in the heart. Over 5,000 publications describe the diagnosis and clinical consequences of human cardiac amyloidosis, with over a 1,000 on TTR amyloidosis alone. One group has reported on isolated cardiomyocytes in the presence of misfolded TTR fibrils and shown the fibrils and preceding oligomeric species exhibit cytotoxicity, increased reactive oxygen species (ROS) production, altered cytoplasmic calcium levels, and impaired rate of repolarization (10). Surprisingly, as of yet, there is no work published on how TTR affects the structure and function of individual cardiac fibroblasts, the major cell type in the cardiac ECM and known mediators of cardiac inflammation in heart failure (11). In addition, there are no validated transgenic rodent models for wtATTR, necessitating the need for alternative approaches (12). The focus here is to understand the effects of underlying TTR fibrils or tetramers on isolated primary cardiac fibroblast behavior in culture when grown on glass or a microtopographic culture substratum. This study reports how fibroblast migration, proliferation, cytoskeletal architecture, and gene expression alter in response to TTR on the surface underlying the cells in culture.

METHODS

TTR Fibril Production and Substrate Preparation

Recombinant human transthyretin protein (Invitrogen Cat. No. LF-P0054, Carlsbad, CA) was induced to a fibrillar state according to previously described protocols (13). Briefly, 1 M acetic acid and 2 M NaCl were added to TTR protein reconstituted in water at 0.5 mg/mL, to a final concentration of 50 mM acetic acid and 100 mM NaCl (pH 3.0). Fibril self-assembly was allowed to occur for at least 72 h and fibrils were stored at 4°C for subsequent use in experiments. To study TTR fibril or tetramer effects on fibroblasts, glass or polystyrene culture dishes were coated with 10 µg/mL fibronectin with or without the addition of 50 µg/mL TTR fibrils or native tetramer in PBS solution. Dishes were coated for 2 h in a 37°C, 5% CO₂ incubator before cell plating. Micropost substrates were prepared by molding 400 kPa polydimethylsiloxane (PDMS) from a parylene template to produce spaced posts as done previously, with microposts 15 µm high, 25 µm wide, in a 75 µm tetragonal array (14, 15). Before coating with fibronectin or TTR fibrils, PDMS micropost substrates were functionalized with 3-aminopropyl triethoxysilane (Sigma–Aldrich, St. Louis, MO).

Although the commercially available recombinant human TTR is over 95% pure via SDS-PAGE, possible lipopolysaccharide (LPS) contamination resulting from the Escherichia coli expression system was tested with the Pierce Endotoxin Quant Kit (Thermo Fisher, Cat. No. A39553). Cell culture dishes were prepared with fibronectin or fibronectin plus TTR fibrils to mimic the solution seen by the fibroblasts in these experiments, and both conditions were in the 0.12–0.13 endotoxin units (EU)/mL range indicating that TTR addition did not affect overall endotoxin concentration. Furthermore, this is well below the stated 1.0 EU/mL limit of detection and in line with commercially available “low endotoxin” recombinant human TTR that is instead expressed in HEK cells rather than E. coli.

Immunofluorescence of Coated Substrates

To ensure even coverage of culture substrates with both fibronectin and TTR fibrils, substrates were fixed with 10% formalin post-coating and stained with antibodies for fibronectin (Invitrogen Cat. No. MA5-11981) and transthyretin (Invitrogen Cat. No. PA5-35315) at 1:50 dilution in 1% BSA, 0.1% Tween-20 solution, and imaged on a Zeiss LSM880 confocal microscope. Fluorescently labeled features were manually measured with FIJI software. Largest feature (or group of features) was in the ∼8-μm-diameter range. However, the majority of features were qualitatively in the 0.5–2-μm range or smaller.

Scanning Electron Microscopy

Native tetrameric TTR or TTR fibrils (50 μg/mL) prepared as previously described were deposited onto glass substrates for 2 h in a 37°C, 5% CO₂ incubator. Substrates were rinsed with PBS and allowed to air dry completely. Samples were sputter coated at low vacuum with platinum/palladium using a Cressington 208HR (High Resolution) coater, 6.0 nm. They were mounted on aluminum stubs by use of double-sided sticky carbon tape. Sample surfaces were imaged using a secondary electron detector (SED) with a JEOL JSM-6320F field emission scanning electron microscope (FESEM).

Atomic Force Microscopy

Twenty microliters of 450 μg/mL of TTR fibrils prepared as previously described were plated onto oxygen-plasma-treated glass slides before scanning. Atomic force microscopy (AFM) was performed using a NanoWizard Ultra Speed A AFM in soft tapping mode, using a NCHV-A Bruker Tip (320 kHz) (Bruker, Billerica, MA) and projections were visualized using the Gwyddion software (Czech Metrology Institute, Jihlava, Czechia).

Rat Cardiac Primary Fibroblast Cell Culture

All research animals were obtained and used in accordance with the guidelines of the NIH [National Research Council (US) Institute for Laboratory Animal Research, 1996]. Animal studies were approved by UIC Institutional Animal Care and Use Committee and conducted according to the NIH’s Guide For The Care And Use Of Laboratory Animals. Hearts were removed and cells isolated from 1- to 2-day-old Sprague–Dawley rats using collagenase type II (Worthington, Lakewood, NJ) as previously described (16). Cells were plated in 10-cm tissue culture dishes and neonatal rat ventricular fibroblasts were given 1 h to attach before removing surrounding media along with unattached cells. Fibroblasts were incubated in DMEM (10% FBS, high glucose, pyruvate) (Thermo Fisher, Waltham, MA), harvested with trypsin/EDTA (Corning Inc., Corning, NY), and plated at desired density on fibronectin-coated or fibronectin with TTR fibril-coated dishes. Experiments were conducted on cells after one passage.

Barrier Removal and Single-Cell Migration Assays

Barrier removal.

Following glass substrate coating with fibronectin and TTR fibrils, removable polydimethylsiloxane chambers with a cell-free exclusion zone (IBIDI, Martinsried, Germany) were added. Fibroblasts were plated in DMEM + 10% FBS into each chamber and allowed to reach confluency over 48 h of culture. At this time, the insert was manually removed, leaving a zone of cell exclusion between chambers. Dishes were cultured for 8 h in a 37°C, 5% CO₂ incubator. Phase microscopy images were taken at barrier removal and after 8 h. The area of the cell-free exclusion zone was measured using ImageJ, allowing calculation of cell front velocity. Velocity was determined across five independent biological runs, with three random regions analyzed per dish.

Single-cell migration.

Fibroblasts were plated at low density (1,300 cells/cm²) on glass substrates coated with fibronectin or fibronectin with either fibrillar or tetrameric TTR. Cells were allowed to attach for 2 h, at which point cell nuclei were stained using Hoechst 33342 (Thermo Fisher) in DMEM + 10% FBS. Plates were placed on a Zeiss LSM 880 confocal microscope equipped with a 37°C, 5% CO₂ incubation chamber and an automated stage. Images of each well were acquired every 10 min for 6 h. Stacks of images were analyzed using ImageJ plugin TrackMate (17). Identified nuclei in consecutive time frames were interpolated and the ratio between the found displacement and the known time passed was calculated as a velocity for each single displacement. Velocities over 75 µm/h were considered artifact and ignored. The sum of the single displacements for an individual cell led to the identification of the total track. The mean velocity for each cell was calculated averaging the velocities of all given total tracks. All the tracks for each cell measured across the three experiments were grouped (reaching n > 300 tracks) and analyzed.

Quantitative Cell Morphology: Actin Organization and Nuclear Shape

Fibroblasts grown on glass substrates coated with fibronectin or fibronectin with TTR fibrils were fixed in 10% formalin after 48 h of culture. Cells were probed with 1:500 rhodamine phalloidin (Thermo Fisher) in 1% BSA, 0.1% Tween-20 solution, and counterstained with DAPI-containing mounting medium (Vector Laboratories, Burlingame, CA) and imaged on a Zeiss AxioObserver microscope and AxioVision software. To measure actin fiber polarity, the Matlab program GTFiber (18) was used to generate skeletonized actin fibers with a color-coded orientation for each cell to calculate the decay of the orientational order parameter (S2D) of the fibers. Two-dimensional nuclear size and three-dimensional nuclear volume was determined by imaging DAPI-stained nuclei on a Zeiss LSM880 confocal microscope with 0.5-μm Z-stack slices and analysis with a custom CellProfiler pipeline (19). Fifty cells pooled across five independent biological experiments were used to calculate actin polarity, nuclear shape, and nuclear volume.

Analysis of Focal Adhesion Formation

Fibroblasts grown on glass substrates coated with fibronectin or fibronectin with TTR fibrils were fixed in 10% formalin after 48 h of culture. Cells were permeabilized with 0.1% Triton X-100 (Sigma-Aldrich) and probed with a 1:250 antibody for paxillin (Abcam Cat. No. ab32084) in 1% BSA, 0.1% Tween-20 solution. Cells were counterstained with secondary antibody (AlexaFluor 488, Thermo Fisher), 1:500 rhodamine phalloidin (Thermo Fisher), and DAPI-containing mounting medium (Vector Laboratories, Burlingame, CA) and imaged on a Zeiss AxioObserver microscope and AxioVision software. A custom CellProfiler pipeline was used to identify formed focal adhesions at the cell/substrate interface by identifying paxillin spots, and number of formed focal adhesions per parent cell was quantified. At least fifty cells pooled across five independent biological experiments were used to calculate focal adhesions.

Proliferation on Flat and Micropost Substrates, and Attachment on Micropost Substrates

Fibroblasts were plated at 12,500 cells/cm² in 96-well plate (with or without micropost substrates as previously described) coated with fibronectin or fibronectin plus TTR fibrils and allowed to grow for 48 h. To assess relative cell proliferation between groups, the CellTiter Aqueous One Solution Cell Proliferation Assay (MTS assay) (Promega, Madison, WI) was used. Six independent biological runs with technical triplicates in each experiment were used to determine proliferation rate.

Attachment to microposts.

Fibroblasts grown on fibronectin or fibronectin with TTR fibril micropost substrates were fixed in 10% formalin after 48 h of culture. Cells were probed with 1:500 rhodamine phalloidin (Thermo Fisher) in 1% BSA, 0.1% Tween-20 solution, and counterstained with DAPI-containing mounting medium (Vector Laboratories, Burlingame, CA) and imaged on a Zeiss AxioObserver microscope and AxioVision software. For all fields of view captured, a stack of images containing the nucleus (DAPI), the cytoskeleton (rhodamine), and the micropost surface (phase contrast) was acquired. CellProfiler was used to identify the micropost grid, and expand the diameter of all posts by a 10-μm annulus. A nucleus was scored as “near” a micropost if it was >50% contained in any post + 10-μm annulus region, and calculations were performed to determine percent of all nuclei that were near a post. For calculation, 25 random fields of view were acquired across five independent biological experiments.

Gene Expression

Cardiac fibroblasts were grown on glass substrates with fibronectin or fibronectin and TTR fibrils for 48 h. Total RNA was purified using Maxwell RSC simplyRNA Cells (Promega) with inclusion of DNAse treatment step. RNA samples were quantified using NanoDrop One Spectrophotometer (Thermo Scientific) and analyzed for integrity using Agilent 4200 TapeStation. Levels of remaining DNA were checked using Qubit fluorometer (Invitrogen). DNA amounts did not exceed 10% of the total amount of nucleic acid. Sequencing libraries for Illumina sequencing were prepared in one batch in 96-well plate using Stranded CORALL total RNAseq library prep kit with RiboCop rRNA Depletion Kit by Lexogen. In brief, 150–250 ng of total RNA was used for the first rRNA depletion step, then followed by library generation initiated with random oligonucleotide primer hybridization and reverse transcription. No prior RNA fragmentation was done, as the insert size was determined by proprietary size restricting method. Next, the 3′ ends of first-strand cDNA fragments were ligated with a linker containing Illumina-compatible P5 sequences and Unique Molecular Identifiers (UMIs). During the following steps of second strand cDNA synthesis and ds cDNA amplification, i7 and i5 indices as well as complete adapter sequences required for cluster generation were added. The number of PCR amplification cycles was 12, as determined by qPCR using a small preamplification library aliquot for each individual sample. Final amplified libraries were purified, quantified, and average fragment sizes confirmed to be 310 bp by gel electrophoresis using TapeStation. Concentration of the final library pool was confirmed by qPCR, and then subjected to test sequencing in order to check sequencing efficiencies and adjust accordingly proportions of individual libraries. Sequencing was carried out on NovaSeq 6000 (Illumina), SP flowcell, 1×100 nt reads. Details for bioinformatic analysis of RNA-Seq data can be found in the Supplemental Information.

Cytokine Expression Analysis

To confirm cytokine expression changes observed via RNA-Seq, cardiac fibroblasts were grown on glass substrates with fibronectin or fibronectin with fibrillar or tetrameric TTR for 48 h. Cell culture supernatant was collected and pooled, and secreted cytokines were quantified with a commercially available proteome profiler that measures expression of 29 common rat cytokines (R&D Systems No. ARY008). Blots were imaged on a Chemidoc MP imager with multiple exposure times, and mean pixel density of each duplicate array spot was compared with the cells on fibronectin only using ImageJ. Targets with greater than fivefold increase on either TTR species relative to fibronectin only are presented. No cytokines were expressed higher on fibronectin only.

Statistical Analysis

Data for all experiments were collected and organized using Microsoft Excel software (Microsoft, Redmond, CA), and histograms, box plots, and statistical analysis were performed using GraphPad Prism software (GraphPad Software, San Diego, CA). Data are expressed as means ± SE for histograms or box plots with minimum to maximum whiskers with a line at the median. Statistical significance was determined via calculation of two-tailed Student’s t test or ordinary one-way ANOVA when appropriate.

RESULTS

TTR Deposition on Cell Culture Surfaces

The first objective was to establish the uniformity of the surface coverage and determine the nanotopographic features of deposited TTR. Previous morphological work did not view the TTR fibrils on culture surfaces, but showed negatively stained TTR fibrils with transmission electron microscopy after adsorption to formvar/carbon-coated nickel mesh grids, or by atomic force microscopy (AFM) on a freshly cleaved mica substrate (10). Here, a conventional glass dish was prepared for cell culture with fibronectin alone, or with both fibronectin and TTR in tetrameric or fibrillar form. Fibronectin has a uniform coating (red stain) over the glass surface regardless of the presence of TTR (Fig. 1A, right). The glass surface coated with both fibronectin and TTR fibrils showed uniformly dispersed green speckles which were mostly under 1 µm with a few of the largest spots ∼8 µm, as measured in FIJI (Fig. 1A, bottom left), but without any background stain when no TTR was present (Fig. 1A, top left). Scanning electron microscopy (SEM) (Fig. 1B, top low- and high magnification) show TTR fibrillar architecture whereas no features were seen for glass coated with native tetrameric TTR (Fig. 1B, bottom). Furthermore, AFM scans (Fig. 1C) corroborate the SEM images of deposited TTR fibrils revealing nanoscale topographic features as high as 90 nm. Thus, the method for preparation of TTR fibrils produced a uniform coating of nanofibrils over the glass surface whereas the tetrameric molecules had no detectable nanotopography.

Figure 1.

Uniform deposition of nanofibrillar or tetrameric layer of transthyretin (TTR) on cell culture surface. A: immunofluorescent images of glass surface coated with fibronectin (FN) only (top) or fibronectin and TTR fibrils (bottom). Left is stained by antibody for TTR (green), and right for fibronectin (red) showing uniformity of both fibronectin and TTR fibrils (small green speckles). Scale bar 150 µm. B: scanning electron microscope images of glass coated with TTR fibrils (top) or TTR tetramer (bottom). Scale bar 1 µm (left) and 100 nm (right). C: atomic force microscopy projections in three-dimensional (3-D) (top) and in two-dimensional (2-D) (bottom) of glass coated with TTR fibrils showing filament structure.

Deposited TTR Fibrils Induce Cytoskeletal and Nuclear Architecture Dysregulation in Cardiac Fibroblasts

A quantitative morphological study was performed to determine if the underlying TTR nanofibrillar array on a glass surface was affecting primary cardiac fibroblasts. Compared with fibronectin-only controls, the actin cytoskeleton of cells grown on underlying TTR was highly disorganized as shown by representative images of fluorescent F-actin stained by phalloidin (Fig. 2, A and B). This was quantified in 50 cells using a MATLAB program to calculate the decay of the orientational order parameter of the filaments. Briefly, images of phalloidin-stained fibroblasts were filtered and skeletonized, with color-coded overlays applied to actin fibers with respect to their relative orientation. The orientation of these fibers is used to calculate the decay rate of the orientational order parameter (S2D) which can range from 0.0 (i.e., two perpendicular lines) to 1.0 (i.e., two parallel lines). The fibroblasts on TTR fibrils had a greater decay indicating less polarity (Fig. 2C). When fibroblasts were cultured on TTR fibrils, the nuclear surface area was increased (Fig. 2D) but three-dimensional (3-D) nuclear volume determined by confocal microscopy (Fig. 2E) and cell area (Fig. 2F) were unchanged. Thus, the cytoskeleton was less organized whereas the nucleus was flattened.

Figure 2.

Fibroblasts cultured on transthyretin (TTR)-coated surfaces have disorganized actin cytoskeleton and flattened nuclear shape. Phalloidin-stained fibroblasts with digitally overlaid color-coded F-actin skeleton and corresponding orientational order parameter (S2D) measurements grown on fibronectin (untreated, UT) (A) or fibronectin + TTR fibril-coated (TTR) glass surface (B). C: the organization (S2D) is significantly decreased for cells cultured on TTR fibril-containing surfaces. Nuclear two-dimensional (2-D) surface area is increased (D) but the three-dimensional (3-D) nuclear volume and overall cell area (E) is unchanged for fibroblasts cultured on TTR fibrils (F). All shown as box plots with minimum to maximum and line at median. n = 50 cells. **P < 0.01 (two-tailed Student’s t test). ns, not significant.

Cardiac Fibroblasts Have Fewer Focal Adhesions When Cultured on Deposited TTR Fibrils

The structure and function of fibroblasts was studied further to determine whether the attachment of the cell to the glass substrate was changed and whether the key functional parameter of migration was affected. The number of formed focal adhesions as defined by paxillin spots in a fibroblast grown with or without underlying TTR fibrils (Fig. 3, A and B) was significantly decreased when grown on TTR.

Figure 3.

Transthyretin (TTR) decreases focal adhesion number and increases migration velocity of fibroblasts. A: fluorescent images of fibroblasts grown on fibronectin (left) or fibronectin + TTR fibrils-coated glass substrates (right) were stained for actin (phalloidin, red), focal adhesion (paxillin, green), and nuclei (DAPI, blue). Scale bar = 25 μm. B: quantification of formed focal adhesions (FA), as defined by number of paxillin (green) events per cell. Fibroblasts grown on TTR fibrils (TTR) have fewer formed focal adhesions than for cells on fibronectin only (untreated, UT). Box plots with minimum to maximum and line at median, n = 50 cells, *P < 0.05 (two-tailed Student’s t test). C: leading cell front velocity of fibroblasts in confluent monolayers was faster when cultured on TTR fibrils in a barrier removal assay. Box plots with minimum to maximum with line at median, n = 13 dishes,*P < 0.05. (two-tailed Student’s t test). D: single-cell velocity of fibroblasts plated at low density and live-imaged was increased when grown on TTR fibrils or tetramers. Box plots with minimum to maximum and line at median, n > 300 cell tracks, ****P < 0.0001 (Ordinary one-way ANOVA with multiple comparisons of each TTR group to UT to determine significance). ns, not significant.

Cardiac Fibroblasts Have Increased Migration Velocity When Cultured on Deposited TTR Fibrils or Tetramers

Since cytoskeletal rearrangement could affect cell migration, two methods to measure the migration velocity were used. The wound healing “gap closure” method used barrier removal to assess the collective cell front migration, and results show a significant increase in velocity on the TTR fibril-coated glass surface (Fig. 3C). Furthermore, fibroblasts migrated significantly faster on both TTR tetramers and fibrils compared with cells on the untreated surface using the single-cell migration speed via live cell imaging [Fig. 3D, movies in Supplemental Videos S1 (UT) and S2 (TTR) (all Supplemental material is available at https://doi.org/10.5281/zenodo.4732410)]. Thus, altered fibroblast cytoarchitecture is reflected in the functional increase of cell migration velocity.

TTR Fibrils Lead to Increased Proliferation on Flat and Microtopographic Substrates, with Decreased Attachment to Microposts

Fibroblast proliferation plays a critical role in fibrosis, which is a major clinical feature seen in patients with TTR amyloid deposition (20). Microtopography has a powerful effect on migration and proliferation of several cell types, including fibroblasts (16, 21, 22). However, it is not clear what impact the TTR fibrils nanotopography might have versus microscale topography such as would be found in vivo, which is relevant to formation of the fibrotic scar tissue seen in both the heart and connective tissue around nerves. It is possible the local TTR nanofibrillar surface is more dominant in regulating fibroblast behavior than the microtopography. Therefore, studies on proliferation were done on TTR-coated flat or micropost PDMS substrates. Similarly, here fibroblasts on untreated surfaces cling to the micropost with strong focal adhesions identified by paxillin but cells near the TTR-coated microposts appear to have fewer focal adhesions, implying weaker attachment (Fig. 4A). A method to quantity the percentage of cells attached to microposts is shown (Fig. 4B). Briefly, phase contrast images allow identification of microposts, which are expanded by 10 μm circular annulus with CellProfiler. Identified fibroblast nuclei are scored as “near” a micropost if the nucleus is >50% contained within an identified micropost and its annulus. Cells grown on TTR nanofibrils were less likely to be near a given micropost (Fig. 4C). The standard tetrazolium reduction proliferation assay (MTS assay) showed fibroblasts grown on both flat and micropost substrates coated with TTR had a significantly increased rate of proliferation (Fig. 4D).

Figure 4.

Proliferation is increased with transthyretin (TTR) coating on flat and microtopographic substrates, with reduced attachment to microposts. A: fibroblasts grown on micropost polydimethylsiloxane (PDMS) substrates fibronectin (untreated, UT) (top) or fibronectin + TTR-fibril coated (TTR) (bottom) in fluorescent images, stained for actin (phalloidin, red), focal adhesion (paxillin, green), and nuclei (DAPI, blue). Micropost outline is shown with white dashed circle. Scale bar = 25 μm. B: phase contrast image of tetragonal micropost array (left) and CellProfiler overlay of microposts plus circular annulus of 10 μm. Identified nuclei outside of this designated zone are shown in color, whereas nuclei contained in the zone are identified via purple outlines. Scale bar = 100 μm. C: percentage of nuclei designated as near any given micropost + 10 μm annulus. Cells grown on TTR fibrils were less likely to be distributed near a micropost. Means ± SE, n = 25 random fields of view across five independent experiments, *P < 0.05 (two-tailed Student’s t test. D: fibroblast proliferation by MTS assay, measured at ABS 490, on flat and micropost substrates; cell number relative to UT is increased when cells are cultured on TTR fibrils on both flat or microtopographic substrates. Means ± SE, n = 6 independent experiments, **P < 0.01 (two-tailed Student’s t test).

Deposited TTR Fibrils and Tetramers Induce Inflammatory Gene Expression by RNA-Seq and Protein Cytokine Screen

To understand how TTR fibrils might influence cardiac fibroblast gene expression, RNA-Seq transcriptomic analysis was performed on primary cardiac fibroblasts after 48 h of culture on untreated or TTR-fibril coated glass substrates. Comparison of normalized counts per million values between groups showed that 69 genes were differentially expressed, with 51 upregulated and 18 downregulated (Fig. 5A). Functional enrichment analysis of these differentially expressed genes through g:Profiler revealed over-represented terms and pathways, largely pertaining to innate immune response (Supplemental Fig. S3A). This was further seen when examining the top terms representative of the most differentially expressed genes in databases for Gene Ontology: Biological Processes (Fig. 5B), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (Fig. 5C), Gene Ontology: Molecular Function (Supplemental Fig. S3B), and reactome pathways (Supplemental Fig. S3C). Across all databases, cytokine and chemokine signaling terms were among the most over-represented, indicative of innate immune cell recruitment and migration, which was also seen to be highly represented. Cataloging KEGG pathways specifically, IL-17 signaling, TNF signaling, and NF-κB signaling pathways were enriched alongside other disease pathways associated with strong proinflammatory responses. See Supplemental data for normalized Log₂ counts per million of differentially expressed genes (Supplemental Fig. S4), differential analysis (Supplemental Table S5) and Ingenuity Pathway Analysis (Supplemental Table S6).

Figure 5.

Transcriptomic and cytokine protein analyses suggests that transthyretin (TTR) fibrils induce an inflammatory response in fibroblasts. A: volcano plot reveals differentially expressed genes in primary cardiac fibroblasts when cultured on glass coated with TTR fibrils compared with fibronectin control, with genes of interest labeled. B: enriched Gene Ontology: Biological Processes (GO:BP) terms for differentially expressed genes, selected by largest fraction of genes associated with each term. C: enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathways for differentially expressed genes selected by largest fraction associated with each pathway. [n = 3 biological replicates per condition, normalized by trimmed-mean of M-values scaling in edgeR, false discovery rate (FDR)-adjusted P < 0.05]. D: cytokine profile blots showing increased secretion of cytokines when cardiac fibroblasts are cultured on TTR fibrils or tetramers. E: relative pixel density to untreated (UT) for all cytokines with greater than fivefold increased expression relative to UT controls. No cytokines were more highly expressed on UT relative to either TTR fibrils or tetramers. Bars are relative mean pixel density. Pooled cell culture medium supernatants from n = 3 biologically independent experiments. Bars are mean relative pixel density compared to UT for each cytokine, adjusted for background signal.

To confirm that the RNA expression profile was translated into functional protein, culture medium supernatant was collected and a cytokine antibody screen was used to assess the secreted cytokine profile. Cytokine secretion was increased on both TTR species relative to the fibronectin-only control (Fig. 5D). Multiple cytokines were secreted at greater than fivefold higher amounts on either TTR species when normalized to the untreated pixel density, with similar or slightly lower expression seen on TTR tetramers compared with TTR fibrils (Fig. 5E).

DISCUSSION

This is the first report of altered structure, function, and inflammatory gene expression of primary cardiac fibroblasts cultured with TTR on various substrates. Fibroblasts cultured on TTR-coated surfaces have a disorganized actin cytoskeleton, decreased focal adhesion number, and a flattened nuclear shape. Both proliferation and migration velocity of fibroblasts on flat, glass TTR-coated surfaces are significantly increased. Furthermore, fibroblasts have reduced attachment to microtopographic substrates with concomitant increased proliferation. Transcriptomic and proteomic analysis of cytokines for fibroblasts grown on TTR-coated surfaces showed significant upregulation of inflammatory genes after 48 h. These basic cell behaviors exhibited in vitro may play an underlying role in the maladaptive diseases that occur when amyloidogenic TTR is deposited in the aging human connective tissue of the heart or peripheral nerves.

Both nanotopographical and chemical cues are known to influence a variety of cell behaviors in multiple cell types. For example, cell proliferation and gene expression of fibroblasts from different tissues exhibit varied levels of response to nanofibrils (23), and the fibril diameter can directly influence cell spreading and gene expression of corneal fibroblasts (24). For TTR fibrils in particular, a previous study attempted to decouple the chemical and topographical effects on epithelial cells, demonstrating that both cues play significant roles in regulating cell behavior (25). The system explored in this work, using the nanoscale features of TTR fibrils and tetramers and microscale post architecture, suggests a role for both micro- and nanoarchitecture and the chemical properties of deposited TTR in affecting observed changes in cardiac fibroblast structure, function, and gene expression. TTR fibril formation is a dynamic process that involves exchange of unfolded oligomers with actively forming fibrils suggesting there is a mixture of various forms of TTR monomers and fibrils (13), therefore there will always be some contribution from monomers and tetramers in vivo, and in this culture system.

Previous in vitro studies of TTR fibril effects have shown that extracellular TTR aggregates interact with the RAGE receptor and lead to F-actin remodeling in endothelial cells (26) to induce toxicity in neurons (27). In cardiomyocytes, various TTR species have been shown to interact with the plasma membrane and induce cytotoxicity and cellular dysfunction (10, 28, 29). Fibroblasts were found to endocytose and degrade deposited TTR fibril aggregates in a cell line in vitro, and also in a skin model in vivo (30). Primary cardiac fibroblasts, a major resident cell type in the myocardium that regulates cardiac fibrosis remodeling and inflammatory signaling (31), have not previously been studied in vitro.

Cells in tissues or confluent cultures have controlled rates of proliferation and migration. Cell anchorage, proliferation, and migration are intimately connected and essential for fibroblasts and all normal cells; there are serious consequences when these functions become disrupted in cancer cells (32). Proliferation rate is low in well-anchored cells but is higher in the migrating population where the cell stalls, divides, and daughter cells continue to migrate. In the presence of TTR nanofibrils, the number of focal adhesions is reduced, the actin cytoskeleton is reorganized, and cells migrate at a higher velocity, permitting further cell division. Both mechanical and topographical cues act through mechanotransduction signaling pathways to form focal adhesions (33). Much is known about these pathways that sense mechanical microenvironmental cues to internal biochemical cellular signals (14, 34). At the micron scale, fibroblasts anchored to microposts had blunted migration and cell proliferation of fibroblasts, while similarly shaped unanchored objects did not (16). Our group has previously shown that fibroblasts, stem cells, and cardiac myocytes in culture anchor firmly to microposts by increased focal adhesion formation (15, 21, 35). Herein, TTR-coated microposts lose their ability to form these strong focal adhesions so that fibroblasts migrate away and proliferate. In adult human connective tissue, the consequence of loose binding, more migration, and the subsequent cell division could result in increased fibrosis. In the heart, activation of resident cardiac fibroblasts causes excessive collagen deposition and stiffening of the myocardium (36), which then alters cardiomyocyte mechanosignaling pathways leading to whole heart dysfunction (37).

Fibroblast nuclei have the same volume on all culture substrates, but their two-dimensional surface area is significantly larger when grown on TTR. This suggests the nuclear shape is flattened presumably by the altered cytoskeleton in these cells. Much is known about the forces exerted by the cytoskeleton on the two uniformly separated lipid nuclear membranes, which have unique mechanical properties due to their distinct proteins, including lamins and lamin-associated proteins (38). Deformation of nuclei by external force has a direct effect on the genome, which is sensitive to the force transduced by the nuclear lamina governing chromatin remodeling (39). It is possible that this nuclear deformation caused by growing fibroblasts on TTR has some role in the fibroblast phenotype and gene expression. Together, these measurements suggest that the cytoskeletal and nuclear deformation are due to sensing of the amyloid nanofibril deposition of the altered underlying microenvironment. Note that cardiac fibroblasts grown on stiff surfaces already express α-smooth muscle actin (α-SMA) (40). In this study, no α-SMA gene expression changes were found, likely due to both the control and TTR groups being grown on stiff glass substrates for 48 h before RNA isolation.

The effects of TTR on primary cardiac fibroblasts were examined by RNA-Seq for transcriptional changes and confirmed for cytokine expression by an antibody screen. Together, these analyses point to an unexplored role for cardiac fibroblasts in wtATTR disease progression, coordinating the immune response following exposure to TTR. The most upregulated gene was ACOD1, a negative regulator of inflammation (41) and found expressed under proinflammatory conditions (42). Upregulation of inflammatory genes has been positively correlated with disease severity in patients with familial amyloid polyneuropathy (FAP) (1). Furthermore, genes associated with ECM remodeling in MMP-9 and LCN2 in patients with FAP had increased expression (8), which could relate to clinical findings that the serum levels of MMP-9 are elevated in diseased patients in the context of cardiac amyloidosis and subsequent cardiomyopathy (43). Thus, the heightened immune responses and transcriptional changes observed in models of FAP models are similar to those seen in primary cardiac fibroblasts in response to deposited TTR fibrils in this study.

The cytokine screen confirmed that fibroblasts express CXCL1, 2, 3, and 5 as well as CCL3, 5, and 20 at fivefold levels higher levels than without TTR confirming an innate inflammatory response in the presence of TTR tetramers or fibrils. CXCL1, 2 are major factors for recruiting and inducing inflammation in neutrophils (44, 45), and also play a key role in inflammasome activation (46). MMP12 was also upregulated and is known to play a cardioprotective role in resolving inflammation following cardiac injury (47). Collectively, this transcriptomic and proteomic analyses strongly point to a new role for primary cardiac fibroblasts in augmenting inflammatory cascades while decreasing anti-inflammatory regulators in response to TTR, suggesting that they may be important contributors and regulators to inflammatory cascades that may worsen cardiac wtATTR progression.

In further investigating transcriptomic data to identify a potential mechanism by which TTR affects cytoskeletal arrangement, downregulation in gene expression of CAP2 was noted. CAP2 is a regulator of the actin cytoskeleton and interestingly, CAP2 knockouts have resulted in sarcomere disarray, dilated cardiomyopathy, and increased fibrosis (48). Furthermore, mice with CAP2 deficiencies have also been shown to have higher risk of conduction delays as well as increased myocardial fibrosis (49). Overall, CAP2 downregulation may point to an interesting pathway by which TTR further contributes to cardiac dysregulation in wtATTR.

Summary

Cardiac fibroblasts are an abundant cell type in the myocardium and the key mediators of extracellular matrix homeostasis and pathological fibrosis in the adult heart. This work identifies how deposited TTR dysregulates neonatal cardiac fibroblasts and their ability to maintain homeostasis. Fibroblasts cultured on deposited TTR had disorganized cytoskeletal and nuclear structure, with fewer focal adhesions and decreased adhesion to micropost substrates. Functionally, TTR deposition induced increased rates of proliferation and migration, and transcriptional sequencing and cytokine proteomics revealed upregulation of inflammatory genes linked to progression of other TTR-based diseases. This may suggest that cardiac fibroblasts, when exposed to deposited TTR in the underlying ECM, transition to a state that may enhance inflammation and subsequent fibrosis. As cardiac fibrosis is a common occurrence in wtATTR, these discoveries may lead to further understanding of fibrosis progression and may help guide future research into therapeutic targets for this widespread cardiac disease.

SUPPLEMENTAL DATA

Supplemental Videos S1 and S2, Supplemental Tables S5 and S6, and Supplemental Figs. S3 and S4: https://doi.org/10.5281/zenodo.4732410.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant HL 62426 (to B. Russell and T. A. Desai), American Heart Association Predoctoral Fellowship 18PRE3403027 (to P. Mohindra), and a National Science Foundation (NSF) Graduate Research Fellowship 1650113 (to J. X. Zhong). Bioinformatics analysis in the project described was performed by the UIC Research Informatics Core, supported in part by National Center for Advancing Translational Sciences (NCATS) through Grant UL1TR002003. The use of JPK Instruments, AFM, and supporting equipment at SF State University are gratefully acknowledged and were supported under NSF MRI Award 1626611.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.T.D. and B.R. conceived and designed research; K.T.D., A.I., and P.M. performed experiments; K.T.D., A.I., J.X.Z., P.M., and B.R. analyzed data; K.T.D., A.I., J.X.Z., P.M., and B.R. interpreted results of experiments; K.T.D. and J.X.Z. prepared figures; K.T.D., J.X.Z., and B.R. drafted manuscript; K.T.D., A.I., J.X.Z., P.M., T.A.D., and B.R. edited and revised manuscript; K.T.D., A.I., J.X.Z., P.M., T.A.D., and B.R. approved final version of manuscript.