The protocol describes a method for the study of extracellular matrix viscoelasticity and its dependence on protein composition or environmental factors. The matrix system targeted is the mouse zonule. The performance of the method is demonstrated by comparing the viscoelastic behavior of wild-type zonular fibers with those lacking microfibril-associated glycoprotein-1.
Elasticity is essential to the function of tissues such as blood vessels, muscles, and lungs. This property is derived mostly from the extracellular matrix (ECM), the protein meshwork that binds cells and tissues together. How the elastic properties of an ECM network relate to its composition, and whether the relaxation properties of the ECM play a physiological role, are questions that have yet to be fully addressed. Part of the challenge lies in the complex architecture of most ECM systems and the difficulty in isolating ECM components without compromising their structure. One exception is the zonule, an ECM system found in the eye of vertebrates. The zonule comprises fibers hundreds to thousands of micrometers in length that span the cell-free space between the lens and the eyewall. In this report, we describe a mechanical technique that takes advantage of the highly organized structure of the zonule to quantify its viscoelastic properties and to determine the contribution of individual protein components. The method involves dissection of a fixed eye to expose the lens and the zonule and employs a pull-up technique that stretches the zonular fibers equally while their tension is monitored. The technique is relatively inexpensive yet sensitive enough to detect alterations in viscoelastic properties of zonular fibers in mice lacking specific zonular proteins or with aging. Although the method presented here is designed primarily for studying ocular development and disease, it could also serve as an experimental model for exploring broader questions regarding the viscoelastic properties of elastic ECM's and the role of external factors such as ionic concentration, temperature, and interactions with signaling molecules.
The eye of a vertebrate contains a living optical lens that helps focus images on the retina1. The lens is suspended on the optical axis by a system of delicate, radially-oriented fibers, as illustrated in Figure 1A. At one end, the fibers attach to the lens equator and, at the other, to the surface of the ciliary body. Their lengths span distances ranging from 150 µm in mice to 1 mm in humans. Collectively, these fibers are known as the zonule of Zinn2, the ciliary zonule, or simply the zonule. Ocular trauma, disease, and certain genetic disorders can affect the integrity of the zonular fibers3, resulting in their eventual failure and accompanying loss of vision. In mice, the fibers have a core comprised mostly of the protein fibrillin-2, surrounded by a mantle rich in fibrillin-14. Although zonular fibers are unique to the eye, they bear many similarities to elastin-based ECM fibers found elsewhere in the body. The latter are covered by a fibrillin-1 mantle5 and have similar dimensions to zonular fibers6. Other proteins, such as latent-transforming growth factor β-binding proteins (LTBPs) and microfibril-associated glycoprotein-1 (MAGP-1), are found in association with both types of fibers7,8,9,10,11. The elastic modulus of zonular fibers is in the range of 0.18-1.50 MPa12,13,14,15,16, comparable to that of elastin-based fibers (0.3-1.2 MPa)17. These architectural and mechanical similarities lead us to believe that any insight into the roles of zonule-associated proteins may help elucidate their roles in other ECM elastic fibers.
The main purpose of developing the method described here is to gain insights into the role of specific zonular proteins in the progression of inherited eye disease. The general approach is to compare the viscoelastic properties of zonular fibers in wild-type mice with those of mice carrying targeted mutations in genes encoding zonular proteins. While several methods have been used previously to measure the elasto-mechanical properties of zonular fibers, all were designed for the eyes of much larger animals12,13,14,15,16. As such models are not genetically tractable; we sought to develop an experimental method that was better suited to the small and delicate eyes of mice.
The method we developed for assessing the viscoelasticity of mouse zonular fibers is a technique we refer to as the pull-up assay4,18, which is summarized visually in Figure 1. A detailed description of the pull-up method and the analysis of the results is provided below. We begin by describing the construction of the apparatus, including the three-dimensional (3D)-printed parts used in the project. Next, we detail the protocol used for obtaining and preparing the eyes for the experiment. Lastly, we provide step-by-step instructions on how to obtain data for the determination of the viscoelastic properties of zonular fibers. In the Representative Results section, we share previously unpublished data obtained with our method on the viscoelastic properties of zonular fibers from mice lacking MAGP-119 as well as a control set obtained from age-matched wild-type animals. Finally, we conclude with general remarks on the advantages and limitations of the method, and suggestions for potential experiments that may elucidate how environmental and biochemical factors affect the viscoelastic properties of ECM fibers.
All animal experiments were approved by the Washington University Animal Studies Committee and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.
1. Fabrication of specialized parts and construction of apparatus
2. Sample preparation and data acquisition
The pull-up technique described here provides a straightforward approach for determining viscoelastic properties of zonular fibers in mice. In brief, the mouse eye is first preserved by injection of a fixative at physiological intraocular pressure. This approach maintains the natural inflation of the eye and keeps the fibers properly pre-tensioned (fixation was deemed acceptable after preliminary experiments demonstrated it did not alter the elasticity or strength of the fibers significantly). The back of the mouse eye is then removed by dissection to expose the lens and the zonular fibers that suspend it. The front of the eye is affixed to a platform and placed inside a Petri dish that rests on a digital scale. Next, a glass capillary attached to a micromanipulator is cemented to the posterior surface of the lens. The lens is then raised in 50 µm increments while the force on the scale is recorded. A reduction in the apparent weight of the preparation provides information about the forces that stretch the fibers. Each displacement is followed by an equilibration period lasting about 1 min to observe any relaxation of the stress induced by the displacement. Finally, the results are analyzed using a quasi-linear viscoelastic model designed specifically for the geometry of the mouse zonular fibers and the direction of the pull in for the assay4.
Typical viscoelastic data obtained with our method are shown in Figure 3. The curve appears inverted (negative) since the lift force on the lens reduces the weight of the dish/platform/eye assembly on the scale by an equivalent amount. The response includes instantaneous force spikes during each of the 50 µm vertical displacements of the lens, followed by a relaxation phase with a lifetime on the order of 10 s. A similar stress relaxation has been observed for bovine zonular fibers12. The magnitude of the instantaneous and relaxed forces increases with each step up to about 1000 s (~800 µm total displacement) and then begins to drop as fibers start to fail. Zonule failure is complete by the 1,500 s time point (~1.25 mm total displacement). Note that due to the evaporation of buffer in the course of the experiment, the curve does not return to the initial reading after the lens is freed from the eye.
Figure 4 contrasts the responses obtained for a Magp-1 knockout mouse (red curve) and an age-matched wild-type animal (blue curve). These curves have been corrected for evaporation, inverted, and the raw measurements of mass (see Figure 3) are now expressed as force (with units of mN). The initial viscoelastic response of the Magp-1-depleted zonule (time 0-600 s) closely resembles that of the wild-type, suggesting that the viscoelastic properties of the zonule were not significantly altered by the absence of Magp-1. However, the fibers appear to rupture at a much lower tension compared to their wild-type counterparts.
To illustrate the reliability of the method, we gathered data from multiple animals on the maximum instantaneous force applied to the eyes before their fibers ruptured. The results are shown in Figure 5. The data for 1-month-old mice exhibit very small values for the standard error of the mean (SEM) despite the relatively low number of samples used (n = 5 or 6), suggesting high reproducibility. The results indicate that the strength of the fibers differs significantly between the two genotypes (p-value = 2.4 x 10-6). Results not shown in the figures also suggest that there is a subtle but statistically significant increase in breaking force strength with age for wild-type animals (p-value = 0.024).
The pull-up method also can generate quantitative estimates of the viscoelastic parameters that account for the observed variations in temporal responses. Table 1 summarizes best-fit parameters to our MAGP-1 data, obtained with a quasi-linear viscoelastic model described previously4. The results show that both MAGP-1 deletion and aging can have highly significant impacts on some of the mechanical properties of zonular fibers.
Figure 1: A visual summary of the pull-up method. (A) Cross-sectional view of a vertebrate eye showing the lens and the zonular fibers that suspend it. (B) A general approach for determining viscoelastic behavior in zonular fibers by displacing the lens upward (away from the cornea). (C) Actual view of a dissected eye glued onto a platform with its lens being pulled upward by a glass probe attached to a micromanipulator. (D) Schematic of the entire apparatus. Please click here to view a larger version of this figure.
Figure 2: Fabrication of various parts. (A) Fabrication of the glass probe. A glass capillary is held at an angle and a flame is applied at a spot roughly 2 cm from one end. Within a few seconds, the end of the capillary starts to fall. The flame is removed when the end of the capillary is bent at about 90°. (B) Fabrication of the eye platform. The part is fabricated with a 3D stereolithography (SLA) printer. It measures 30 x 30 x 5 mm and contains three hemispherical indentations with 2.0, 2.5, and 3.0 mm diameters into which dissected eyes of various sizes are glued. (C) Fabrication of probe holder. This part was also fabricated with a 3D SLA printer. It consists of two orthogonal, 7.3 mm diameter rods. The lower rod contains a 1.5 mm bore and two 2.5 mm through holes on the outer surface to accommodate metal screws that secure the capillary probe in place. (D) Negative lens assembly. Images captured by the side microscope contain an astigmatic distortion due to the curvature of the Petri dish and the buffer solution. The lens assembly is designed to compensate for the distortion, allowing the side microscope to capture images in sharp focus. Please click here to view a larger version of this figure.
Figure 3: Typical raw data obtained with the assay. The graph shown was recorded with logging software that records data from a digital scale with a 0.01 g accuracy. The left edge of the graph (time 0) reflects the weight of the sample without a lifting force. The y-axis depicts mass in g. The lens is then lifted in 50 µm steps until all zonular fibers are broken and the Petri dish again rests fully on the scale. Note that the end reading is offset from the initial reading. The offset is due to gradual evaporation of the buffer solution during the course of the experiment and can be accounted for during data analysis, as outlined in step 2.2.14. Please click here to view a larger version of this figure.
Figure 4: Representative zonular force-displacement curves for wild-type and MAGP-1-deficient mice. The graph compares the viscoelastic response obtained after discrete displacements of the lens away from its equilibrium position. The response of an eye from a MAGP-1 knockout (KO) mouse tracks that of an age-matched wild-type animal up to the point where the fibers in the knockout mouse break prematurely. Please click here to view a larger version of this figure.
Figure 5: Zonular fiber breaking forces obtained with the pull-up method for MAGP-1 KO versus wild-type mice and at two ages. All measurements shown are based on n = 5 or 6 eyes, with error bars representing the standard error of the mean (SEM). Abbreviations: WT = wild-type; KO = MAGP-1 knockout. Please click here to view a larger version of this figure.
Genotype/age | G0 (Pa) | G∞ (Pa) | Ʈ (sec) | σf (Pa) | |
WT 1-month | MEAN | 2.34E+05 | 9.33E+04 | 16.3 | 9.61E+05 |
SD | 2.83E+04 | 2.94E+04 | 3.4 | 1.25E+05 | |
95% Cl | 5.55E+04 | 5.76E+04 | 6.7 | 2.45E+05 | |
KO 1-month | MEAN | 2.73E+05 | 6.74E+04 | 17.6 | 4.44E+05 |
SD | 6.30E+04 | 2.06E+04 | 3.8 | 7.85E+04 | |
95% Cl | 1.23E+05 | 4.03E+04 | 7.5 | 1.54E+05 | |
p values | 0.25 | 0.12 | 0.58 | 0.000022 | |
WT 1-year | MEAN | 1.98E+05 | 7.42E+04 | 17 | 1.41E+06 |
SD | 1.17E+05 | 2.39E+04 | 9.1 | 2.44E+05 | |
95% Cl | 2.29E+05 | 4.69E+04 | 17.9 | 4.79E+05 | |
KO 1-Year | MEAN | 1.70E+04 | 2.46E+04 | 12.9 | 5.05E+05 |
SD | 9.06E+03 | 8.04E+03 | 7.4 | 1.48E+05 | |
95% Cl | 1.78E+04 | 1.58E+04 | 14.4 | 2.91E+05 | |
p values | 0.0063 | 0.001 | 0.41 | 0.000014 | |
p values, age | WT | 0.46 | 0.23 | 0.85 | 0.002 |
KO | 0.0007 | 0.0068 | 0.26 | 0.44 |
Table 1: Viscoelastic properties obtained with a quasi-linear viscoelastic (QLV) model. Data scans like those shown in Figure 4 were analyzed with a QLV model developed specifically for the pull-up assay and the mouse zonule. Best fit parameters for the instantaneous (G0) and equilibrium (G∞) stiffnesses, the relaxation time constant (τ), and the ultimate tensile strength (σf) are shown. Abbreviations: SD = standard deviation; CI = confidence interval.
The zonule is an unusual ECM system where fibers are arranged symmetrically and can be manipulated identically by displacing the eye lens along the optical axis. The space can also be readily accessed without cellular disruption, allowing the fibers to be studied in an environment close to their native state. The pull-up technique takes advantage of this ECM presentation to manipulate the delicate fibers from mice, a genetically tractable system, and accurately quantify their mechanical properties. This has allowed us to examine the contribution of key ECM proteins (fibrillin-118, LTBP-24, and MAGP-1 reported here) to the biomechanical properties of the zonular fibers. Our analysis of fibrillin-1-deficient mice revealed that zonular fibers lacking fibrillin-1 weaken with age and eventually rupture, leading to the displacement of the lens within the eye (in humans, a condition known as ectopia lentis). Significantly, lens dislocation is also a common occurrence in patients with Marfan syndrome, a disease caused by mutations in the FBN1 gene20. Thus, the pull-up assay offers an opportunity to model aspects of human connective tissue disease in mice. In mice lacking LTPB-2 (a protein thought to be involved in the genesis of microfibrils), we were able to demonstrate that zonular fibers were produced in the absence of that protein, but ruptured at significantly lower stresses and eventually disintegrated with age4. These results suggest that LTBP-2 contributes to the longevity of the fibers rather than their synthesis. In the current study, we determined that MAGP-1-deficient fibers had similar viscoelastic properties to wild-type fibers but ruptured at lower stresses, with no sign of further age-related degradation. This would be consistent with a model in which fibers lacking MAGP-1 are intrinsically weaker as soon as they develop.
We note that the ultimate tensile strengths listed in Table 1 are estimated under the assumption that the fibers break somewhere mid-span. However, we cannot rule out the possibility that fiber failure is due to detachment from anchorage points on the lens surface or the ciliary body. If this were the case, the fracture tensile strength of the fiber could be higher than the values listed in Table 1. Microscopic analysis will be required to differentiate between these possibilities. Such analysis is far from trivial since the fibers involved are very thin (~0.5-0.6 µm in width) and nearly index-matched to water, making them essentially invisible. In the absence of this additional information, we can only state that the ultimate tensile strengths listed in Table 1 represent their lower limits. It would also be interesting, in principle, to check whether the force measurements differ depending on which direction the lens is pulled. In practice, however, pulling the lens from the anterior side would require removing the iris without damaging the zonular fibers lying immediately beneath. Such precise dissection is beyond what we can currently achieve with the mouse eye.
The relative simplicity of the method and the high reproducibility of its results are desirable qualities for comparative studies of ECM mechanical properties. Additionally, as demonstrated here, it is also possible to use the pull-up assay to obtain absolute values of viscoelastic parameters by assuming a viscoelastic model and fitting the time curves to it. For instance, using a standard quasi-linear viscoelastic (QLV) model, we were able to extract values for instantaneous (G0) and equilibrium (G∞) stiffnesses, the relaxation time constant (τ), and the ultimate tensile strength (σf) of zonular fibers from wild-type mice, as well as those lacking LTBP-24 or MAGP-1. The G0 and G∞ values obtained for wild-type animals in both studies vary from 6.7 x 104 Pa to 2.3 x 105 Pa, a range broadly comparable to those found in much larger fibers derived from human, bovine, and porcine zonules (1.8 x 105-1.5 x 106 Pa)12,13,14,15,16. This agreement among species suggests that these are universal features of these fibers, and gives us confidence that meaningful viscoelastic parameters can be extracted with our method.
A critical step for obtaining quality viscoelastic responses is the orientation of the dissected eye glued to the platform (step 2.1.9). Minor tilting (less than 10°) does not appear to affect the results significantly. Experiments performed outside this limit can generate curves with shapes that deviate from the ones shown in Figure 4. For example, some of those curves may possess two broad peaks instead of one.
Ideally, the procedure outlined in this paper would have been performed without fixation of the eyes, which limits our ability to assess the true viscoelastic parameters of fresh zonular fibers. However, after our preliminary experiments showed no significant difference between the paraformaldehyde-fixed and fresh samples, we decided to adopt fixation as it afforded several advantages. As alluded to in the Protocol, the use of fixed tissues helps to preserve the native stretch of the fibers for the pull-up experiments. Additionally, we found that fixation promoted greater adhesion of the UV glue to the eye capsule, thus reducing the chances of the probe detaching from the lens during the pull-up action, as commonly experienced with fresh samples (probe detachment can be readily recognized as a sudden return of the force to a baseline level). The fixation also prevented buckling of the eyewall in the direction of the pull. Despite this limitation, our method provides a robust approach for determining the relative contribution of protein components to the viscoelastic properties of the zonular fibers.
Although our work to date has focused on the contribution of specific proteins, the method could readily be adapted to study the effect of factors external to the fibers on their mechanical properties. Such factors include temperature, pH, calcium concentration, and the presence or absence of cross-linking enzymes. High-precision measurements could be achieved using our method in differential mode, i.e., by pre-tensioning the zonular fibers with an initial stress/strain, and then reading the difference in tension that ensues when external conditions are altered. Some of these interventions may conceivably affect the elasticity of the tissues that surround the zonule and hence produce changes in tension that compete with those generated in the zonule. Control measurements with isolated tissues would be needed to assess their relevance to the proposed experiments. We expect that such effects may be negligible, based on observations with the side camera showing that contiguous tissues behave as highly stiff materials that undergo essentially no deformation even when the zonular fibers are fully stretched.
The authors have nothing to disclose.
This work was supported by NIH R01 EY029130 (S.B.) and P30 EY002687 (S.B.), R01 HL53325 and the Ines Mandl Research Foundation (R.P.M.), the Marfan Foundation, and an unrestricted grant to the Department of Ophthalmology and Visual Sciences at Washington University from Research to Prevent Blindness. J.R. also received a grant from the University of Health Sciences and Pharmacy in support of this project.
1/4-20 hex screws 3/4 inch long | Thorlabs | SH25S075 | |
1/4-20 nut | Hardware store | ||
3D SLA printer | Anycubic | Photon | |
4-40 screws 3/8 inch long, 2 | Hardware store | ||
Capillaries, OD 1.2 mm and 3 inches long, no filament | WPI | 1B120-3 | |
Cyanoacrylate (super) glue | Loctite | ||
Digital Scale accurate to 0.01 g | Vernier | OHAUS Scout 220 | |
Excel | Microsoft | Spreadsheet | |
Gas cigarette lighter | |||
Inspection/dissection microscope | Amscope | SKU: SM-4NTP | Working distance ~ 15 cm |
Micromanipulator, Economy 4-axis | WPI | Kite-L | |
Motorized micrometer | Thorlabs | Z812B | |
Negative cylindrical lens | Thorlabs | LK1431L1 | -75 mm focal length |
Petri dishes, 50 mm | |||
Post holder, 3 inches | Thorlabs | PH3 | |
Post, 4 inches | Thorlabs | TR4 | |
Scale logging software | Vernier | LoggePro | |
Servo motor controller | Thorlabs | KDC101 | |
Servo motor controller software | Thorlabs | APT | |
Slotted base, 1 | Thorlabs | BA1S | |
Slotted bases, 2 | Thorlabs | BA2 | |
Stand for micromanipular | WPI | M-10 | |
USB-camera for microscope | Amscope | SKU: MD500 | |
UV activated glue with UV source | Amazon |