- Email: [email protected]
Preliminary investigation of a novel technique for the quantification of the ex vivo biomechanical properties of the vocal folds
Materials Science and Engineering C 45 (2014) 333–336
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Preliminary investigation of a novel technique for the quantification of the ex vivo biomechanical properties of the vocal folds Paulo G. Coelho a,⁎, Michael Sobieraj b, Nick Tovar a, Kenneth Andrews c, Benjamin Paul c, Nandini Govil c, Seema Jeswani c, Milan R. Amin c, Malvin N. Janal d, Ryan C. Branski c a
Biomaterials and Biomimetics, College of Dentistry, New York University, New York, NY, USA Orthopedic Surgery, Hospital for Joint Diseases, New York University, New York, NY, USA NYU Voice Center, Department of Otolaryngology-Head and Neck Surgery, New York University School of Medicine, New York, NY, USA d Departments of Public Health and Epidemiology, New York University, New York, NY, USA b c
a r t i c l e
i n f o
Article history: Received 27 October 2013 Received in revised form 21 July 2014 Accepted 29 August 2014 Available online 17 September 2014 Keywords: Vocal fold Mechanical testing Quasi-static Dynamic Rabbit
a b s t r a c t The human vocal fold is a complex structure made up of distinct layers that vary in cellular and extracellular matrix composition. Elucidating the mechanical properties of vocal fold tissues is critical for the study of both acoustics and biomechanics of voice production, and essential in the context of vocal fold injury and repair. Both quasistatic and dynamic behavior in the 10–300 Hz range was explored in this preliminary investigation. The resultant properties of the lamina propria were compared to that of the nearby thyroarytenoid muscle. Er, quantified via quasistatic testing of the lamina propria, was 609 ± 138 MPa and 758 ± 142 MPa in the muscle (p = 0.001). E′ of the lamina propria as determined by dynamic testing was 790 ± 526 MPa compared to 1061 ± 928 MPa in the muscle. Differences in E′ did not achieve statistical significance via linear mixed effect modeling between the tissue types (p = 0.95). In addition, frequency dependence was not significant (p = 0.18). © 2014 Elsevier B.V. All rights reserved.
1. Introduction The human vocal fold (VF) is a complex structure made up of distinct layers that vary in cellular and extracellular matrix density and composition. Elucidating the mechanical properties of VF tissues is critical to our increased understanding of both the acoustics and the biomechanics of voice production and essential in the context of the inherent tissue response to injury and repair. In healthy humans, these tissues vibrate at rate of 80–500 Hz during speech and much higher during singing. Given the unique biophysical demands placed upon these structures, quantitative methods for measuring vocal fold pliability are essential to our collective insight regarding pathological conditions, such as edema, scar, and mass lesions, and their impact on the dynamics of phonation. Specifically, the need to quantify the response(s) to novel, experimental treatments using ex vivo models is lacking; this paucity of literature serves as the primary motivation for the current study. Historically, such investigation involved qualitative or semiquantitative evaluation of histological sections. The correlation between tissue architecture as determined by histology and the actual biomechanical properties of the VFs is unclear. Regardless, several have employed ⁎ Corresponding author at: Department of Biomaterials and Biomimetics, College of Dentistry, New York University, 345 E 24th Street, 804s, New York, NY 10010. Tel.: +1 212 998 9214. E-mail address: [email protected] (P.G. Coelho).
http://dx.doi.org/10.1016/j.msec.2014.08.051 0928-4931/© 2014 Elsevier B.V. All rights reserved.
various bulk measurements of tissue properties [1–4], including forceelongation experimentation [5], parallel plate rheometry [6], or linear skin rheometry [7,8]. These methods, although theoretically sound, largely fail to capture the true frequency demands placed upon the tissues. Furthermore, clinically, a relatively small region of fibrosis or other architectural aberration may have a profound impact on phonatory physiology (e.g., voice quality). These bulk measurements likely lack adequate sensitivity to sufficiently quantify these subtle tissue changes. The VFs have been modeled as being viscoelastic in previous studies [1–4]. In a viscoelastic material, the strain (a measure of deformation) of the material is dependent on not only the applied stress but also the frequency at which it is applied [9]. The modulus of such a viscoelastic material can be represented as a complex modulus, E*, which has both real and imaginary components in its mathematical representation. E* is composed of two parts, the storage modulus, E′, which represents the energy stored in the material, and the loss modulus, E″, which represents the viscous portion or the energy lost as heat. In a perfectly viscoelastic material, the values of the complex modulus should be independent of frequency. Furthermore, in a perfectly elastic material the phase lag, δ, is 0; therefore, there is no loss modulus. The complex modulus reduces simply to the storage modulus and is known as the elastic modulus (E). Indentation-based methods are newer in laryngology. Chhetri et al. [10] recently provided empiric data on the elastic properties of the VFs
334
P.G. Coelho et al. / Materials Science and Engineering C 45 (2014) 333–336
using indentation techniques involving a variety of indenter diameters in several anatomic conditions with reasonable accuracy. In the current study, a novel technique utilizing the dynamic mechanical analysis capabilities of a nano-indenter (nano-DMA) was employed. In nanoindentation, a small probe is used to indent the surface of a material of interest a small amount in a non-destructive manner and the resulting force and material displacement are recorded. Both the quasistatic behavior (behavior where inertial effects of the material can be ignored) and the dynamic behavior in the 10–300 Hz range was explored in this preliminary investigation. The resultant properties of the lamina propria were compared to that of the nearby muscle tissue. It is hypothesized that nano-indentation will differentiate between these two tissue types based on mechanical properties as quantified via nanoindentation. We sought to provide foundational, proof-of-principle data to support nano-indentation as a viable means to quantify changes in the biomechanical properties of the vocal folds.
2. Materials and methods Whole larynx specimens from untreated, adult rabbits (n = 4) were harvested and embedded in poly-methyl-methacrylate (PMMA). Sections were cut at 100 μm sequentially in the coronal plane and were polished to a surface smoothness suitable for nano-indentation testing. Slides were examined under light microscopy and those deemed to have adequately captured the lamina propria were used. Each animal yielded 2–4 such slides.
2.1. Mechanical testing Both quasistatic and dynamic testing was performed on both the lamina propria and the ipsilateral thyroarytenoid muscle in the samples using a Hysitron TI 950 TriboIndenter (Hysitron Inc., Minneapolis, MN) with a 100 μm radius conospherical (polycarbonate, E = 2.2 GPa) tip and nano-DMA III software (Fig. 1). All testing was conducted at room temperature, with thermal drift correction routines enabled, and under load control. Quasistatic testing was performed using a 4 s ramp to 1000 μN, a 5 s hold, and then 1 s unloading, with the unloading curve used to calculate the quasistatic reduced elastic modulus of the material, Er. For the quasistatic testing, 15 indentations per area of interest per slide were performed. Indents were spaced 100 μm apart from one another to ensure that previously indented tissue was not being tested.
For dynamic testing, a ramp to 1000 μN was used, and then a 1 μN amplitude sinusoidal wave was applied at multiple frequencies between 10 and 300 Hz. Storage (E′) and loss (E″) modulus were determined via nano-DMA III software (Hysitron Inc., Minneapolis, MN). For dynamic testing, 5 indentations per area of interest per slide were performed, with indents once again spaced at 100 μm apart from one another.
2.2. Statistical analyses Linear mixed effects modeling was performed using IBM SPSS Statistics software (IBM, Armonk, NY) on the quasistatic data using E as the dependent variable. Tissue type (lamina propria vs. muscle) was the independent predictor, and the animal from which the tissue came was considered a random effect. The number of animals was used as the count for statistical analyses and not the number of indentations throughout the study. For the dynamic testing, due to the large variation in the data, the results were first ranked, and then linear mixed effects modeling was performed on the ranked data with rank E′ and rank E″ as the dependent variables. Both tissue type and frequency were considered independent predictors and the random effects included both the animal (to account for differences between subjects) and the slide (to account for differences that were slide preparation dependent). An alpha of 0.05 was used for all statistical analyses. As an internal control, the laterality of the vocal fold was recorded (left vs. right), and laterality was tested as an independent predictor of the dependent variables mentioned above in separate analyses of both the quasistatic and the dynamic data. The same methods as described above were used on a reduced data set that only included lamina propria data, and with laterality replacing tissue type in model.
3. Results Quasistatic testing found Er of the lamina propria to be 609 ± 138 MPa and the muscle to be 758 ± 142 MPa (Table 1). This difference was statistically significant (p = 0.001). The analysis of the dependence of Er on the laterality of the VF yielded values of 642 ± 135 MPa and 566 ± 134 MPa for the left and right VF, respectively. This difference was not statistically significant (p = 0.14). Dynamic testing found E′ for the lamina propria to be 790 ± 526 MPa and the muscle to be 1061 ± 928 MPa (Table 1, Fig. 2). However, linear mixed effects modeling of rank E′ did not confirm this difference to be significant (p = 0.95). Frequency dependence also did not achieve statistical significance (p = 0.18). Regarding E″, the lamina propria was 125 ± 157 MPa compared to 181 ± 241 MPa for muscle (Table 1, Fig. 3). Again, linear mixed effects modeling of rank E″ did not find the difference between tissues to be significant (p = 0.85); similarly, frequency dependence was also not significant (p = 0.41). Testing of the dependence of E′ and E″ on the laterality of the VF was not significant (p = 0.91 and 0.77, respectively).
Table 1 Mean ± standard deviation for the measured mechanical properties in MPa. For the lamina propria, the top row is the combined data (used in the models comparing vocal fold to muscle), and the separate values for the left and right are shown in the bottom row for each dependent variable (used in the models comparing laterality). The difference in Er between lamina propria and muscle was statistically significant. The differences in E′ and E″ were not. In none of the analyses was the laterality of the lamina propria significant. Vocal fold Er (static) E′ (storage) Fig. 1. Typical view of section of interest for nano-indentation. Indenter diameter is included as well as typical indent diameter for reference. The indenter diameter pictured is 200 μm, and the typical indent was 20–25 μm in diameter.
E″ (loss)
609 642 790 667 125 119
± ± ± ± ± ±
138 135 526 495 157 149
Muscle 758 ± 142 566 ± 134 1061 ± 928 857 ± 445 181 ± 241 128 ± 139
P.G. Coelho et al. / Materials Science and Engineering C 45 (2014) 333–336
Fig. 2. Storage modulus vs. frequency for lamina propria and muscle. Notice the apparent lack of dependence upon frequency.
4. Discussion This study used a novel method to quantify the mechanical properties of vocal fold tissue via nano-indentation. Previously, nano-indentationbased methods have been applied to other soft tissues and viscoelastic materials [11–14]. Biological tissues tend to be anisotropic by nature due to their structure there can be significant variance in tissue properties tested in different locations even within the same tissue type due to changes in orientation, making testing biological tissues particularly challenging. Despite these known difficulties, this study showed that the lamina propria can be differentiated from muscle using a quasistatic nano-indentation loading protocol based on the resulting reduced modulus (Er) values for coronally sectioned specimens indented in the anterior to posterior plane. It is possible that different orientations of sectioning may influence the outcome for the aforementioned reasons, and slight differences could contribute to the wide variances seen in this work in the dynamic data. Additionally, the study did not find frequency dependence over the 10–300 Hz range for either the storage (E′) or loss (E″) moduli for either the lamina propria or muscle tissue but did find both of these values to be greater than zero across tissue types. This finding lends support to the VF being modeled as a viscoelastic solid. There was, unfortunately, no significant difference found between these two tissues in either of these dynamic properties when examined using ranked data. However, it should be noted (Figs. 2 and 3) that the deviations for the measurements were large, and that only four animals were included in this preliminary study. This method, in its current implementation, has limitations. The first is that the values for the moduli are in the 100 s of MPa range in this study, while previous studies have been in the kPa range [1–4,10]. Given this finding, the assumption of an indenter that does not significantly deform during indentation may not necessarily hold true since
Fig. 3. Loss modulus vs. frequency for lamina propria and muscle. Notice the apparent lack of dependence upon frequency.
335
the modulus of the polycarbonate indenter was only 4–5 times that of the tested material. Additionally, the modulus of the PMMA in which the samples were embedded is 1.8–3.1 GPa; one may hypothesize that the embedding material may have contributed to the stiffness to some degree and could potentially suggest that infiltration of the embedding material could be problematic. However, this limitation must be considered in the context that the current methods captured statistically significant differences between tissue types within each sample. In this regard, perhaps the greatest strength of this method in its current development stage is not its ability to acquire exact values for mechanical properties because values will vary between individuals but its ability to be a reproducible method for comparison of tissues. Due to its increased spatial resolution over previously published methods, nano-indentation would allow for the comparison of the quasistatic and dynamic mechanical properties of repaired tissue to native tissue within an individual specimen. One could imagine that this would be a powerful adjunct to histological analyses in determining the effectiveness of different surgical procedures or novel treatments with the caveat that further research into appropriate embedding techniques for testing is required. 5. Conclusion In summary, we present preliminary data regarding the use of nanoindentation methods to quantify both static and dynamic mechanical properties of vocal folds. Using a ramp-hold-unload and a frequency sweep loading pattern, both the quasistatic reduced modulus (Er) and the dynamic storage (E′) and loss (E″) moduli were determined. Quasistatic analysis differentiated the lamina propria from surrounding muscle based on a reduced modulus. No differences between the lamina propria and the muscle were observed with respect to the two dynamic moduli; however, both were non-zero, confirming the viscoelastic nature of the system. The values found for the moduli in the current study were orders of magnitude higher than previously reported values. This difference was likely due to the embedding technique used in sample preparation and will be further refined in future work. It is hypothesized that this method, while not yet refined to provide absolute values, will likely find significant utility not in providing exact values for the mechanical properties of this complex tissue, but as a quantitative comparison tool. Furthermore, these data will provide a critical adjunct to traditional histological analyses for the comparison of vocal folds undergoing various surgical/medical treatment regimens and native healthy vocal folds to aid in determining effective treatments to restore tissue dynamics. References [1] R.W. Chan, M.L. Rodriguez, A simple-shear rheometer for linear viscoelastic characterization of vocal fold tissues at phonatory frequencies, J. Acoust. Soc. Am. 124 (2) (2008 Aug) 1207–1219. [2] R.W. Chan, Measurements of vocal fold tissue viscoelasticity: approaching the male phonatory frequency range, J. Acoust. Soc. Am. 115 (6) (2004 Jun) 3161–3170. [3] I.R. Titze, S.A. Klemuk, S. Gray, Methodology for rheological testing of engineered biomaterials at low audio frequencies, J. Acoust. Soc. Am. 115 (1) (2004 Jan) 392–401. [4] R.W. Chan, I.R. Titze, Viscoelastic shear properties of human vocal fold mucosa: measurement methodology and empirical results, J. Acoust. Soc. Am. 106 (4 Pt 1) (1999 Oct) 2008–2021. [5] F. Alipour-Haghighi, S. Jaiswal, S. Vigmostad, Vocal fold elasticity in pig, sheep, and cow larynges, J. Voice 25 (2) (2011) 130–136. [6] R. Chan, I.R. Titze, Viscoelastic shear properties of human vocal fold mucosa: measurement methodology and empirical results, J. Acoust. Soc. Am. 106 (4) (1999) 2008–2021. [7] M. Hess, F. Mueller, J.B. Kobler, S.M. Zeitels, E. Goodyer, Measurements of vocal fold elasticity using the linear skin rheometer, Folia Phoniatr. Logop. 58 (3) (2006) 207–213. [8] S. Dailey, I. Tateya, D. Montequin, N. Welham, E. Goodyer, Viscoelastic measurements of vocal folds using the linear skin rheometer, J. Voice 23 (2) (2009) 143–150. [9] N.E. Dowling, Mechanical Behavior of Materials: Engineering Methods for Deformation, Fracture, and Fatigue, 2nd ed. Upper Saddle River, New Jersey, Prentice Hall, 1999.
336
P.G. Coelho et al. / Materials Science and Engineering C 45 (2014) 333–336
[10] D.K. Chhetri, Z. Zhang, J. Neubauer, Measurement of Young's modulus of vocal folds by indentation, J. Voice 25 (1) (2011 Jan) 1–7. [11] V.T. Nayar, J.D. Weiland, C.S. Nelson, A.M. Hodge, Elastic and viscoelastic characterization of agar, J. Mech. Behav. Biomed. Mater. 7 (2012 Mar) 60–68. [12] V.T. Nayar, J.D. Weiland, A.M. Hodge, Macrocompression and nanoindentation of soft viscoelastic biological materials, Tissue Eng. Part C Methods 18 (12) (2012 Dec) 968–975.
[13] R. Akhtar, N. Schwarzer, M.J. Sherratt, R.E.B. Watson, H.K. Graham, A.W. Trafford, et al., Nanoindentation of histological specimens: mapping the elastic properties of soft tissues, J. Mater. Res. 24 (3) (2009 Mar) 638–646. [14] D.M. Ebenstein, L.A. Pruitt, Nanoindentation of soft hydrated materials for application to vascular tissues, J. Biomed. Mater. Res. A 69 (2) (2004 May 1) 222–232.
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Preliminary investigation of a novel technique for the quantification of the ex vivo biomechanical properties of the vocal folds Paulo G. Coelho a,⁎, Michael Sobieraj b, Nick Tovar a, Kenneth Andrews c, Benjamin Paul c, Nandini Govil c, Seema Jeswani c, Milan R. Amin c, Malvin N. Janal d, Ryan C. Branski c a
Biomaterials and Biomimetics, College of Dentistry, New York University, New York, NY, USA Orthopedic Surgery, Hospital for Joint Diseases, New York University, New York, NY, USA NYU Voice Center, Department of Otolaryngology-Head and Neck Surgery, New York University School of Medicine, New York, NY, USA d Departments of Public Health and Epidemiology, New York University, New York, NY, USA b c
a r t i c l e
i n f o
Article history: Received 27 October 2013 Received in revised form 21 July 2014 Accepted 29 August 2014 Available online 17 September 2014 Keywords: Vocal fold Mechanical testing Quasi-static Dynamic Rabbit
a b s t r a c t The human vocal fold is a complex structure made up of distinct layers that vary in cellular and extracellular matrix composition. Elucidating the mechanical properties of vocal fold tissues is critical for the study of both acoustics and biomechanics of voice production, and essential in the context of vocal fold injury and repair. Both quasistatic and dynamic behavior in the 10–300 Hz range was explored in this preliminary investigation. The resultant properties of the lamina propria were compared to that of the nearby thyroarytenoid muscle. Er, quantified via quasistatic testing of the lamina propria, was 609 ± 138 MPa and 758 ± 142 MPa in the muscle (p = 0.001). E′ of the lamina propria as determined by dynamic testing was 790 ± 526 MPa compared to 1061 ± 928 MPa in the muscle. Differences in E′ did not achieve statistical significance via linear mixed effect modeling between the tissue types (p = 0.95). In addition, frequency dependence was not significant (p = 0.18). © 2014 Elsevier B.V. All rights reserved.
1. Introduction The human vocal fold (VF) is a complex structure made up of distinct layers that vary in cellular and extracellular matrix density and composition. Elucidating the mechanical properties of VF tissues is critical to our increased understanding of both the acoustics and the biomechanics of voice production and essential in the context of the inherent tissue response to injury and repair. In healthy humans, these tissues vibrate at rate of 80–500 Hz during speech and much higher during singing. Given the unique biophysical demands placed upon these structures, quantitative methods for measuring vocal fold pliability are essential to our collective insight regarding pathological conditions, such as edema, scar, and mass lesions, and their impact on the dynamics of phonation. Specifically, the need to quantify the response(s) to novel, experimental treatments using ex vivo models is lacking; this paucity of literature serves as the primary motivation for the current study. Historically, such investigation involved qualitative or semiquantitative evaluation of histological sections. The correlation between tissue architecture as determined by histology and the actual biomechanical properties of the VFs is unclear. Regardless, several have employed ⁎ Corresponding author at: Department of Biomaterials and Biomimetics, College of Dentistry, New York University, 345 E 24th Street, 804s, New York, NY 10010. Tel.: +1 212 998 9214. E-mail address: [email protected] (P.G. Coelho).
http://dx.doi.org/10.1016/j.msec.2014.08.051 0928-4931/© 2014 Elsevier B.V. All rights reserved.
various bulk measurements of tissue properties [1–4], including forceelongation experimentation [5], parallel plate rheometry [6], or linear skin rheometry [7,8]. These methods, although theoretically sound, largely fail to capture the true frequency demands placed upon the tissues. Furthermore, clinically, a relatively small region of fibrosis or other architectural aberration may have a profound impact on phonatory physiology (e.g., voice quality). These bulk measurements likely lack adequate sensitivity to sufficiently quantify these subtle tissue changes. The VFs have been modeled as being viscoelastic in previous studies [1–4]. In a viscoelastic material, the strain (a measure of deformation) of the material is dependent on not only the applied stress but also the frequency at which it is applied [9]. The modulus of such a viscoelastic material can be represented as a complex modulus, E*, which has both real and imaginary components in its mathematical representation. E* is composed of two parts, the storage modulus, E′, which represents the energy stored in the material, and the loss modulus, E″, which represents the viscous portion or the energy lost as heat. In a perfectly viscoelastic material, the values of the complex modulus should be independent of frequency. Furthermore, in a perfectly elastic material the phase lag, δ, is 0; therefore, there is no loss modulus. The complex modulus reduces simply to the storage modulus and is known as the elastic modulus (E). Indentation-based methods are newer in laryngology. Chhetri et al. [10] recently provided empiric data on the elastic properties of the VFs
334
P.G. Coelho et al. / Materials Science and Engineering C 45 (2014) 333–336
using indentation techniques involving a variety of indenter diameters in several anatomic conditions with reasonable accuracy. In the current study, a novel technique utilizing the dynamic mechanical analysis capabilities of a nano-indenter (nano-DMA) was employed. In nanoindentation, a small probe is used to indent the surface of a material of interest a small amount in a non-destructive manner and the resulting force and material displacement are recorded. Both the quasistatic behavior (behavior where inertial effects of the material can be ignored) and the dynamic behavior in the 10–300 Hz range was explored in this preliminary investigation. The resultant properties of the lamina propria were compared to that of the nearby muscle tissue. It is hypothesized that nano-indentation will differentiate between these two tissue types based on mechanical properties as quantified via nanoindentation. We sought to provide foundational, proof-of-principle data to support nano-indentation as a viable means to quantify changes in the biomechanical properties of the vocal folds.
2. Materials and methods Whole larynx specimens from untreated, adult rabbits (n = 4) were harvested and embedded in poly-methyl-methacrylate (PMMA). Sections were cut at 100 μm sequentially in the coronal plane and were polished to a surface smoothness suitable for nano-indentation testing. Slides were examined under light microscopy and those deemed to have adequately captured the lamina propria were used. Each animal yielded 2–4 such slides.
2.1. Mechanical testing Both quasistatic and dynamic testing was performed on both the lamina propria and the ipsilateral thyroarytenoid muscle in the samples using a Hysitron TI 950 TriboIndenter (Hysitron Inc., Minneapolis, MN) with a 100 μm radius conospherical (polycarbonate, E = 2.2 GPa) tip and nano-DMA III software (Fig. 1). All testing was conducted at room temperature, with thermal drift correction routines enabled, and under load control. Quasistatic testing was performed using a 4 s ramp to 1000 μN, a 5 s hold, and then 1 s unloading, with the unloading curve used to calculate the quasistatic reduced elastic modulus of the material, Er. For the quasistatic testing, 15 indentations per area of interest per slide were performed. Indents were spaced 100 μm apart from one another to ensure that previously indented tissue was not being tested.
For dynamic testing, a ramp to 1000 μN was used, and then a 1 μN amplitude sinusoidal wave was applied at multiple frequencies between 10 and 300 Hz. Storage (E′) and loss (E″) modulus were determined via nano-DMA III software (Hysitron Inc., Minneapolis, MN). For dynamic testing, 5 indentations per area of interest per slide were performed, with indents once again spaced at 100 μm apart from one another.
2.2. Statistical analyses Linear mixed effects modeling was performed using IBM SPSS Statistics software (IBM, Armonk, NY) on the quasistatic data using E as the dependent variable. Tissue type (lamina propria vs. muscle) was the independent predictor, and the animal from which the tissue came was considered a random effect. The number of animals was used as the count for statistical analyses and not the number of indentations throughout the study. For the dynamic testing, due to the large variation in the data, the results were first ranked, and then linear mixed effects modeling was performed on the ranked data with rank E′ and rank E″ as the dependent variables. Both tissue type and frequency were considered independent predictors and the random effects included both the animal (to account for differences between subjects) and the slide (to account for differences that were slide preparation dependent). An alpha of 0.05 was used for all statistical analyses. As an internal control, the laterality of the vocal fold was recorded (left vs. right), and laterality was tested as an independent predictor of the dependent variables mentioned above in separate analyses of both the quasistatic and the dynamic data. The same methods as described above were used on a reduced data set that only included lamina propria data, and with laterality replacing tissue type in model.
3. Results Quasistatic testing found Er of the lamina propria to be 609 ± 138 MPa and the muscle to be 758 ± 142 MPa (Table 1). This difference was statistically significant (p = 0.001). The analysis of the dependence of Er on the laterality of the VF yielded values of 642 ± 135 MPa and 566 ± 134 MPa for the left and right VF, respectively. This difference was not statistically significant (p = 0.14). Dynamic testing found E′ for the lamina propria to be 790 ± 526 MPa and the muscle to be 1061 ± 928 MPa (Table 1, Fig. 2). However, linear mixed effects modeling of rank E′ did not confirm this difference to be significant (p = 0.95). Frequency dependence also did not achieve statistical significance (p = 0.18). Regarding E″, the lamina propria was 125 ± 157 MPa compared to 181 ± 241 MPa for muscle (Table 1, Fig. 3). Again, linear mixed effects modeling of rank E″ did not find the difference between tissues to be significant (p = 0.85); similarly, frequency dependence was also not significant (p = 0.41). Testing of the dependence of E′ and E″ on the laterality of the VF was not significant (p = 0.91 and 0.77, respectively).
Table 1 Mean ± standard deviation for the measured mechanical properties in MPa. For the lamina propria, the top row is the combined data (used in the models comparing vocal fold to muscle), and the separate values for the left and right are shown in the bottom row for each dependent variable (used in the models comparing laterality). The difference in Er between lamina propria and muscle was statistically significant. The differences in E′ and E″ were not. In none of the analyses was the laterality of the lamina propria significant. Vocal fold Er (static) E′ (storage) Fig. 1. Typical view of section of interest for nano-indentation. Indenter diameter is included as well as typical indent diameter for reference. The indenter diameter pictured is 200 μm, and the typical indent was 20–25 μm in diameter.
E″ (loss)
609 642 790 667 125 119
± ± ± ± ± ±
138 135 526 495 157 149
Muscle 758 ± 142 566 ± 134 1061 ± 928 857 ± 445 181 ± 241 128 ± 139
P.G. Coelho et al. / Materials Science and Engineering C 45 (2014) 333–336
Fig. 2. Storage modulus vs. frequency for lamina propria and muscle. Notice the apparent lack of dependence upon frequency.
4. Discussion This study used a novel method to quantify the mechanical properties of vocal fold tissue via nano-indentation. Previously, nano-indentationbased methods have been applied to other soft tissues and viscoelastic materials [11–14]. Biological tissues tend to be anisotropic by nature due to their structure there can be significant variance in tissue properties tested in different locations even within the same tissue type due to changes in orientation, making testing biological tissues particularly challenging. Despite these known difficulties, this study showed that the lamina propria can be differentiated from muscle using a quasistatic nano-indentation loading protocol based on the resulting reduced modulus (Er) values for coronally sectioned specimens indented in the anterior to posterior plane. It is possible that different orientations of sectioning may influence the outcome for the aforementioned reasons, and slight differences could contribute to the wide variances seen in this work in the dynamic data. Additionally, the study did not find frequency dependence over the 10–300 Hz range for either the storage (E′) or loss (E″) moduli for either the lamina propria or muscle tissue but did find both of these values to be greater than zero across tissue types. This finding lends support to the VF being modeled as a viscoelastic solid. There was, unfortunately, no significant difference found between these two tissues in either of these dynamic properties when examined using ranked data. However, it should be noted (Figs. 2 and 3) that the deviations for the measurements were large, and that only four animals were included in this preliminary study. This method, in its current implementation, has limitations. The first is that the values for the moduli are in the 100 s of MPa range in this study, while previous studies have been in the kPa range [1–4,10]. Given this finding, the assumption of an indenter that does not significantly deform during indentation may not necessarily hold true since
Fig. 3. Loss modulus vs. frequency for lamina propria and muscle. Notice the apparent lack of dependence upon frequency.
335
the modulus of the polycarbonate indenter was only 4–5 times that of the tested material. Additionally, the modulus of the PMMA in which the samples were embedded is 1.8–3.1 GPa; one may hypothesize that the embedding material may have contributed to the stiffness to some degree and could potentially suggest that infiltration of the embedding material could be problematic. However, this limitation must be considered in the context that the current methods captured statistically significant differences between tissue types within each sample. In this regard, perhaps the greatest strength of this method in its current development stage is not its ability to acquire exact values for mechanical properties because values will vary between individuals but its ability to be a reproducible method for comparison of tissues. Due to its increased spatial resolution over previously published methods, nano-indentation would allow for the comparison of the quasistatic and dynamic mechanical properties of repaired tissue to native tissue within an individual specimen. One could imagine that this would be a powerful adjunct to histological analyses in determining the effectiveness of different surgical procedures or novel treatments with the caveat that further research into appropriate embedding techniques for testing is required. 5. Conclusion In summary, we present preliminary data regarding the use of nanoindentation methods to quantify both static and dynamic mechanical properties of vocal folds. Using a ramp-hold-unload and a frequency sweep loading pattern, both the quasistatic reduced modulus (Er) and the dynamic storage (E′) and loss (E″) moduli were determined. Quasistatic analysis differentiated the lamina propria from surrounding muscle based on a reduced modulus. No differences between the lamina propria and the muscle were observed with respect to the two dynamic moduli; however, both were non-zero, confirming the viscoelastic nature of the system. The values found for the moduli in the current study were orders of magnitude higher than previously reported values. This difference was likely due to the embedding technique used in sample preparation and will be further refined in future work. It is hypothesized that this method, while not yet refined to provide absolute values, will likely find significant utility not in providing exact values for the mechanical properties of this complex tissue, but as a quantitative comparison tool. Furthermore, these data will provide a critical adjunct to traditional histological analyses for the comparison of vocal folds undergoing various surgical/medical treatment regimens and native healthy vocal folds to aid in determining effective treatments to restore tissue dynamics. References [1] R.W. Chan, M.L. Rodriguez, A simple-shear rheometer for linear viscoelastic characterization of vocal fold tissues at phonatory frequencies, J. Acoust. Soc. Am. 124 (2) (2008 Aug) 1207–1219. [2] R.W. Chan, Measurements of vocal fold tissue viscoelasticity: approaching the male phonatory frequency range, J. Acoust. Soc. Am. 115 (6) (2004 Jun) 3161–3170. [3] I.R. Titze, S.A. Klemuk, S. Gray, Methodology for rheological testing of engineered biomaterials at low audio frequencies, J. Acoust. Soc. Am. 115 (1) (2004 Jan) 392–401. [4] R.W. Chan, I.R. Titze, Viscoelastic shear properties of human vocal fold mucosa: measurement methodology and empirical results, J. Acoust. Soc. Am. 106 (4 Pt 1) (1999 Oct) 2008–2021. [5] F. Alipour-Haghighi, S. Jaiswal, S. Vigmostad, Vocal fold elasticity in pig, sheep, and cow larynges, J. Voice 25 (2) (2011) 130–136. [6] R. Chan, I.R. Titze, Viscoelastic shear properties of human vocal fold mucosa: measurement methodology and empirical results, J. Acoust. Soc. Am. 106 (4) (1999) 2008–2021. [7] M. Hess, F. Mueller, J.B. Kobler, S.M. Zeitels, E. Goodyer, Measurements of vocal fold elasticity using the linear skin rheometer, Folia Phoniatr. Logop. 58 (3) (2006) 207–213. [8] S. Dailey, I. Tateya, D. Montequin, N. Welham, E. Goodyer, Viscoelastic measurements of vocal folds using the linear skin rheometer, J. Voice 23 (2) (2009) 143–150. [9] N.E. Dowling, Mechanical Behavior of Materials: Engineering Methods for Deformation, Fracture, and Fatigue, 2nd ed. Upper Saddle River, New Jersey, Prentice Hall, 1999.
336
P.G. Coelho et al. / Materials Science and Engineering C 45 (2014) 333–336
[10] D.K. Chhetri, Z. Zhang, J. Neubauer, Measurement of Young's modulus of vocal folds by indentation, J. Voice 25 (1) (2011 Jan) 1–7. [11] V.T. Nayar, J.D. Weiland, C.S. Nelson, A.M. Hodge, Elastic and viscoelastic characterization of agar, J. Mech. Behav. Biomed. Mater. 7 (2012 Mar) 60–68. [12] V.T. Nayar, J.D. Weiland, A.M. Hodge, Macrocompression and nanoindentation of soft viscoelastic biological materials, Tissue Eng. Part C Methods 18 (12) (2012 Dec) 968–975.
[13] R. Akhtar, N. Schwarzer, M.J. Sherratt, R.E.B. Watson, H.K. Graham, A.W. Trafford, et al., Nanoindentation of histological specimens: mapping the elastic properties of soft tissues, J. Mater. Res. 24 (3) (2009 Mar) 638–646. [14] D.M. Ebenstein, L.A. Pruitt, Nanoindentation of soft hydrated materials for application to vascular tissues, J. Biomed. Mater. Res. A 69 (2) (2004 May 1) 222–232.