Material Properties of Aged Human Mitral Valve Leaflets

Abstract

Objective

To characterize the mechanical properties of aged human anterior (AML) and posterior (PML) mitral leaflets.

Materials and Methods

The AML and PML samples from explanted human hearts (n = 21, mean age of 82.62 ± 8.77 years old) were subjected to planar biaxial mechanical tests. The material stiffness, extensibility and degree of anisotropy of the leaflet samples were quantified. The microstructure of the samples was assessed through histology.

Results

Both the AML and PML samples exhibited a nonlinear and anisotropic behavior with the circumferential direction being stiffer than the radial direction. The AML samples were significantly stiffer than the PML samples in both directions, suggesting that they should be modeled with separate sets of material properties in computational studies. Histological analysis indicated the changes in the tissue elastic constituents, including the fragmented and disorganized elastin network, the presence of fibrosis and proteoglycan/glycosaminoglycan infiltration and calcification, suggesting possible valvular degenerative characteristics in the aged human leaflet samples. Overall, stiffness increased and areal strain decreased with calcification severity. In addition, leaflet tissues from hypertensive individuals also exhibited a higher stiffness and low areal strain than normotensive individuals.

Conclusion

There are significant differences in the mechanical properties of the two human mitral valve leaflets from this advanced age group. The morphologic changes in the tissue composition and structure also infer the structural and functional difference between aged human valves and those of animals.

INTRODUCTION

Mitral valve (MV) disease is one of the most common valvular lesions. At least moderate mitral regurgitation (MR) occurred at a frequency of 1.7% in the US adult population [1]. The prevalence of MV disease has been shown to be strongly associated with age of the patients, particularly in mitral regurgitation patients, increasing from 0.5% in patients aged 18 to 44 years to 9.3% in patients aged ≥75 years [1, 2]. Clearly, MV disease mainly affects the aged population.

It is well known that aging has a substantial impact on soft tissue mechanical properties. Studies have shown that the stiffness of aortic vessels is much higher in aged patients compared with that of young patients [3-7]. However, our knowledge on the mechanical properties of mitral valve tissues in aged patients has been very limited. Quantification of native tissue properties has many implications in the development of medical treatment and prosthetic device, and the understanding of physiological function and disease progression. For example, mitral valve repair, benefited from the improved understanding of mitral valve biomechanics, has now been acknowledged as desirable and superior to mitral valve replacement in virtually all pathologies of mitral disease [8].

Our current knowledge of mitral valve mechanical properties is mainly derived from that of porcine mitral valves [9-17]. The assumption that porcine MV is similar to that of human ones is frequently adopted, even though it has not been rigorously validated. Indeed, there are numerous computational models of human mitral valve that had utilized porcine MV tissue properties [18-24]. Moreover, for prosthetic valve device approval, the Food and Drug Administration (FDA) mandates pre-clinical animal trials, using either porcine or ovine models, to demonstrate sufficient safety including performance and handling, as well as to study the efficacy of new valve devices [25]. While many disparate results between human and animal trials have been observed, the current assumption taken for heart valve device trials (such as for transcatheter valve intervention [26-31]) - porcine and ovine animal models are similar to those of aged humans – is still prevailing.

In this study, the mechanical properties of aged human MV tissues, both mitral anterior (AML) and posterior (PML) leaflets, were characterized using the planar biaxial testing method. The MV tissues were obtained from individuals above 65 years old with no prior history of valvular disease. The material stiffness, extensibility and degree of anisotropy of the leaflet samples were quantified. Histological analysis was performed to examine tissue microstructural properties. The presence of abnormal valvular structures such as calcific deposits and disrupted fiber structure was assessed. Finally, the material properties of aged human mitral leaflets were compared with those of animal mitral leaflets published in the literature.

MATERIALS AND METHODS

Materials and sample preparation

A total of 21 hearts from 82.62 ± 8.77 years old human cadavers were obtained from the National Disease Research Interchange (NDRI, Philadelphia, PA). An approval was obtained from the Institutional Review Board at the University of Connecticut for human tissue research in this study. The selection criteria for the hearts were no history of valvular disease or any valvular repair or replacement. See Table 1 for characteristics and medical history of all patients. All hearts were fresh frozen within a post-mortem recovery of less than 24 hours and kept frozen during delivery. Upon arrival, the hearts were defrosted in a hot water bath, and the MV leaflets were dissected out. A total of 20 AML and 18 PML specimens were removed from the hearts and used in this study. All chordae tendineae were carefully cut off from the leaflets. The MV leaflets that could not be tested immediately were cryopreserved in 10% DMSO [32] and stored in the −80°C freezer until testing. It has been showed that connective tissues can preserve their structural architecture [33] and mechanical properties [32, 34, 35] when store in a cryoprotectant agent at low temperature (−80°C). The cryopreserved specimens were thawed using a four-step process [32] to remove cryoprotectant agent prior to testing. Specimen thickness was measured and averaged at three locations in the belly region and two in the free-edge region (Fig. 1) using a thickness gauge (Mitutoyo, Model 7301).

Table 1.

Patients’ medical history.

Specimen	Age	Gender	Smoking	Diabetes	Heart Weight (g)	Cause of death	Heart related disease(s)	Calcification grade
								AML	PML
1	69	M	Unknown	Unknown	733	CPA secondary to HA	HTN	1	1
2	88	F	n	Unknown	342	Alzheimer	CAD	2	no data
3	78	M	Unknown	Unknown	396	CPA secondary to HA	HBP, HA	1	0
4	79	F	y*	Unknown	545	Respiratory failure	CHF	2	1
5	81	F	n	y	330	Unkonwn	Stroke	2	1
6	96	F	y	n	545	CPA	None	2	1
7	95	F	y*	n	330	Respiratory arrest	None	2	1
8	79	F	n	Unknown	404	CPA	HTN	1	1
9	98	F	n	y	303	Respiratory arrest	None	1	1
10	82	M	y*	y	714	COPD	HTN, CHF, CAD	2	1
11	75	M	y*	n	425	HA	None	1	1
12	63	M	y*	n	369	Unknown	HTN	2	0
13	75	F	y*	n	575	CPA	HTN	2	1
14	82	M	y*	n	510	CPA	HTN	2	1
15	83	M	y	n	412	Vascular Dementia	HTN, CHF	2	no data
16	87	F	y*	n	295	Unknown	None	2	no data
17	80	F	y*	n	397	Unknown	None	1	1
18	82	M	y	n	544	Aspiration Pneumonia	None	3	3
19	95	M	y	n	425	Natural Causes (Age)	None	2	1
20	87	F	n	n	490	Unknown	None	1	1
21	81	F	y*	n	621	Alzheimer's related	None	1	0
Mean	82.00	F:10	--	--	464.00	--	--	--	--
SD	8.51	M:7	--	--	130.86	--	--	--	--

COPD - Chronic obstructive pulmonary disorder, HTN - Hypertension, HBP – High blood pressure, CAD - coronary artery disease, CHF - Congestive heart failure, CPA - chronic pulmonary aspergillosis, HA - heart attack.

^*Smoking = quit more than 5 years prior to death.

Calcification 1 = focal, calcification 2 = densely distribution over the belly region, calcification 3 = complete calcification

Figure 1.

a) An image of one excised mitral valve from a cadaver heart. The testing regions are located in the dash square boxes, and the above and below regions of the dashed curved lines on each leaflet are belly and edge, respectively; b) five regions of thickness measurements; and c) mounting position and alignment of the leaflet specimen.

Biaxial mechanical test

Details on planar biaxial mechanical testing techniques and methods of analysis can be found in Sacks and Sun 2003 [36]. Briefly, all specimens were mounted in a trampoline fashion by continuous strands of silk suture looped along each of the four specimen sides, with the circumferential (CIRC) and radial (RAD) directions aligned to the loading axes of the biaxial test fixtures. In each specimen side, four barbless fish hooks (size 22) were punctured through the tissue thickness and spaced out evenly near the leaflet edge (Fig. 1-c). All specimens were tested in a phosphate buffered saline solution bath at 37°C. The four markers were affixed to the lower-belly region of the tissue above the line demarcating the edge and the belly regions (Fig. 1-b). This region was chosen due to its relatively homogenous fiber structure where collagen fibers are aligned to the circumferential direction. A stress-controlled test protocol was utilized [37], wherein the ratio of the normal Lagrangian stress components T₁₁ : T₂₂ was kept constant with T₁₂ = T₂₁ = 0. Preconditioning was performed to minimize tissue hysteresis. Each tissue specimen was preconditioned for at least 40 continuous cycles with T₁₁ : T₂₂ = 1 : 1. The seven successive protocols were performed using the ratios T₁₁ : T₂₂ = 1 : 0.3, 1 : 0.5, 1 : 0.75, 1 : 1, 0.75 : 1, 0.5 : 1, and 0.3 : 1. The unloaded reference markers were obtained after preconditioning cycles and used to analyze all the consecutive protocols.

Biomechanical analysis

The mechanical properties of valve leaflets were compared by means of secant modulus at three equibiaxial membrane tension levels: 1) at 10 N/m for tissue responses under small load, 2) at 60 N/m to produce responses at physiological level [38], and 3) at high load of 120 N/m for high stress state. From the equibiaxial protocols, the areal strain, e, was also obtained , e = λ_θθ* λ_zz – 1, where λ_θθ and λ_zz are the CIRC and RAD stretch values at the three equibiaxial membrane tension levels, respectively. In addition, the degree of anisotropy (DA) was calculated as the ratio of CIRC strain to RAD strain under equibiaxial stretching. With the value of DA approaching 1, the tissue response is considered isotropic, while others represent various degrees of anisotropy. The correlations between the mechanical property data and clinical variables were also examined.

Microstructural analysis

The microscopic anatomy of representative AML and PML tissues used in mechanical testing were examined. Specimens were fixed in 10% formalin for 24 hours prior to histological process and analysis. After dehydrated and embedded in paraffin, the specimens were sectioned into 5 μm thick sections. Tissues were stained with Verhoeff Van-Gieson (VVG) for elastic fiber (as black), collagen fiber (as pink) and muscle cell (as dark brown/gray) visualization. Crimping of collagen fibers were also assessed using Picro-sirius red stain. The calcified deposits on the leaflets were identified with von-Kossa stain. Histological images were obtained from an Olympus U-TVO.5xC digital camera coupled to an Olympus BX40 light microscope.

Constitutive Modeling

The valve leaflet materials are assumed to be homogeneous and incompressible, and exhibited nonlinear, hyperelastic and anisotropic responses. The Fung-type model was utilized to model the biaxial mechanical responses of these valve samples. The in-plane 2 ^nd Piola-Kirchhoff stresses, S_ij, were derived from a strain energy function (SEF), W,

$${\mathrm{S}}_{\mathrm{ij}}=\frac{\partial \mathrm{W}}{\partial {\mathrm{E}}_{\mathrm{ij}}}$$ (1)

where E_ij is the Green-Lagrange strain tensor and W is expressed as:

$$\mathrm{W}=\frac{\mathrm{c}}{2}({\mathrm{e}}^{\mathrm{Q}}-1)$$ (2)

$$\mathrm{Q}={\mathrm{A}}_1{\mathrm{E}}_{11}^2+{\mathrm{A}}_2{\mathrm{E}}_{22}^2+2{\mathrm{A}}_3{\mathrm{E}}_{11}{\mathrm{E}}_{22}+{\mathrm{A}}_4{\mathrm{E}}_{12}^2+2{\mathrm{A}}_5{\mathrm{E}}_{11}{\mathrm{E}}_{12}+2{\mathrm{A}}_6{\mathrm{E}}_{22}{\mathrm{E}}_{12}$$ (3)

with c and A_1-6 are material parameters. The constitutive model was fit to the seven stress controlled protocols for each individual specimen to obtain the material parameters. The goodness of the fit was determined by a correlation coefficient, R-square value, based on the nonlinear regression Levenberg-Marquardt algorithm in SYSTAT 10 (Systat Software Inc., Chicago, IL).

Statistical analysis

Data were tested for normality prior to statistical tests. Statistical differences between the AML and PML variables were determined using the Student's t-test. A paired-sample t-test was performed to compare the directional responses. Non-parametric tests, including the Wilcoxon signed-rank test (for paired data) and Mann-Whitney rank sum test, were used for non-normal distributed sample groups. Correlations between the parameters and clinical variables were determined using the Pearson's (for normal) and Spearman (for non-parametric data) correlation coefficient (r), and p-values were calculated based on the aforementioned statistical tests. A p-value less than 0.05 was considered statistical significant, with high significance indicated by p < 0.001. Statistical analyses were performed using SigmaPlot (V11.0, Systat Software Inc., San Jose, CA) and SYSTAT 10 (Systat Software Inc., Chicago, IL). All measurements were presented as mean ± standard deviation (SD).

RESULTS

Valve leaflet thickness and degree of calcification

The general observations of the leaflet samples were: thickening of leaflets, calcification and chordae tendineae fusion. As showed in Fig. 2, calcified regions on the leaflets were classified into four groups: 1) no calcification, defined as CALC 0, 2) focal or spotted dense mass of calcium deposition, defined as CALC 1, 2) dense and uniform calcium deposition over a large continuous region, defined as CALC 2, and 3) completely calcified and thickened leaflet, defined as CALC 3. From the 20 AML samples, 35% (n = 7) had CALC 1, 60% (n = 12) had CALC 2, and 0.05% (n = 1) had a complete calcification or CALC 3. Therefore, all AML samples had some degree of calcification. The CALC 2 or a uniform and continuous distribution of calcification in the AML samples were observed mainly at the aortomitral curtain or the anterior annular region that is anatomically coupled to the aortic annulus [39], and focal calcified spots were observed at the regions of chordal insertions. There was no uniform calcification in 18 PML samples, 78% (n = 14) exhibited CAL 1, 0.06% (n = 1) with CAL 2 and 17% (n = 3) with no calcification. Similar to the AML samples, the focal calcification in PML samples were mostly found at the annular and chordal insertion regions.

Figure 2.

The anterior and b) posterior mitral valve leaflets showing calcification (CALC) degrees: (1) - focal or spotted dense mass of calcium deposit, (2) - dense and uniform calcium deposition over large continuous regions, and (3) completely calcified and thickened leaflet.

One valve had AML and PML thickness of 4.28 mm and 4.0 mm, respectively, due to severe calcification, which was identified as outliers and removed from the subsequent data analysis. The mean thickness of AML and PML samples were then 1.74 ± 0.37 (n = 19) and 1.83 ± 0.34 (n = 17) mm, p = 0.51, respectively. Both the interquartile and overall ranges of both leaflet data sets are similar, see Fig. 3. The mean belly and free-edge thickness values are also examined. It can be seen in Fig. 3 that the belly (denoted as AB) regions were significantly thinner than the free-edge (denoted as AE) regions of AML samples, 1.62 ± 0.34 mm versus 2.00 ± 0.54 mm, p < 0.05. Fusion of the chords and calcification most likely contributed to a higher value in thickness in the free-edge regions of all AML samples. No difference between the free-edge (PE) and belly (PB) regions of the PML leaflets, 1.90 ± 0.44 mm versus 1.70 ± 0.39 mm, p = 0.09. No correlation between age and thickness was observed in both leaflet groups.

Figure 3.

Boxplots of the thickness of anterior mitral leaflet (AML, n = 19) and posterior mitral leaflet (PML, n = 17) samples (left side) and of thicknesses of belly and edge regions of AML and PML samples (right side). The black lines inside the boxes represent the median. The red dashes and the numbers represent the mean values. (*) indicates a statistical significant difference between two groups. AB – Anterior Belly, AE – Anterior Edge, PB – Posterior Belly, PE – Posterior Edge.

Biomechanical responses

In general, the stress-strain response curve of leaflet tissues exhibits an anisotropic behavior, with the CIRC direction being the stiffer one. This MV leaflet characteristic has been observed and reported in studies of porcine [9] and human [40, 41] mitral leaflets. However, in our sample set, several AML (n = 6) samples had the RAD direction being stiffer than the CIRC direction, see Fig. 4-b. This reverse directional response could be possibly due to the calcium deposits within the leaflet, which might dominate these leaflet responses. In our data analysis, we considered the samples with reverse directional response atypical for normal aged mitral leaflet tissues, and thus, they was excluded the subsequent data analysis. The mean equibiaxial membrane tension-strain curves for the AML (n = 13) and PML (n = 14) samples in the CIRC and RAD directions after removal of reverse response data at 60 N/m are illustrated in Fig. 4-a.

Figure 4.

a) Mean equibiaxial protocols of the anterior (AML) and posterior (PML) mitral leaflet samples in circumferential (CIRC) and radial (RAD) directions prior to data removal, and b) the AML (n = 6) samples with reverse directions .

The mean secant moduli of mitral leaflet samples were calculated (Fig. 5). The AML samples were significantly stiffer than the PML samples at the three membrane tension levels (10 and 60 N/m, p = 0.002, and 120 N/m, p = 0.006), see Fig. 6-a. Similarity in the mean areal strains at 60 and 120 N/m membrane tensions indicated that both AML and PML samples reached to a high stiffening region on a stress-strain curve at 60 N/m, where the collagen fibers may have already been straightened. Both leaflet groups were anisotropic with mean DA values of 0.46 ± 0.25 and 0.47 ± 0.26 at 10 N/m, respectively, and the values did not change significantly at higher membrane tension levels, 0.48 ± 0.18 and 0.48 ± 0.27 at 60 N/m and 0.53 ± 0.13 and 0.46 ± 0.20 at 120 N/m, respectively, as shown in Fig. 6-b.

Figure 5.

Comparing stiffness between the circumferential and radial directions (short bar) and between the AML and PML samples (long bar). (*) indicates p < 0.05 and (**) indicates p < 0.001.

Figure 6.

Comparison of a) the extensibility and b) the degree of anisotropy at 10, 60 and 120 N/m of the AML and PML samples. No significant variation in degree of anisotropy was found. (*) indicates statistical significant difference with p < 0.05.

Correlation with patients’ medical history

No correlation was observed between the mechanical properties and patients’ characteristic data (Table 1) of age, smoking, and diabetes. The leaflet samples were further divided into subgroups of hypertension (HTN) versus normotensive (NTN), and calcification levels, and compared between the groups. Of note, for calcification levels, we compared CALC 1 versus CALC 2 for the AML samples, and CALC 0 versus CALC 1 for the PML samples. Table 3 listed the stiffness, the areal strain and the DA values of each subgroup at 60 N/m membrane tension. An increase in tissue stiffness was observed when CALC 1 was increased to CALC 2 and when NTN was increased to HTN. All samples remained anisotropic in all disease states.

Table 3.

Differences in stiffness, areal strain and degree of anisotropy (DA) between tissue valve leaflets.

		n	SM-CIRC (MPa)	p	SM-RAD (MPa)	p	Areal strain	p	DA	p
AML	NTN	6	2.03 ± 1.51	0.30	0.77 ± 0.57	0.10	0.19 ± 0.19	0.07	0.42 ± 0.20	0.24
	HTN	7	2.48 ± 1.28		1.28 ± 0.64		0.08 ± 0.03		0.54 ±0.15
	CALC 1	7	1.96 ± 1.27	0.45	0.88 ± 0.55	0.35	0.17 ± 0.18	0.45	0.46 ± 0.18	0.65
	CALC 2	6	2.64 ± 1.46		1.23 ± 0.73		0.09 ± 0.03		0.51 ± 0.19
PML	NTN	8	1.22 ± 0.99	0.57	0.42 ± 0.20	0.30	0.25 ± 0.09	0.29	0.52 ± 0.32	0.63
	HTN	6	1.39 ± 0.81		0.56 ± 0.29		0.19 ± 0.10		0.44 ± 0.21
	CALC 0	2	0.94 ± 0.35	/	0.28 ± 0.09	/	0.29 ± 0.09	/	0.30 ± 0.01	/
	CALC 1	12	1.35 ± 0.95		0.51 ± 0.24		0.21 ± 0.09		0.52 ± 0.28

NTN – normotensive, HTN – hypertensive, CALC – calcification, SM – Secant Modulus, CIRC – circumferential, RAD – radial, DA – Degree of Anisotrophy, AML – Anterior Mitral Leaflet, PML – Posterior Mitral Leaflet.

Constitutive modeling

A representative seven-protocol stress-strain curves with the Fung-elastic model fits in each direction of a set of leaflet samples are illustrated in Fig. 7-a&b. The patient-specific material parameters obtained from fitting the multi-protocol experiment data to the seven-parameter Fung-elastic model of Eq. 2 for both leaflet groups are shown in Table 3. The model demonstrated an excellent fit with the mean R² values of about 0.93 for AML and 0.92 for PML.

Figure 7.

Representative experimental data and Fung's model fit of multiple biaxial data in the circumferential (CIRC) and radial (RAD) directions of the AML (top) and PML (bottom) samples.

Microstructural analysis

Figure 8 illustrates the Movat stain sections of the belly and free-edge regions of the representative AML and PML samples in both CIRC and RAD directions. Both AML and PML samples appeared to undergo the pathological changes: accumulation of fibrous components (arrows in Fig. 8), multiple replicated elastic lamellae in the atrialis layer and disorganized and fragmented elastin fibers (Fig. 8-b). In the belly region, infiltration of the massive amount of proteoglycan (PG) and glycosaminoglycan (GAG) (Fig. 8-a) together with disorientation of the collagen networks, appeared as marble-like structure as reported by Stephens et al. [42] in aged porcine mitral valve leaflets, were observed in both leaflet groups. In the free-edge region, collagen content was abundant in the fibrosa layer and also evident in spongiosa layer of the AML sample while the amount of PG/GAG was merely visible in the same region. Fibrin was also evident within the fibrosa layer and highly expressed in the belly regions in both AML and PML samples. In addition, the von Kossa stains showed the calcified deposits in the belly regions with a high distribution in the fibrosa layer of the AML sample, see Fig. 9. The calcification content decreased toward the free-edge region and disappeared near the belly and free-edge junction. For the PML samples, only small calcified spots were observed, which scattered in the spongiosa layer along the radial direction of the valve leaflet.

Figure 8.

Representative histological results of the AML and PML samples from one patient. The images on the top row show the cross sections through thickness in the circumferential (CIRC) direction, the one on the left is at the belly and the right is at the free-edge region. The bottom row displays images of sections cut in the radial direction. The pop-up image A on the bottom left shows the proteoglycan (PG) and glycosaminoglycan (GAG) infiltration in the belly region, and the pop-up image B on the right shows fragmented elastin fibers in the radial direction within the atrialis layer. The pop-up image C shows the reduction in undulation of collagen fibers, stained with Picro-sirius red stain. Red arrows indicate the accumulation of fibrous layers in the atrialis layers. (*) indicates the presence of fibrin (intense red color) in the fibrosa layers. Movat-pentachrome stains: collagen – yellow, elastin – black, PG/GAG – blue/greenish, red – muscle fibers, intense red – fibrinoid, fibrin. All histology slides were captured using a standard light microscope. Bar = 400 0m unless indicated otherwise.

Figure 9.

Histological results of calcification (dark brown color indicated by the arrows) in the belly and belly/freeedge regions in the radial direction of representative AML and PML samples. Dense and thick calcified deposits were observed in the spongiosa and fibrosa layers of the AML belly region and spotted calcification in the fibrosa of the belly/free-edge region. Calcification were scattered throughout the thickness and along the radial direction of the PML sample. Von Kossa stains. Bar is 400 ,m.

DISCUSSION

Tissue properties related to aging

Heart valve disease mainly affects senior patients. Therefore, a study of normal and diseased valve tissue properties in advance aged patients is more relevant than a study of younger age groups, and it could enhance our understanding of the heart valve disease progression and the development of clinical treatment techniques. In this study, we presented, to our knowledge, the first comprehensive biaxial characterization of aged (82.62 ± 8.77 years old) human MV leaflets using a relatively large sample size (n = 21). A comparative study of age-dependent valve property changes was conducted by Stephens et al. [42] using a porcine model. They compared the three age groups, i.e., 6 weeks, 6 months and 6 years old of porcine mitral and aortic valves and observed an elevated stiffness and a higher collagen fiber distribution with increasing age. Another animal study by Stephen et al. [43] showed that changes in valve PGs and GAGs with age are complex and distinct within valve type. In human mitral valve leaflets, the GAG content and many different PG subclasses were found to be significantly decreased with advancing age [44]. Other studies have also shown that age has a significant impact on the composition of soft tissues [45, 46], resulting in significant differences in both material and microstructural properties of human aortic tissues [47]. Interestingly, our results suggest that the mechanical properties of leaflet tissues remain unchanged for persons older than 65 years of age. A broad age range of 30 – 90 years could possibly reveal significant differences as previously reported by others [47].

Comparison with human MV tissue properties reported in the literature

An early study of human mitral leaflet mechanics performed in 1970s by Clark [48] showed the unidirectional non-linear stress-strain behavior of generalized leaflets, though with unknown leaflet types, patient background and medical history. Recently, the uniaxial data of mitral valve tissues from a 88 year-old patient in the study by Prot et al. [40] showed the CIRC and RAD maximum stretches were approximately 4% and 12% for the AML sample and 23% and 69% for the PML sample, respectively. Our mean maximum strains were similar to their data for the AML samples, 4% and 10% for CIRC and RAD directions, respectively, but our PML samples were much stiffer, with 6% and 17% for CIRC and RAD directions, respectively. Barber and coworkers [41] showed that normal posterior mitral valve leaflets could stretch up to 15.5 ± 6.2% and 24.1 ± 4.0% strain in the CIRC and RAD directions, respectively. They cut the tissues into strips, stretched until failure (with higher load range 200-400g) and reported the maximum tangent modulus as stiffness, which were 6.1 ± 1.6 kN/m (CIRC) and 6.4 ± 0.3 kN/m (RAD). Because of different testing methods (i.e., uniaxial versus biaxial tests), a direction comparison cannot be made across studies. However, in general, the mechanical behaviors of our samples seem to be stiffer than others. In our biaxial testing protocols, we performed rigorous pre-conditioning of at least 40 cycles and analyzed the experimental data using the post-preconditioning reference state. As shown in the studies by Martin and Sun [49] and Carew et al. [50], valve tissues exhibited significant residual strains upon unloading, even after 20 cycles of preconditioning. Therefore, in this study, the tissue response was evaluated after 40 cycles of preconditioning to obtain reproducible mechanical testing data. By applying rigorous preconditioning, the residual strain was removed when analyzed using the post-preconditioning state. Thus, the tissue responses obtained in this study were stiffer than ones analyzed using a less number of pre-conditioning cycles, as shown in the Figure 8 in Martin and Sun [49].

Valve microstructure

Even though the hearts collected from patients who were not diagnosed with valvular diseases, we observed that all AML samples and 88% of the PML samples exhibited a certain degree of calcification, some with severe calcification. Several retrospective studies of mitral valve calcification found that it is a degenerative process that commonly occurring in elderly [51, 52]. Woodring and West [53] found a prevalence of valvular calcification increased with increasing age, 36% of patients over 70 years and 75% over 80. Hence, the mean age of our samples was 83 which could explain the present of calcification in all of our valves. Mechanically, the samples with more calcified deposits had higher stiffness and lower areal strain comparing to ones with less calcification. Furthermore, calcification was mostly found on the annulus and chordae insertion points. Recently, a computational study of mitral leaflet dynamics has shown that the annulus and chordae insertion locations are associated with high stress concentration [54]. Thus, the high stress could be one of the factors contributing to tissue calcification.

Our histological data shows that the microstructure of our human MV leaflets was altered with evidences of massive calcified deposition, marbling of PG/GAGs, straightened collagen fibers, high collagen contents and fragmented elastin fibers. Similar observations were reported by Martin et al. [55] who compared porcine and human aortic valve leaflets and also found straighter collagen fibers that corresponding to higher stiffness in human tissues compared to porcine ones. The study by Pham and Sun [56] also found significantly stiffer mechanical properties of their human coronary sinus vessels, which exhibited a high collagen content in the intima layer, compared to those of porcine and ovine models.

We also examined the difference in mechanical properties of MV tissues between individuals with and without hypertension. Hypertension has been shown to alter the structure and function of valve leaflets to accommodate an increase in blood pressure and consequently correlates with valvular disease such as degenerative and calcification [57-60]. We found that MV leaflets from hypertensive hearts exhibited a higher stiffness and lower extensibility values compared to that of normotensive hearts. Yap et al. [61] studied the dynamic deformation of aortic valve during hypertensive loading conditions and showed that valves experienced an increased diastolic stretch with high pressure conditions. A long-term exposure to an increased tensile strain may lead to structural damages resulting in high stiffness and low extensibility.

Modeling of MV leaflets

Although the overall thickness of the AML samples was similar to that of PML samples, AML free-edge regions were thicker compared to the belly regions, whereas PML tissues had a relatively uniform thickness. Mechanically, both of the AML and PML exhibited a nonlinear and anisotropic behavior. However, the AML leaflets were stiffer than the PML leaflets in both circumferential and radial directions. Differences between the anterior and posterior leaflets found in this study may indicate that AML and PML might need to be modeled with separate sets of material properties in computational studies. Finite element modeling (FEM) has been proven useful in predicting the biomechanics of mitral valve at pathological [40, 62-64] conditions as well as the mechanical behaviors in the edge-to-edge surgical repair [65, 66 , 67] and chordal replacement [68] techniques. Up to now, the patient-specific material properties of heart valve have yet to be assessed in vivo. Thus, our ex vivo human biaxial material properties data and the constitutive models will be useful in computational studies of aged human mitral valve and provide more accurate prediction of valvular mechanical responses.

Difference in the mechanical properties between animal and human MV leaflet tissues

It is well known that some of bioprosthetic aortic valves are made from chemically-treated porcine aortic valve and the pre-clinical animal studies are required by the FDA. In addition, there are numerous studies in the literature that use either porcine or ovine valve tissue as a surrogate for human one to study MV tissue structure and mechanical properties. For example, planar biaxial mechanical data of porcine mitral valve leaflets by May-Newman and Yin [9] and Kunzelman and Cochran [10] were widely used in constitutive material modeling [18, 19] and simulation of human mitral valve dynamics at physiological condition [20-24]. However, compared to our data, their tissue responses were much compliant. For instance, May-Newman and Yin [9] reported stretches of 17% and 20% for AML and 23% and 24% for PML at 20 N/m in the CIRC and RAD directions, respectively. This suggests that human MV leaflets were much stiffer than porcine ones and that porcine valve models might not reflect the functionality of aged human mitral valves.

Limitations

Studies in the literature have shown the differences between fresh and preserved tissues [69-77]. Particularly, Clark [48] showed that freezing at −70 °C without any medium markedly altered the mechanical behavior of mitral valve tissues such that an increase in pre-transition modulus and a markedly increase in post-transition modulus in frozen samples compared with fresh ones. Due to the limited access to fresh human tissues, the obtained data from frozen tissue samples could be different from those of fresh tissues. The additional freezing method using cryoprotectant at a very cold temperature at our laboratory, however, has been shown to provide a minimal effect in the mechanical properties [32]. The structures of both anterior and posterior valve leaflets were not homogeneous, particularly at the free-edge region where the tissue was thicker and exhibited disorganized fiber microstructures throughout the layers. An assumption of tissue homogeneity in the central region of the leaflet samples might not represent the material properties of the entire leaflet area. Thus, regional heterogeneity in mechanical properties of leaflet needs to be further studied. In addition, the degree of calcification was classified based on structural observation, thus no quantitative correlation with the mechanical data was assessed. A more detailed quantitative study of valvular pathogenesis including changes in the ECM components as well as immunohistochemical analysis is needed to further characterize and correlate with the mechanical data. Another limitation is our small range of ages. A larger sample size including younger human mitral valves would facilitate comparison and correlation data with age and further study the mechanics of pathologic valves. For instance, the myxomatous degenerative disease of the mitral valve leaflet is a common cause of mitral regurgitation [78, 79]. Myxomatous disease occurs more frequently in younger age group [80], and mitral valve repair is a common treatment technique [81]. Thus, material property data obtained for a younger patient population with myxomatous disease would be more valuable in FE simulations for evaluating diseased valvular structure and the efficacy of various surgical procedures.

CONCLUSION

In this study, the mechanical properties of the human mitral valve leaflet tissues aged older than 65 years of age were characterized using a multiple planar biaxial testing method. The overall nonlinear anisotropic valve leaflet tissue properties were quantified. We found that for mitral leaflets, the radial direction was significantly more compliant than the circumferential direction. The anterior leaflets were significantly stiffer and less compliant than the posterior leaflets in both directions. From the histological results, the changes in the elastic components including the fragmented and disorganized elastin network, high collagen content and the presence of fibrosis and PGs/GAGs infiltration were observed, suggesting possible valvular degenerative characteristics in the valve samples. Additionally, calcification was observed in most of the leaflet samples. Overall, stiffness increased and areal strain decreased with calcification severity. Leaflet tissues from hypertensive hearts also exhibited a higher stiffness and low areal strain than normotensive hearts. This study showed that the mechanical properties of aged human mitral tissues might not be equivalent to that of animal tissues. The anterior and posterior leaflet mechanical data herein could be implemented into finite element models to study the structure and function of aged human mitral valve tissues.

Table 2.

Material parameters of mitral valve leaflets.

Specs	c	A1	A2	A3	A4	A5	A6	R2	c	A1	A2	A3	A4	A5	A6	R2
1	0.87	328.54	220.00	58.72	270.79	−26.74	15.92	0.98	0.71	145.59	65.70	42.81	104.35	−8.67	−10.32	0.96
3	0.85	354.72	294.61	115.29	13.94	0.59	4.96	0.98	-	-	-	-	-	-	-	-
4	1.55	371.86	328.02	37.72	339.71	−33.78	65.17	0.95	5.30	389.56	66.31	92.15	13.65	−1.57	−4.02	0.82
5	0.37	672.03	94.99	80.02	42.72	−7.12	−4.07	0.93	0.33	99.93	85.12	0.03	13.85	4.33	1.81	0.97
6	-	-	-	-	-	-	-	-	0.58	30.75	37.56	8.26	−28.74	5.34	−11.22	0.88
7	-	-	-	-	-	-	-	-	0.76	227.88	57.53	7.95	13.80	−31.37	17.13	0.88
8	0.26	292.65	250.12	112.10	66.92	7.42	28.11	0.88	0.73	147.93	54.29	−20.76	10.39	−2.65	−3.97	0.98
9	0.61	188.25	74.98	1.68	1.21	0.01	0.71	0.96	0.65	26.67	23.90	8.94	5.47	1.26	1.71	0.89
10	-	-	-	-	-	-	-	-	1.21	391.98	34.42	37.73	84.78	48.47	−5.05	0.79
11	-	-	-	-	-	-	-	-	0.28	326.87	62.51	50.23	94.13	0.61	2.53	0.95
12	-	-	-	-	-	-	-	-	0.46	83.72	19.45	11.85	1.39	−2.59	0.23	0.95
13	2.22	168.85	72.49	42.84	13.90	−8.62	1.89	0.98	0.71	327.54	163.94	35.08	34.74	5.47	7.39	0.95
14	0.98	974.11	221.35	−23.19	449.04	17.23	32.84	0.86	-	-	-	-	-	-	-	-
15	1.10	351.79	1017.67	344.84	178.34	75.48	−9.12	0.85	-	-	-	-	-	-	-	-
16	0.30	368.61	329.88	12.17	34.12	8.36	−3.04	0.93	-	-	-	-	-	-	-	-
17	0.13	56.48	11.50	3.77	8.61	−8.30	−1.57	0.98	0.84	7.80	246.06	24.88	15.57	−1.89	−7.91	0.87
20	3.01	484.73	452.11	251.34	60.38	−41.49	32.19	0.86	0.34	228.46	201.21	0.65	11.11	3.87	−17.88	0.97
21	0.71	460.83	159.80	1.11	183.35	60.34	51.44	0.96	0.22	223.13	57.99	27.18	39.44	−5.10	1.58	0.96
Mean	1.00	390.26	271.35	79.88	127.92	3.34	16.57	0.93	0.94	189.84	84.00	23.36	29.57	1.11	−2.00	0.92
SD	0.83	233.15	256.84	107.55	145.67	33.40	23.49	0.05	1.28	133.21	69.15	27.74	38.59	16.48	8.68	0.06