Myocardium From Patients With ATTR Amyloidosis Produces Less Force Secondary to Increased Fibrosis

Visual Abstract

Highlights

• •Cardiac amyloidosis is an infiltrative cardiomyopathy that results in wall thickening and diastolic dysfunction.
• •Contractile assays, histology, and biochemical assays were performed on myocardial samples from patients with ATTR amyloidosis and organ donors without heart failure.
• •Amyloidosis myocardium displayed decreased maximum force with an increase in calcium sensitivity.
• •Phosphorylation of troponin I and myosin binding protein-C was decreased in amyloidosis myocardium.
• •A larger proportion of total myocardial stiffness was attributable to the extracellular matrix in amyloidosis myocardium.
• •Amyloidosis myocardium had more microcalcifications than nonfailing myocardium.

Summary

Amyloid transthyretin cardiac amyloidosis is one of the most common infiltrative cardiomyopathies. Contractile, biochemical, and histological assays were performed on myocardium from patients with and without amyloid transthyretin amyloidosis. Force was reduced in amyloidosis, but calcium sensitivity was increased. The change in calcium sensitivity may reflect dephosphorylation of troponin I. The proportion of stiffness attributable to the extracellular matrix was larger in amyloidosis. Septal fibrosis and amyloid burden correlated with measurements from LV samples. Technetium pyrophosphate scans may detect increased microcalcifications in amyloidosis myocardium. Replacement of myocytes with extracellular matrix is the most important factor depressing contractile force in amyloidosis myocardium.

Cardiac amyloidosis is an infiltrative cardiomyopathy wherein proteins are abnormally deposited in the myocardium. One of the common forms is amyloid transthyretin (ATTR) amyloidosis which can be age related or secondary to sequence variants in the TTR gene. Normally, the transthyretin protein circulates as a tetramer; however, in ATTR amyloidosis transthyretin dissociates into monomers which form fibrils. These fibrils deposit in the heart and lead to increased fibrosis, ventricular remodeling, diastolic dysfunction, and systolic dysfunction.^1,2 Traditionally, cardiac amyloidosis has been thought of as a rare disease, but this view may be caused by underdiagnosis.³ Increased awareness of the disease and utilization of noninvasive technetium-99m pyrophosphate (PYP) scans has shown it may be more common than thought, particularly in older patients.^4,5

Treatment of cardiac amyloidosis with traditional heart failure medications has limited efficacy.⁶ Pharmaceuticals that modify the pathogenesis of ATTR by stabilizing the TTR tetramer,⁷ decreasing synthesis of TTR,^8,9 and increasing clearance of TTR fibrils in tissues¹⁰ have shown potential for improving cardiovascular outcomes. As the prevalence of ATTR amyloidosis increases caused by increased screening and an aging population, improvements in the diagnosis and management of patients with ATTR amyloidosis will require additional clinical and preclinical studies.

Research on cardiac amyloidosis has primarily focused on clinical or pharmacological studies. In contrast to hypertrophic, dilated, and ischemic cardiomyopathies, which have extensive literature on the molecular drivers of disease, investigation of cardiac amyloidosis at the molecular level has been minimal. To better describe the pathophysiology of cardiac amyloidosis at the molecular level, this study conducted biochemical and biophysical assays on myocardial samples from patients with ATTR amyloidosis. Because amyloidosis results in diastolic dysfunction and is thought to stiffen the myocardium,11, 12, 13 this study measured tissue-level stiffness and molecular regulators of myocardial stiffness such as titin, tubulin, and collagen.

Additionally, few studies have investigated whether cardiomyocytes and sarcomeres remodel with amyloidosis. To address this, the intracellular impacts of amyloidosis on the sarcomere were examined by measuring sarcomere mechanical properties and phosphorylation of sarcomere regulatory proteins. This provides a unique opportunity to integrate biochemical assays, histology, and muscle mechanics to examine how the myocardium of patients with ATTR amyloidosis is impacted at multiple scales.

Methods

Patient samples

Hearts from 10 patients, 5 diagnosed with ATTR amyloidosis and 5 organ donors without cardiovascular disease, were collected from the operating room by a member of the research team and transported on ice to the laboratory where samples were taken from the midwall of the right ventricle (RV), septum, and left ventricle (LV). Each sample was flash frozen in liquid nitrogen and stored in the vapor phase of liquid nitrogen until needed for experimentation. Total time from explanation from the patient to freezing the sample was approximately 30 minutes. Previous studies have shown that this method of sample preparation retains many of the contractile and biochemical properties of the myocardium.^14,15 A more detailed description of the collection protocol has been published.¹⁶ All procedures were approved by the University of Kentucky Institutional Review Board (IRB#46103), and the subjects or their legally authorized representative gave written informed consent.

Myocardial strip preparation

Multicellular myocardial strips of LV midmyocardium were obtained by mechanical disruption followed by chemical permeabilization with 1% v/v Triton, as described.¹⁷ Myocardial strips were attached between a force transducer (model 403, Aurora Scientific) and motor (model 312B, Aurora Scientific) as described.¹⁸ The strip’s sarcomere length was adjusted to 2.13 ± 0.02 μm in pCa 9.0 solution ($\text{pCa}=-{\log }_{10}\left[\mathrm{C}{\mathrm{a}}^{2+}\right]$). Muscle length and cross-sectional area were measured. To correct for subtle variations in strip dimensions, force measurements and length changes were normalized to cross-sectional area and muscle length (L₀) respectively. Myocardial strip dimensions did not differ between groups (Supplemental Figure 1).

Contractile function measurements

For every trial, myocardial strips were moved from pCa 9.0 to preactivation solution for 2 minutes. The strip was then moved into a solution with a defined pCa value until it reached a steady state force, at which time a series of length changes were imposed. The strip was then returned to pCa 9.0 and allowed to relax. This cycle was repeated in a random order for pCa values ranging from 9.0 to 4.5 to evaluate the force-pCa relationship (Supplemental Figure 2). All contractile assays were performed at 37 °C.

Force-pCa curves were generated from each strip by fitting a four-parameter Hill equation:

$$\mathrm{F}={\mathrm{F}}_{\text{pas}}+{\mathrm{F}}_{\text{act}}\frac{{\left[\mathrm{C}{\mathrm{a}}^{2+}\right]}^{{\mathrm{n}}_{\mathrm{H}}}}{{\left[\mathrm{C}{\mathrm{a}}^{2+}\right]}^{{\mathrm{n}}_{\mathrm{H}}}+{\left[\mathrm{C}{\mathrm{a}}_{50}^{2+}\right]}^{{\mathrm{n}}_{\mathrm{H}}}}$$

where ${\mathrm{F}}_{\text{pas}}$ is the passive force, ${\mathrm{F}}_{\text{act}}$ is the maximum active force, ${\mathrm{n}}_{\mathrm{H}}$ is the Hill coefficient, and $\mathrm{C}{\mathrm{a}}_{50}^{2+}$ is the concentration of free calcium required to generate half-maximum force. The rate of force recovery, $k_{\text{tr}}$, was measured by fitting a single exponential function of the form F = A (1 − B × e^{(−ktr × t)}) to the force record after a rapid shortening/restretch protocol (20% L₀, 20-ms duration).¹⁹ Myocardial strips were excluded from the study if they displayed >30% rundown, calculated as the percent difference in maximum force from the beginning to end of the experiment.

Stretch response analysis

To examine additional tissue properties and myosin kinetics, step stretches (1% muscle length) were imposed at both maximal (pCa 4.5) and minimal activation (pCa 9.0). The active component of the force response measured in pCa 4.5 solution was obtained by subtracting the pCa 9.0 response from the experimental record. As shown in Supplemental Figure 3 and prior publications,20, 21, 22 the active component exhibits an initial steep increase in force to a peak (P₁) that reflects the strain imposed on bound cross-bridges. Then force decays to a nadir (P₂). The decay phase can be fitted with a single exponential equation, F = A × e^{−krel × t}, where A is the amplitude of P₁ and k_rel is the rate at which force decays to P₂. Force then redevelops to a new steady state force (P₃) at a rate k_df, which is typically calculated as the linear transformation of the half-time of force redevelopment using the formula: k_df = −ln(0.5) × (t_1/2)⁻¹ where ${\mathrm{t}}_{1/2}$ is the time after the nadir to redevelop one-half of the new steady state force.^23,24

Intracellular and extracellular passive stiffness

To examine the stiffness of the myocardium, myocardial strips were passively stretched 10% of their initial length over 1 second, held for 500 ms, and then returned to the initial length. The strip was then placed in high salt relax solutions (0.6 mol/L KCl and 1.0 mol/L KI) for 30 minutes each to depolymerize the myofilaments and remove the intracellular contributions to passive stiffness.²⁵ The myocardial strip was then restretched using the same protocol. The initial stretch reflects the total stiffness of the myocardium and the second stretch reflects the properties of the extracellular matrix (ECM).²⁶ From these stretches, stiffness was calculated as the slope of force plotted against length. To account for differences in myocardial strip dimensions, Young’s modulus was calculated from the stiffness and strip dimensions for the total and extracellular components as:

$$\mathrm{E}=\frac{\raisebox{1ex}{$\Delta \mathrm{F}$}\!\left/ \!\raisebox{-1ex}{$\mathrm{A}$}\right.}{\raisebox{1ex}{$\Delta \mathrm{L}$}\!\left/ \!\raisebox{-1ex}{${\mathrm{L}}_0$}\right.}$$

where $E$ is Young’s modulus, $\Delta \mathrm{F}$ is the change in force, $\mathrm{A}$ is the cross-sectional area of the strip, $\Delta \mathrm{L}$ is the change in length of the strip with stretch, and ${\mathrm{L}}_0$ is the initial strip length.

Histology

Tissue samples were trimmed over dry ice into rough cubes, placed into plastic molds filled with optimal cutting temperature compound, and frozen by submerging in liquid nitrogen cooled isopentane. These tissue blocks were then cryosectioned at a thickness of 10 μm in triplicate and used for picrosirius red, Congo red, and Von Kossa staining.

Picrosirius red is commonly used to stain collagen in tissues and measure cardiac fibrosis.^27,28 Sections from the LV, RV, and septum were and stained using picrosirius red as previously described.²⁹

Congo red stains amyloid deposits and is used clinically to diagnose cardiac amyloidosis.^30,31 Sections from the LV, RV, and septum were stained with Congo red using a commercially available kit (Abcam: ab150663) according to manufacturer specifications.

Last, Von Kossa stain was used to visualize calcium deposits within the myocardium. Sections from the LV were stained using a commercially available kit (Sigma Aldrich: 1003620001) according to manufacturer specifications.

Image acquisition and analysis

Tissue sections were imaged on a Zeiss Axioscan Z1 slide scanner. Picrosirius red and Congo red stained images were analyzed with a k-means clustering algorithm. This segments an image by clustering its pixels based on similarity of color and allows for algorithmic quantification (Supplemental Figures 4 and 5). A Calinski-Harabasz index was used to unbiasedly evaluate the optimal number of clusters for each image. Relative fibrosis and amyloid were calculated as the ratio of fibrosis or amyloid stained tissue to the total tissue area (myocardium plus amyloid/fibrosis). Slides stained with Von Kossa were segmented by thresholding the saturation-channel of the image in HSV color-space using Otsu’s method.³² These segmented Von Kossa images were then overlaid with serially sectioned Congo red images and aligned using intensity-based image registration to observe whether microcalcifications aligned with amyloid deposits. These image segmentation techniques and quantification were performed using custom-written routines in MATLAB (MathWorks).

Titin gels

Frozen tissue samples were pulverized in a liquid-nitrogen cooled Dounce homogenizer, solubilized in sample buffer, and run on a 1% SDS-agarose gel as previously described.³³ Gels were stained with Oriole (BioRad: 161-0496) per manufacturer instructions and imaged on a BioRad ChemiDoc. Images were analyzed using GelBox, software developed by our laboratory, which fits Gaussian curves to the optical density profiles of protein bands to quantify protein abundance.³⁴ The N2BA isoform percentage was calculated as the optical density of the N2BA band divided by the sum of the optical densities of the N2BA and N2B bands.

Immunoblots

Frozen tissue samples were homogenized in a bead blender (MP FastPrep-24), solubilized in a 4-mol/L urea sample buffer, and run on an 8% SDS-acrylamide gel. Gels were transferred to polyvinylidene difluoride (PVDF) membranes, blocked for 60 minutes at room temperature, and incubated in primary antibodies overnight at 4 °C. Membranes were then incubated in secondary antibody (1:10,000, Invitrogen, 31460) for 60 minutes at room temperature, washed, developed in SignalFire ECL (Cell Signaling: 6883S), and then imaged using a BioRad ChemiDoc. Membranes were then stripped and reprobed.

Myosin binding protein-C (MyBPC) phosphorylation was measured using phosphospecific antibodies for MyBPC at serines 273, 282, 302 (1:10,000, ProSci Inc) normalized to total MyBPC (1:5,000, Santa Cruz Biotechnology, sc-137237) as previously described.³⁵ Regulatory light chain (RLC) phosphorylation was measured using a phosphospecific antibody for RLC at serine 15 (1:1,000, Affinity BioSciences, AF8618) normalized to total RLC (1:5,000, Abcam, ab92721). Tubulin abundance was measured by probing tubulin (1:5,000, Abcam, ab4074) and normalizing to actin using trichloroethanol.^36,37 Band densities were measured as previously described using GelBox software.

Phosphate affinity gels were used to measure troponin I (TnI) phosphorylation. This method is similar to traditional immunoblotting but with the addition of Phos-tag (FujiFilm) in the gel, which allows for the separation of proteins by both size and phosphorylation status.³⁸ Detailed protocols of this method have been described.^39,40 These gels were transferred as previously described and probed for TnI (1:5000, HyTest, 4T21).

Statistical analysis

Data were analyzed using linear mixed models in SAS version 9.4 (SAS Institute Inc). Linear mixed models used disease and/or region as fixed effects with the patient’s deidentified code as a random effect and assumed a compound symmetry covariance structure. Normality of data was determined using Shapiro-Wilk tests in SAS. Tukey-Kramer post hoc tests were used for additional pairwise comparisons. This statistical approach allows for repeated measurements and accounts for samples coming from different regions of the same hearts. The data are presented in superplots with the mean measurement for each patient represented as an opaque symbol and repeated measures as transparent symbols.⁴¹ Demographics and echocardiogram measurements of patients were analyzed by unpaired Student's t-tests in MATLAB (MathWorks). Linear regression plots comparing septal and LV measurements and curve fitting were completed with MATLAB. Analysis of continuous variables by linear regression were presented with slope P value, 95% CIs, and Pearson’s correlation coefficients (r). P values <0.05 were considered statistically significant. Demographics and echocardiogram measurements are presented as the mean ± SD, and mixed model results are presented as the mean ± SEM.

Results

Patient characteristics

This study used samples collected from 5 patients with ATTR amyloidosis receiving a heart transplant and 5 nonfailing donors. Efforts were made to age and sex match these patient groups (age ± 7 years and same sex). Age and body mass index were not significantly different between groups (Table 1). Patients in the amyloidosis cohort had increased interventricular septal thickness and LV posterior wall thickness compared with control subjects (Table 2). The patients with amyloidosis had reduced ejection fraction (EF) without significant changes in LV internal diameter at end-systole or end-diastole (Table 2).

Table 1.

Demographics of Patients With Amyloidosis and Nonfailing Donors

Patient ID		Sex	Race	Age, y	BMI, kg/m2	ATTR Variant	Cause of Death
Amyloid
	5CCF6	Male	White	68.2	24.9	wtATTR	—
	AFAA5	Male	Black	67.8	29.2	wtATTR	—
	0C492	Male	White	72.4	34.7	wtATTR	—
	2B20A	Male	Black	72.2	28.2	wtATTR	—
	0D377	Male	Black	59.8	22.7	His76Arg	—
Mean ± SD				68.1 ± 5.1	27.9 ± 4.6
Nonfailing
	30B2B	Male	White	71.2	24.2	—	Anoxic brain injury
	BE497	Male	Black	56.1	36.1	—	Anoxic brain injury
	778BB	Male	White	64.4	23.0	—	Brain aneurysm
	3066B	Male	White	63.3	23.4	—	Anoxic drug intoxication
	15B03	Male	White	61.2	30	—	Stroke
Mean ± SD				63.2 ± 5.5	27.3 ± 5.7

A brief overview of the demographics of patients with amyloidosis (top) and nonfailing organ donors (bottom). Symbols by each patient correspond to the symbols in subsequent figures. Data were analyzed with unpaired Student's t-tests.

Arg = arginine; ATTR = amyloid transthyretin; BMI = body mass index; His = histidine; wtATTR = wildtype amyloid transthyretin.

Table 2.

Patient Echocardiography Measurements

Patient ID		Ejection Fraction, %	IVSd, cm	LVIDd, cm	LVIDs, cm	LVPWd, cm	ESV, mL	EDV, mL
Amyloid
	5CCF6	40	2.0	3.3	2.6	2.2	25.2	45.3
	AFAA5	24	1.5	4.9	4.5	1.4	65.9	82.1
	0C492	20	1.1	5.1	5	1.1	118	148
	2B20A	21	2.5	4.4	4.8	2.2	56	71
	0D377	26	1.4	4.8	4.1	1.4	118	164
Mean ± SD		26.2 ± 8.1a	1.7 ± 0.6b	4.5 ± 0.7	4.2 ± 1.0	1.7 ± 0.5b	76.6 ± 40.6	102.1 ± 51.3
Nonfailing
	30B2B	55	1.1	4.0	3.2	1.2	49.2	—
	BE497	30	1.4	4.6	3.8	1.4	60.7	—
	778BB	45	1.0	4.4	3.6	1.0	—	—
	3066B	52	1.1	4.2	3.4	1.2	—	—
	15B03	55	0.8	4.2	3.2	1.0	111.0	—
Mean ± SD		47.4 ± 10.5	1.1 ± 0.2	4.3 ± 0.2	3.4 ± 0.3	1.2 ± 0.2	73.6 ± 32.9	-

A brief overview of the echocardiogram measurements of patients with amyloidosis (top) and nonfailing organ donors (bottom). Symbols by each patient correspond to the symbols in subsequent figures. Data were analyzed with unpaired Student's t-tests.

IVSd = Interventricular septal thickness; LVIDd = left ventricular internal diameter at end diastole; LVIDs = left ventricular internal diameter at end systole; LVPWd = left ventricular posterior wall thickness; ESV = end systolic volume; EDV = end diastolic volume.

^aP < 0.01 compared with nonfailing donors.

^bP < 0.05.

Amyloidosis myocardium produces less force with subtle changes in cross-bridge kinetics

Myocardial preparations from patients with amyloidosis exhibited significantly decreased maximum and minimum force per cross-sectional area compared with nonfailing myocardium (Figures 1C and 1D, respectively). Additionally, there was a modest but statistically significant increase in calcium sensitivity (Figure 1E) in the amyloidosis myocardium. Cooperativity (n_H) was not significantly different between groups. Further analysis of step stretch response displays showed the amyloid group had a slower k_rel (Figure 2B) and faster k_df (Figure 2D) compared with nonfailing control subjects. No significant differences were measured in the other stretch response parameters or k_tr.

Figure 1.

Amyloidosis Myocardium Exhibits Decreased Maximum and Minimum Force

(A) Absolute and (B) relative force as a function of Ca²⁺ concentration. Superplots show (C) maximum force per cross-sectional area, (D) minimum force per cross-sectional area, (E) calcium sensitivity (pCa₅₀), and (F) Hill coefficient. Symbols correspond to those shown in Table 1 and identify each patient. Data were analyzed using linear mixed models. ∗P < 0.05, ∗∗∗P < 0.001.

Figure 2.

Amyloidosis Myocardium Displays Largely Unchanged Cross-Bridge Kinetics

Parameters obtained from the force transient in response to step-length change or ktr maneuver. Superplots show (A) P₁, (C) P₂, and (E) P₃ normalized to P0 and (B) k_rel, (D) k_df, and (F) k_tr. Symbols correspond to those shown in Table 1 and identify each patient. Data were analyzed using linear mixed models. ∗P < 0.05.

Phosphorylation of MyBPC and TnI is decreased in amyloidosis myocardium

MyBPC phosphorylation was decreased at serine 273 and 282 in amyloidosis myocardium (Figures 3B and 3C, respectively). There was no difference in the phosphorylation of MyBPC at serine 302. TnI phosphorylation was decreased in amyloidosis myocardium (Figure 4C), but there was no change in RLC phosphorylation (Figure 4D).

Figure 3.

MyBP-C Phosphorylation Is Decreased in Amyloidosis Myocardium

(A) Immunoblots of total and phosphorylated myosin binding protein-C (MyBP-C) at serine 273, 282, and 302. Superplots comparing phosphorylation of MyBP-C at serine (B) 273, (C) 282, and (D) 302, relative to total MyBP-C in amyloidosis and nonfailing myocardium. Symbols correspond to those shown in Table 1 and identify each patient. Data were analyzed using linear mixed models. ∗∗P < 0.01, ∗∗∗P < 0.001.

Figure 4.

TnI Phosphorylation Is Decreased in Amyloidosis

(A) Representative troponin I (TnI) Phos-tag gel showing dephosphorylated (P₀), monophosphorylated (P₁), and bisphosphorylated (P₂) TnI. (B) Representative immunoblots for phosphorylated regulatory light chain (pRLC) at serine 15 and total RLC. (C) Plot comparing the degree of phosphorylation in moles of phosphate per mole of TnI in amyloidosis and nonfailing myocardium. (D) Plot comparing relative phosphorylation of RLC at serine 15 in amyloidosis and nonfailing myocardium. Symbols correspond to those shown in Table 1 and identify each patient. Data were analyzed using linear mixed models. ∗∗∗P < 0.001.

ECM-based stiffness is increased in amyloidosis myocardium

Total passive stiffness was significantly higher in nonfailing myocardium (Figure 5B). However, the percent of total myocardial stiffness attributable to the ECM was significantly higher in the patients with amyloidosis (Figure 5C).

Figure 5.

Total Stiffness Is Unchanged In Amyloidosis Myocardium, But a Larger Proportion Is Attributable to the Extracellular Matrix

(A) Representative force-length curves showing increasing force with passive lengthening of the muscle. Total force was measured by stretching the permeabilized myocardial strip. Extracellular matrix (ECM)–based force was measured after the myofilaments had been depolymerized with high-salt solutions. Superplots show (B) a comparison of total and extracellular Young’s Moduli between amyloidosis and nonfailing myocardium and (C) the proportion of total Young’s Modulus attributable to the ECM in amyloid and nonfailing myocardium. Symbols correspond to those shown in Table 1 and identify each patient. Data were analyzed using linear mixed models. ∗∗P < 0.01, ∗∗∗P < 0.001.

N2BA titin is increased in amyloidosis myocardium

The proportion of titin expressed as the N2BA isoform was increased in amyloidosis compared with nonfailing myocardium (Figure 6B). Tubulin abundance was unchanged in amyloidosis myocardium (Figure 6D).

Figure 6.

The Proportion of Titin N2BA Is Increased in Amyloidosis Myocardium

(A) Representative titin gel with individual lanes showing control and myocardium from amyloidosis and nonfailing hearts. (B) Representative immunoblot of tubulin and actin loading control. (C) Plot comparing the proportion of N2BA relative to total titin in amyloidosis and nonfailing myocardium. (D) Plot comparing the tubulin abundance relative to actin in amyloidosis and nonfailing myocardium. Symbols correspond to those shown in Table 1 and identify each patient. Data were analyzed using linear mixed models. ∗P < 0.05. LV = left ventricle; RV = right ventricle; Sept = Septum.

Amyloidosis myocardium has increased fibrosis and similar amyloid deposition in the septum and LV

Relative fibrosis was increased in amyloidosis myocardium compared with nonfailing myocardium (Figure 7B). Every tested region of the heart (RV, septum, LV) had increased fibrosis in the hearts from patients with amyloidosis compared with nonfailing patients. There was no significant difference in amyloid burden between the RV, septum, and LV (Figure 8B). Myocardium from the nonfailing patients did not show Congo red staining for amyloidosis when viewed under brightfield or polarized light. Septal measurements of fibrosis and amyloid burden correlated well with the corresponding measurements in the LV, and these measurements were not significantly different from a 1:1 relationship (Figures 7C and 8C, respectively).

Figure 7.

Fibrosis Is Increased in Amyloidosis Myocardium

(A) Representative picrosirius red staining and programmatic image analysis of myocardium from patients with and without amyloidosis. (B) Superplot comparing RV, Sept, and LV fibrosis from patients with and without amyloidosis. Data were analyzed using linear mixed models. (C) Deming linear regression of each patient’s average left ventricle fibrosis plotted against their average septal fibrosis. This analysis allows for the experimental uncertainty associated with both the x and y variables. The p-value shows the probability of the measured slope differing from 1. The identity line is shown in red. Symbols in both B and C correspond to those shown in Table 1 and identify each patient. ∗∗P < 0.01, ∗∗∗P < 0.001. Abbreviations as in Figure 6.

Figure 8.

Amyloid Deposition Is Homogeneous in the Ventricles and Septum of Patients With Amyloidosis

(A) Representative Congo-Red staining of myocardium from patients with and without amyloidosis. No amyloid deposition was observed nor programmatically detected in any of the nonfailing patients. (B) Superplot comparing amyloid deposition in RV, Sept, and LV from amyloidosis myocardium. Data were analyzed using linear mixed models. (C) Deming linear regression of each patient’s average LV fibrosis plotted against their average septal fibrosis. This analysis allows for the experimental uncertainty associated with both the x and y variables. The P value shows the probability of the measured slope differing from 1. The identity line is shown in red. Symbols in both B and C correspond to those shown in Table 1 and identify each patient. Abbreviations as in Figure 6.

Microcalcifications are increased in the myocardium of patients with amyloidosis

Quantification of calcium deposits with Von Kossa staining showed significantly more microcalcification in ATTR amyloidosis myocardium than in nonfailing myocardium (Figure 9D). Automated image registration was used to align serially sectioned myocardium from patients with amyloidosis stained with Von Kossa and Congo red, which qualitatively show the microcalcifications occupy similar positions as the amyloid deposits (Figure 9B).

Figure 9.

Increased Microcalcifications in Amyloid Myocardium

(A) Representative serial sections from amyloidosis myocardium stained with Congo red and Von Kossa and analyzed using image segmentation algorithms. (B) Automated image registration of Von Kossa and Congo Red stained sections with microcalcifications (labeled green) and ATTR amyloid deposits (labeled blue). (C) Representative Von Kossa staining and image analysis of nonfailing myocardium shows minimal microcalcifications. (D) Plot comparing microcalcifications in amyloidosis and nonfailing myocardium. Symbols correspond to those shown in Table 1 and identify each patient. Data were analyzed using linear mixed models. ∗∗∗P < 0.001.

Discussion

ATTR amyloidosis is one of the more common infiltrative cardiomyopathies and is becoming more frequently diagnosed because of new imaging techniques and increased physician awareness. Clinical and pharmacological studies have helped develop new pharmaceuticals, which may be beneficial in treating the disease, but there has been little investigation at the tissue and sarcomere level.42, 43, 44 To our knowledge, this study is one of the first to use biophysical and biochemical experiments to determine how ATTR amyloidosis affects myocardium and sarcomere-level properties.

Patient characteristics

Amyloidosis has a higher prevalence in men of advanced age. Consistent with this, the patients with amyloidosis included in this study were all men over the age of 55 years. To minimize potential confounding effects of sex and age, nonfailing control subjects were age and sex matched. Examination of included patient echocardiograms showed that the amyloidosis cohort had significantly thicker interventricular septal thickness and LV posterior wall thickness (Table 2) with no changes in LV internal diameter at end-systole or -diastole, similar to prior studies.^45,46 Taken together, the demographic and echocardiographic characteristics of the patients used in this study are comparable to prior studies on ATTR amyloidosis.

Decreased force production in ATTR myocardium is likely caused by decreased contractile tissue per area

To our knowledge no study has examined the biophysical properties of ATTR amyloidosis myocardium at the tissue level. Our multicellular contractile assays showed the myocardium from patients with amyloidosis produces significantly less force than nonfailing myocardium (Figure 1C). On average, amyloidosis myocardium produced 66% less force than nonfailing myocardium. In contrast, similar experiments on human ischemic and nonischemic cardiomyopathy have shown force decreases by approximately 30% compared with nonfailing myocardium.^17,47 These different reductions in force are likely caused by increased ECM expansion in amyloidosis. Imaging studies have shown that patients with cardiac amyloidosis have increased extracellular volume fraction compared with other forms of heart disease.48, 49, 50 Additional studies have shown that increased extracellular volume fraction correlates with increased measurements of tissue fibrosis, amyloid deposition, and ECM expansion.^51,52 Consistent with these studies, our histological data showed decreased contractile tissue per area in amyloidosis myocardium secondary to significantly increased fibrosis (Figure 7B) and amyloid (Figure 8) deposition. The relationship between contractile tissue per area and contractile force is further supported by an inversely proportional relationship between tissue fibrosis and maximum contractile force (r = 0.92; P < 0.001) (Supplemental Figure 6). These contractile and histological assays suggest the decreased force production observed in amyloidosis myocardium is caused by decreased contractile tissue per unit area secondary to fibrosis and amyloidosis deposition.

Altered calcium sensitivity in amyloidosis myocardium may reflect decreased sarcomeric protein phosphorylation

In addition to the decreased force production, amyloidosis myocardium displayed a subtle but significant increase in calcium sensitivity (Figure 1E). This may be a compensatory effect to increase force at calcium concentrations characteristic of normal systole. This increase in the calcium sensitivity of amyloidosis myocardium may be driven by decreased phosphorylation of TnI (Figure 4). Dephosphorylation of TnI increases calcium sensitivity by stabilizing the Ca²⁺ bound state of troponin C (TnC) and decreasing Ca²⁺ dissociation.^53,54 At the organ level, this has been shown to slow myocardial relaxation.^55,56 It is possible that decreased TnI phosphorylation in amyloidosis myocardium may contribute to impaired relaxation and diastolic dysfunction.

In addition to decreased TnI, amyloidosis myocardium had significant dephosphorylation of MyBPC (Figure 3). MyBPC can modulate the activity of myosin by stabilizing the head in a super relaxed state, which cannot directly participate in cross-bridge cycling.⁵⁷ Dephosphorylation has been shown to disrupt these interactions and increase the proportion of myosin heads in the super relaxed state which may lead to depressed force.⁵⁸ The impact on MyBPC dephosphorylation on calcium sensitivity has not been fully resolved. Some studies have shown no change in calcium sensitivity,59, 60, 61, 62 while others show an increase in calcium sensitivity.63, 64, 65 Although these studies report variable impacts of MyBPC phosphorylation on contractile function, it is possible that MyBPC dephosphorylation plays a role in the altered contractile function in amyloidosis myocardium.

Interestingly, in amyloidosis myocardium we only measured decreased phosphorylation at sites targeted primarily by protein kinase A (PKA). There were no significant differences in the phosphorylation of MyBPC at serine 302 or RLC at serine 15 (Figures 3D and 4D). RLC is phosphorylated by myosin light chain kinase and MyBPC serine 302 is phosphorylated by PKA, PKD, and PKCɛ.^66,67 The role of β-adrenergic stimulation and subsequent PKA activation in heart failure is well established. Previous studies on hypertrophic cardiomyopathy, dilated cardiomyopathy, and ischemic heart failure have shown similar decreases in phosphorylation of TnI and MyBPC as we measured in amyloidosis.68, 69, 70 This may suggest that altered phosphorylation at PKA sites is conserved across various forms of heart failure and is independent of etiology. Although PKA plays a prominent role in sarcomere phosphorylation, recent studies show additional kinases may also contribute to altered phosphorylation in the setting of cardiac dysfunction. For example, kinases such as PKG, PKC, and ribosomal S6 kinase II (RSK2) have been shown to phosphorylate MyBPC, and their activity is altered in different disease states.^66,71,72 These protein phosphorylation measurements illustrate that decreased phosphorylation of sarcomeric proteins contribute in part to altered contractile function and may play a role alongside ECM expansion in the pathogenesis of cardiac amyloidosis.

Myosin cross-bridge kinetics are largely unchanged in amyloidosis myocardium

Although amyloidosis myocardium displayed altered calcium sensitivity, cross-bridge kinetics measured from stretch activation or k_tr protocols were largely unchanged. Amyloidosis myocardium only produced significantly slower k_rel and faster k_df (Figures 2B and 2D). The k_rel parameter is thought to be influenced by the detachment of strained cross-bridges and k_df may reflect the cooperative recruitment of myosin heads into a strongly bound state.^73,74 However, given the lack of changes in other kinetics parameters, such as k_tr, any change in cross-bridge attachment and detachment would have to be subtle. Alternatively, these changes may be caused by altered viscoelastic properties of the ECM within amyloidosis myocardium, which become apparent when quickly stretching the muscle.

Amyloidosis myocardium is not stiffer than nonfailing myocardium, but a larger proportion of stiffness is attributable to the ECM

Myocardial tissue stiffness was measured by passively (pCa 9.0) stretching muscle strips before and after decellularization to calculate the total and ECM-based Young’s moduli. The restrictive cardiomyopathy observed in cardiac amyloidosis is often attributed to increased myocardial stiffness and ventricle wall thickness, but the data showed that amyloidosis myocardium was significantly less stiff than nonfailing myocardium (Figure 5B). This is likely caused by expansion of the ECM, which decreases cardiomyocytes per unit area and therefore contributions of titin to total passive stiffness. Potentially compounding this effect, amyloidosis myocardium had a significantly increased proportion of the less stiff titin isoform, N2BA, which would suggest more compliant cardiomyocytes (Figure 6). This has been observed in other forms of heart disease and may be a compensatory effect of increased fibrosis and ECM based stiffness.^75,76

Although the total stiffness was decreased in amyloidosis myocardium, the proportion of ECM-based to total Young’s modulus was significantly increased in amyloidosis myocardium (Figure 5C). This suggests a decrease in the contribution of intracellular sources of stiffness relative to overall stiffness, likely caused by increased fibrosis (Supplemental Figure 7). The relative contribution of the ECM to total passive stiffness increases with sarcomere length.^25,77 Therefore, the relative contribution of collagen to myocardial stiffness likely varies during the cardiac cycle. Experiments on canine and rodent myocardium show in vivo sarcomere lengths ranging from 1.6 to 2.4 in systole and diastole, respectively.78, 79, 80 This indicates that our passive stretches from sarcomere lengths of 2.1 to 2.3 likely reflect end-diastolic stiffness.

Importantly, Young’s modulus quantifies the stiffness of a material after correcting for the preparation’s dimensions. Although amyloidosis myocardium had a significantly smaller total Young’s modulus, these patients have significantly thicker ventricular walls (Table 2). Therefore, increased wall thickness (perhaps driven by ECM deposition) (Supplemental Figure 8) may be the main cause of the restrictive filling observed in patients with amyloidosis rather than changes in the intrinsic stiffness of the myocardium.

Septal myocardial samples accurately predict LV tissue histology

Cardiac amyloidosis can be diagnosed by Congo red staining of septal biopsies collected during a right heart catheterization. Whether these biopsies accurately reflect the histology of the RV and LV free wall has not been well established. Here, we showed significant increases in fibrosis in the myocardium of patients with amyloidosis (Figure 7B), consistent with prior studies and new imaging techniques.^51,81, 82, 83 Fibrosis was relatively constant across the RV, septum, and LV in both amyloid and nonfailing tissue (Figure 7B). Similarly, amyloid burden was consistent and did not depend on the region sampled in hearts with amyloidosis (Figure 8B). This would imply that at least end-stage amyloidosis is a relatively homogenous disease that affects the ventricle walls evenly. In a direct comparison, septal fibrosis and amyloid burden correlated strongly with measurements from the LV and shared a 1:1 relationship (Figures 7C and 8C, respectively). These data suggest that assessment of septal myocardium, as is typically done with endomyocardial biopsies, is a valid predictor of LV fibrosis and amyloid burden in patients with ATTR amyloidosis.

Microcalcifications are increased in amyloidosis myocardium

PYP scans are an increasingly popular method to screen patients for cardiac amyloidosis and potentially improve on biopsy-based approaches. PYP scans were originally used during bone scans, which has led some to hypothesize that increased calcium in the myocardium was the mechanism behind this scan’s sensitivity for ATTR amyloidosis.^84,85 Here we show that the myocardium from patients with ATTR amyloidosis has significantly increased microcalcifications compared to nonfailing, non-amyloidosis myocardium (Figure 9D). Previous studies have shown similar microcalcifications in ATTR myocardium, and to a lesser extent, light chain (AL) amyloidosis, but did not examine nonamyloidosis myocardium.⁸⁶ The current data extends these findings by quantifying both amyloid and nonamyloid containing myocardium with robust image segmentation techniques and show a significant increase in microcalcifications within amyloidosis myocardium. Overlaying serial sections of amyloidosis myocardium stained with Von Kossa and Congo red suggest that these microcalcifications and amyloid deposits may be close spatially (Figure 9B). This was qualitatively observed in serially sectioned slices but robust colocalization studies could be performed in future work to provide a more quantitative test. These data illustrate that the increased microcalcifications may be responsible for the increased uptake of PYP in hearts of patients with ATTR amyloidosis.

Study limitations

Cardiac amyloidosis is highly variable in its presentation and can span heart failure phenotypes from heart failure with reduced to preserved EF. The patients included in this study primarily had reduced EF without LV dilation compared with nonfailing control subjects. As a result, it is possible that the findings we report here may not extend to other presentations of cardiac amyloidosis such as those with heart failure with preserved EF–like phenotypes. Future studies could collect endomyocardial biopsies from patients with different amyloidosis phenotypes to determine whether these findings are unique to this amyloidosis phenotype. As with many studies on amyloidosis myocardium, access to myocardial samples remains a challenge. Our study was limited to 5 patients with primarily wild-type, age-related ATTR amyloidosis. This makes generalizing the findings of this study to heredity ATTR or AL amyloidosis challenging. Additionally, some imaging studies have shown apical sparing in cardiac amyloidosis, but because this study used samples taken from the midwall of the RV and LV, it cannot inform on potential apical to basal changes.⁸⁷

Conclusions

This study shows that, compared with nonfailing tissue, ATTR myocardium develops less contractile force per unit area likely secondary to decreased contractile tissue per unit. Additionally, amyloidosis myocardium has increased calcium sensitivity likely driven by TnI dephosphorylation. This suggests both intracellular and extracellular contributions to the contractile dysfunction in cardiac amyloidosis myocardium. The total stiffness of ATTR myocardium is decreased compared with age-matched control subjects, but the proportion attributed to the ECM is greater. The restrictive filling observed in vivo is therefore likely driven by changes in ventricular wall thickness rather than an increase in the intrinsic stiffness of the myocardium. Improved diagnostic techniques will be vital as awareness and screening of cardiac amyloidosis increases. These data emphasize that septal biopsies of fibrosis and amyloid burden are good predictors of LV histology and that increased microcalcifications in ATTR myocardium may explain the utility of PYP tracers in the diagnosis of ATTR amyloidosis.

Perspectives.

COMPETENCY IN MEDICAL KNOWLEDGE: This study showed that intracellular and extracellular changes in the myocardium of patients with ATTR amyloidosis may contribute to disease. Amyloidosis myocardium produced reduced contractile force with subtle changes in the calcium handling of the myofilaments and cross-bridge kinetics. This suggests that the changes in contractile function were driven by expansion of the ECM via amyloid and collagen deposition and decreased phosphorylation of sarcomeric proteins. Additionally, total stiffness of amyloidosis myocardium was decreased compared with nonfailing myocardium, but the proportion attributable to the ECM increased in amyloidosis myocardium. Therefore, increased wall thickness driven by ECM expansion may play a larger role in restrictive filling and diastolic dysfunction observed in cardiac amyloidosis than changes to the myocardium’s intrinsic stiffness. Additionally, the utility of technetium pyrophosphate scans in diagnosing ATTR amyloidosis may be caused by increased microcalcifications within the myocardium of patients with ATTR amyloidosis.

TRANSLATIONAL OUTLOOK: Additional studies should examine the impacts of disease-modifying drugs on amyloidosis myocardial function at the tissue level. Although these disease-modifying drugs primarily act on extracellular amyloid fibrils, additional studies should examine whether they promote intracellular remodeling. Understanding what pharmaceutical options improve or restore sarcomere and tissue-level properties may contribute to the development of better treatments or management strategies for patients.

Funding Support and Author Disclosures

This study was supported by the National Institutes of Health (HL149164 to Dr Campbell) and (1F31HL170558 to Mr Wellette-Hunsucker) and the American Heart Association (24PRE1191551 to Mr Milburn). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.