Circumferential Tensile Stiffness of Glaucomatous Trabecular Meshwork

Abstract

Purpose.: Our previous work indicated that a larger circumferential Young's modulus (E) of trabecular meshwork (TM) correlated with a higher outflow facility (C) in normal human donor eyes. The current study investigated the influence of glaucomatous TM stiffness and cellularity on C.

Methods.: Eight left eyes from glaucomatous human donors were perfused within 48 hours post mortem. Values of C were determined at pressures of 10, 20, 30, and 40 mm Hg. The TM was then dissected and imaged with optical coherence tomography to determine its cross-sectional area. Uniaxial tensile stress was applied longitudinally to TM segments to determine stress-strain curves. E was calculated at zero strain, representing the circumferential stiffness of the TM at a relaxed state. Confocal images of DAPI-stained TM segments were used to determine cellularity after mechanical stretching.

Results.: C (μL/min/mm Hg) of glaucomatous eyes was 0.18 ± 0.02 (mean ± SE) at 10 mm Hg and decreased to 0.11 ± 0.02 when the pressure was increased to 40 mm Hg. C measured at 30 and 40 mm Hg correlated with TM cellularity. E was 12.5 MPa and 1.4 (geometric mean and SE) and did not statistically correlate with postmortem time, age, C, or cellularity.

Conclusions.: Compared with data of normal eyes observed in a previous study, C in glaucomatous eyes was reduced significantly, and the amount of reduction increased with increasing the pressure. E of glaucomatous TM was approximately one-fifth that of normal TM. Prospective studies are needed to further investigate the influence of TM tensile stiffness on outflow regulation.

Introduction

The stiffness of the trabecular meshwork (TM) may play an important role in the pathogenesis of glaucoma, although mechanisms are still unknown. Recent data show that a stiffer TM in terms of circumferential or axial Young's modulus is more likely to be observed in normal donor eyes with higher outflow facility and less variation in outflow resistance in response to IOP elevation. ¹ On the other hand, microscopic or local stiffness, measured at juxtacanalicular tissue (JCT) and inner wall of Schlemm's canal (SC) with atomic force microscopy (AFM), is higher and more heterogeneous in glaucomatous eyes than in normal eyes, ² suggesting that a higher local stiffness of JCT correlates to a lower outflow facility in glaucomatous eyes. The two studies may appear conflicting to each other, but it has been well established that manner, rate, and direction in which a mechanical load is applied to tissue can result in very different stiffnesses. ³ Tensile stiffness measured by Camras et al. ¹ reflected resistance of the whole TM as an interconnected structure, including all cells and extracellular matrix (ECM), to mechanical stretch, whereas microindentation stiffness determined by Last et al. ² depended only on resistance of a region, smaller than a cell, to mechanical compression by the AFM probe. It is well known that the stiffness of a cell is several orders of magnitude smaller than that of a collagen fibril ^3–6 ; thus, results from the two studies reflect completely different mechanical properties of the TM and cannot be compared directly. In previous studies, the local stiffness has been compared between normal and glaucomatous TM, ² and the circumferential Young's modulus of normal TM has been measured, ¹ but the circumferential Young's modulus of glaucomatous TM is still unknown. 

TM cells may play an important role in regulating tissue stiffness through control of ECM turnover. ⁷ Cellularity of the TM decreases with age in both normal and primary open angle glaucoma (POAG). ^8,9 Additionally, the cellularity of POAG TM specimens as a whole were observed to be lower than that of normal specimens, with the exception of the JCT layer in which there was no significant difference in cellularity between the two groups. ⁸ At present, no studies to our knowledge have directly investigated stiffness of TM tissues with different cell densities. A direct correlation between reduction in TM cellularity and decrease in outflow facility has been hypothesized in the literature, ¹⁰ but has yet to be shown experimentally. 

This study aims to investigate the circumferential Young's modulus of TM, and its relationship to outflow function and TM cellularity in glaucomatous donor eyes. Findings in this study were compared with the data collected from normal donor eyes described previously. ¹ The comparison may provide important information for future analysis on how TM stiffness affects outflow facility. 

Methods

Eye Preparation

Human cadaver eyes with glaucoma were obtained from the Lions Eye Institute of Tampa, Florida, or the San Diego Eye Bank. With the exception of cataract surgery in some eyes, they had no reported history of ocular diseases and/or surgeries. Eyes were kept in saline-wetted gauze, placed in moist chambers, and shipped on ice. The left donor eyes (OS) were prepared for whole globe perfusion as described previously. ^1,11–14 As described in our previous study, ¹ all eyes were perfused within 48 hours post mortem. 

Perfusion System and Outflow Analysis

The perfusion system and methods described by Camras et al. ^1,15 were used to measure outflow facilities at perfusion pressures of 10, 20, 30, and 40 mm Hg. In brief, a syringe filled with an isotonic solution (5.5 mM of glucose in Dulbecco's PBS) ¹⁶ was connected to a three-way stopcock linked to two pressure tubings. One of the tubings was connected to a pressure transducer (Honeywell model 140 PC; Honeywell Sensing and Control, Freeport, IL). The other tubing was attached to a second three-way stopcock linked to a fluid column, which acted as a manometer to set the perfusion pressure through the control of fluid column height. The pressure transducer was connected using an analogue cable to a computer and calibrated by varying the height of the fluid column. The calibration was checked again prior to each experiment by referencing a fluid column height to pressure detected by the transducer. 

To determine the rate of perfusion, we measured the pressure reduction (ΔP) with the pressure transducer over a short period (Δt) such that the decrease in fluid column height was negligible. The perfusion rate was calculated as πr ²ΔP/(ρgΔt), where r is the radius of fluid column (2 mm), ρ is the mass density of fluid (1 g/cm³), and g is gravity (980 cm/s²). To determine the sensitivity and accuracy of the perfusion rate measurement, we blocked the outlet of the system at the second stopcock, and used a syringe pump to infuse fluid into the system via the first stopcock at different rates, ranging from 0.25 to 5 μL/min. Once the rate of infusion-induced pressure increase reached a steady state, infusion rate was independently calculated as πr ²ΔP/(ρgΔt), where ΔP was the pressure increase over a 5-minute period, and Δt = 5 minutes. It was observed that the maximum difference between the preset rate in the syringe pump and the calculated infusion rate was 5% if the flow rate was greater than 1 μL/min, indicating that the flow rate could be determined accurately through the pressure measurement. 

In experiments, the system pressure was referenced to zero corresponding to zero flow rate. A 25-gauge needle was attached to the second stopcock and placed in the posterior chamber of the eye for perfusion. A small moist Kimwipe (Kimberly-Clark, Corp., Roswell, GA) was placed over the cornea to prevent its dehydration. Eyes were set in a water bath maintained at 34°C. PowerLab software (ADInstruments, Inc., Colorado Springs, CO) was used to record the pressure for the duration of the experiment at a sampling rate of 10 Hz. At the end of the experiment, the needle was re-zeroed to verify that no obstructions in the perfusion system had occurred. 

Outflow facilities (C) and outflow resistances (R) were calculated from the stabilized flow rate (F) measured at a set perfusion pressure (P) (see Equation 1),^11,17 as we performed previously for normal eyes.¹ 

 

The perfusion pressure ex vivo is equivalent to the difference between IOP and episcleral venous pressure in vivo. As a result, P was set to 10, 20, 30, or 40 mm Hg in this study. For one eye, the flow rate at 10 mm Hg was too small to be measured. Thus, outflow facility analysis for this eye was only performed at pressures of 20, 30, and 40 mm Hg. The variance in outflow resistance, Var(R), was determined with the four outflow resistances for each eye (see Equation 2).  

 

where R_i with i = 1, 2, 3, or 4 is the outflow resistance measured at four different pressures, and R̄ is the mean of four R_i.  

Trabecular Meshwork Preparation and Mechanical Test

Eyes were disconnected from the perfusion system. The extraocular muscles and optic nerves were used to identify temporal and nasal regions of the eye and a color tissue marker was used to outline these regions prior to dissection. Then, the TM was isolated from the globes as described in our previous paper, ¹ and kept in physiological saline until optical and mechanical measurements. The details of the measurements have been described in our previous paper as well. In brief, the TM was cut into segments of approximately 5 to 10 mm long, with each segment located in a specific region (temporal or nasal). The TM segments were then secured to brackets to avoid prestretching and submerged in PBS to keep them hydrated for imaging and mechanical testing. First, 50 B-scans or cross-sectional views were acquired using optical coherence tomography (OCT) over a 5-mm portion of each TM segment. Volume reconstruction was performed to generate the three-dimensional (3-D) image of this portion using a Matlab program (courtesy of Hansford Hendargo, Department of Biomedical Engineering, Duke University; MathWorks, Natick, MA). Three cross-sectional areas, approximately 1 to 2 mm apart, in the 3-D image were calculated based on the average width and thickness of the three locations. After the OCT measurement, the TM underwent a quasi-static uniaxial tensile mechanical test using microstrain analyzer (MSA; TA Instruments RSA III) at room temperature within 6 hours after the whole eye perfusion. The instrument has a displacement resolution of 1 nm and a force resolution of 0.1 mN within the range of 1 mN to 35 N, which was adequate to accurately measure the minimum force (4.6 mN) observed in this study when TM was stretched by 2% of its original length. The instrument was calibrated with a 97.6-g weight and then offset to zero force before each measurement. 

For stress/strain measurement, the TM was loaded into the MSA by clamping the top and the bottom of the bracket into the upper and the lower fiber/film holders, respectively. Once loaded, the spine of the bracket was cut leaving the two ends of TM secured to the holders. A small stretch was initiated slowly until there was a force spike slightly above zero, showing that the TM was engaged and slightly tightened. The stretch was then stopped and the force quickly returned to zero. The tissue length measured at the force spike was defined as the initial sample length, and used to determine the rate of tissue stretch, which was set at 0.1% strain per second. Afterward, the test was initiated to obtain force versus percent strain (ε) curves with the TA Orchestrator software (TA Instruments, New Castle, DE). The data were finally exported to Microsoft Excel software for offline analysis. To minimize tissue dehydration during the mechanical test, the entire experimental procedure from loading of the TM to completion of the mechanical test was finished in less than 5 minutes. Our previous study¹ revealed that only 2% stretch of the TM was necessary and physiologically relevant for calculation of E of normal TM. Therefore, rather than stretching the glaucomatous TM to mechanical failure, as was done with the normal eyes,¹ the glaucomatous TM was only stretched by up to 5% of its original length. The force measured by MSA was divided by the average cross-sectional area of TM, determined by OCT, to obtain the stress (σ). Similar to the normal TM study,¹ the stress-strain curves with the strain (ε) varying between 0 and 2% were fitted with an exponential function (see Equation 3),¹⁸ 

 

where A and B are constants. Quality of the fit was measured by the coefficient of determination R², and we only accepted the results if R² was greater than 0.9. For most fits, we observed R² greater than 0.95. The Young's modulus (E) is the derivative of the stress with respect to strain,  

 

which depends on ε, as shown in Equation 4. At zero strain (ε = 0), E is equal to A. The constant B in Equation 4 determined how fast E increased with increasing the strain.  

Cellularity and Histological Analysis

Cellularity was evaluated after mechanical stretching of TM. In the experiment, TM was submerged in a solution of DAPI (2 μg/mL) for 20 minutes, then fixed in 2.5% glutaraldehyde and 2% paraformaldehyde. The fixed TM was imaged at excitation/emission wavelengths of 405/440 nm, using a spinning disk confocal microscope equipped with a ×20 objective (Revolution XD; Andor Technologies, South Windsor, CT). A Z-stack of tissue images with section thickness of 1 μm was generated for each TM. To avoid double counting of cells, three optical sections spaced 20 μm apart were selected for cellularity assessment. In each image (see Fig. 1), the tissue area was manually traced with ImageJ software (National Institutes of Health, Bethesda, MD). The number of pixels within the traced region was then converted to the tissue area by a factor of 4.91 × 10⁻⁷ mm²/pixel; and the number of cells in the same region was counted by a masked observer to determine the cellularity. 

Figure 1

 

Typical image of glaucomatous TM segments obtained with confocal microscopy after staining of cells with DAPI. It was used to determine TM cellularity. The original image is shown in (A). (B) The TM tissue area was outlined by the orange curve with ImageJ, in which TM cells were counted (red circles). The cellularity was calculated as the number of cells per unit TM tissue area.

Statistics

Data distributions are shown with box-and-whisker plots. If a distribution was significantly asymmetric about its median, logarithmic transformation would be performed to normalize the distribution. Differences between glaucomatous and normal samples were evaluated with unpaired, two-tailed Student t-tests for normally distributed data using StatView software (SAS Institute, Cary, NC). For data that were not normally distributed, Mann-Whitney U and Wilcoxon signed rank tests were used to compare unpaired and paired data sets, respectively. A difference was considered to be statistically significant if the P value was less than 0.05. Linear regression analysis was performed to evaluate correlation between two variables; and R ² is reported to show the strength of correlation. 

Results

Stiffness of Glaucomatous TM

Eye donors with glaucoma (n = 8) were primarily Caucasian, with ages ranging between 66 and 90 years old (see Tables 1 and 2). The left eyes (OS) were perfused within 48 hours post mortem, and outflow facilities shown in Table 3 were measured at pressures of 10, 20, 30, and 40 mm Hg, respectively. Table 4 shows mean dimensions of cross-section and cellularity of the TM. Stress-strain curves of all glaucomatous TM samples under stretch are shown in Figure 2A; they were used to determine the circumferential Young's modulus (E) of the TM at zero strain, and the exponent, B, defined in Equation 3. These data are also shown in Table 4. The distribution of E was skewed. Thus, it was normalized through logarithmic transformation; and the geometric mean and SE of E were observed to be 12.5 MPa and 1.4, respectively. Both original and transformed distributions of E are shown as box-and-whisker plots in Figure 2B. 

Figure 2

 

(A) Stress-strain curves of glaucomatous TM samples. The solid lines are experimental results for individual TM samples (n = 8), and the dashed line is the geometric average of individual curves. (B) Box and whisker plots of the Young's modulus (E) of TM. Both E and log₁₀(E) are plotted here.

Table 1

 

Glaucoma Donor Eye Information*

Donor	Sex	Ethnicity	Cause of Death	Age, y	PMT, h
1	Male	Caucasian	Pulmonary fibrosis, rectal bleed	82	46
2	Female	Caucasian	Renal cancer	76	41.5
3	Male	Caucasian	Cardiac arrest	84	42
4	Female	Unknown	Pneumonia	81	43.5
5	Female	Caucasian	Pneumonia	90	42
6	Male	Unknown	Septic shock	77	40
7	Male	Caucasian	Cancer-lung	84	42.5
8	Female	Unknown	Intracranial hemorrhage	66	41
Average				80.0	42.3

*  All experiments unless specified used only left eyes (OS) from donors (n = 8).

Table 2

 

Ocular Treatment of Glaucoma Donors

Donor	Year of Diagnosis	Latest IOP, mm Hg	Cup/Disc	Cataract Surgery	Glaucoma Medications
1	2003	13	0.6	Yes (2009)	Travatan (travoprost),* Azopt (brinzolamide)†
2	Unknown	Unknown	Unknown	Yes (2012)	Unspecified glaucoma drops
3	Unknown	13	0.7	Yes (2007)	Xalatan (latanoprost),† Azopt (brinzolamide)
4	Unknown	Unknown	Unknown	Yes (N/A)	Unspecified glaucoma drops
5	Unknown	Unknown	Unknown	Yes (N/A)	Alphagan (brimonidine)†
6	Unknown	Unknown	Unknown	Yes (N/A)	Lumigan (bimatoprost)*
7	1995	11	Unknown	Yes (2008)	Unspecified glaucoma drops
8	2009	Unknown	Unknown	No	Unspecified glaucoma drops

*  Prostaglandin derivatives that increase uveoscleral outflow.

†  Carbonic anhydrase inhibitors or Alpha-2 agonists that decrease aqueous humor production.

Table 3

 

Outflow Facility of Glaucomatous Eyes

Donor Eye, OS	C at 10 mm Hg*	C at 20 mm Hg	C at 30 mm Hg	C at 40 mm Hg	Var(R)†
1	Unreliable measurement	0.076	0.052	0.045	22.9
2	0.154	0.126	0.108	0.086	3.45
3	0.155	0.106	0.096	0.091	0.620
4	0.239	0.135	0.120	0.082	6.61
5	0.079	0.080	0.058	0.049	15.4
6	0.198	0.191	0.187	0.171	0.113
7	0.166	0.191	0.212	0.197	0.069
8	0.233	0.142	0.169	0.125	2.56
Average	0.175	0.131	0.125	0.106	6.47‡

*  C is outflow facility, μL/min/mm Hg.

†  Var(R) is the variance of outflow resistance (mm Hg/[μL/min]) ² .

‡  It is geometric mean of Var(R), (mm Hg/[μL/min]) ² ; other average values are arithmetic means.

Table 4

 

Circumferential Young's Modulus and Morphology of Glaucomatous TM

Donor, OS	B*	E, MPa	Width, μm	Thickness, μm	Cross-Sectional Area, mm2	Cellularity, Cells/mm2
1	20.5	3.0	369.1	154.8	0.0582	471
2	22.2	8.2	134.6	96.5	0.0135	986
3	34.7	7.3	173.3	153.4	0.0245	1710
4	28.2	52.6	110.6	109.1	0.0111	894
5	19.0	12.0	119.0	100.9	0.0117	511
6	39.6	13.2	241.9	64.3	0.0156	1387
7	65.6	12.2	223.3	81.0	0.0183	1696
8	32.4	31.4	212.4	42.2	0.0090	1472
Average	32.8 ± 5.8†	12.5 & 1.4‡	198 ± 30	100 ± 14	0.020 ± 0.006	1141 ± 175.9

*  B is the dimensionless constant defined in Equation 3.

†  Arithmetic mean ± SE.

‡  Geometric mean & SE, where “&” means “×/ ÷” and the geometric SE is dimensionless.

Comparison of Normal and Glaucoma Eyes

Table 5 summarizes differences in various characteristics of tissues between normal (n = 7) and glaucomatous eyes (n = 8). All data from healthy donors have been reported previously. ¹ Normal eyes were phakic, while all but one glaucomatous eye were pseudo-phakic. Comparison of other characteristics between normal and glaucoma eyes showed that there were insignificant differences in postmortem time (PMT), thickness, width and cross-sectional area of the dissected TM, and the exponent, B. Significant differences were observed in donor ages, outflow facilities (C), and E between the two groups (see also Fig. 3). We did not consider that the significant differences observed in C and E were age-related, as there were no significant correlations between age and any other characteristics tested in glaucoma eyes. 

Figure 3

 

(A) Circumferential Young's modulus E of normal and glaucomatous TM. The median E was 12.1 MPa for glaucomatous TM (n = 8), and 42.6 MPa for normal TM (n = 7). This difference was marginally significant with Mann-Whitney U test (P = 0.06). (B) Outflow facility C in normal and glaucomatous eyes. C in glaucomatous eyes decreased with pressure elevation. Its value at 40 mm Hg was significantly lower than those at 20 and 30 mm Hg (*P < 0.05, Wilcoxon Signed Rank test). C in normal eyes was significantly higher than that in glaucomatous eyes at pressures of 20, 30, and 40 mm Hg (**P < 0.05, Mann-Whitney U test).

Table 5

 

Comparison of Normal Versus Glaucomatous Data

Parameter	Normal Eyes, OS, n = 7	Glaucomatous Eyes, OS, n = 8	P Value*
PMT, h	38.8 ± 2.9†	42.3 ± 0.64	0.23
Age, y	61.0 ± 3.2	80.0 ± 2.5	< 0.005
Area, mm2	0.027 ± 0.006	0.020 ± 0.006	0.56
Thickness, μm	119 ± 18	100 ± 14	0.42
Width, μm	202 ± 27	198 ± 30	0.92
B	47.3 ± 22.9	32.8 ± 5.8	0.51
E, MPa	51.5 ± 13.6	17.5 ± 5.8 or 12.5 & 1.4	0.06‡
C at 10 mm Hg§	0.245 ± 0.032	0.175 ± 0.021	0.09
C at 20 mm Hg	0.254 ± 0.030	0.131 ± 0.016	< 0.005
C at 30 mmHg	0.253 ± 0.037	0.125 ± 0.021	< 0.01
C at 40 mm Hg	0.257 ± 0.043	0.106 ± 0.019	< 0.01
Var(R)‖	0.216 & 2.75	1.778 & 2.15	0.11

*  Statistical significance was determined with two-tailed, Student's t-test unless indicated. P values in bold indicate statistical significance.

†  Data are reported as either arithmetic mean ± SE or geometric mean & SE, where “&” is “×/ ÷” and all geometric SE data are dimensionless.

‡  Mann-Whitney U test was used to determine statistical significance.

§  Outflow facility (C, μL/min/mm Hg) measured at different pressures.

 ‖ Variance of outflow resistance (Var(R), [mm Hg/(μL/min)]²).

On average, E of normal TM was 4 times higher than that of glaucomatous TM, but this finding was only marginally statistically significant (see Fig. 3A). Meanwhile, we found that the difference in E between normal and glaucomatous TM became statistically significant (P < 0.01, Mann-Whitney U test) if the outlier data in the normal group was removed (see the Outlier Data Analysis below). The outflow facility (C) was also lower in glaucoma eyes than in normal eyes (Fig. 3B), and the difference was statistically significant at the perfusion pressures of 20, 30, and 40 mm Hg (P < 0.05, Mann-Whitney U test). C was independent of pressure in normal eyes, but its value at 40 mm Hg was significantly lower than those at 20 and 30 mm Hg in glaucomatous eyes (P < 0.05, Wilcoxon signed rank test). Unlike the normal TM, E of glaucomatous TM did not correlate with C or the variance in outflow resistance (Var(R)), but it might weakly (R ² = 0.60, P < 0.05) depend on the cross-sectional area of TM (Fig. 4) (see also the Outlier Data Analysis below). 

Figure 4

 

Correlation between cross-sectional area and Young's modulus of glaucomatous TM. The correlation was negative and statistically significant (P < 0.05, n = 8).

Linear correlation analysis was performed between the outflow facility at a given pressure and the logarithmically transformed Var(R), log₁₀(VAR), measured at four different pressures. The correlation for glaucomatous eyes was negative and statistically significant (P < 0.01, R ² > 0.70, n = 8) at pressures of 20, 30, and 40 mm Hg (see Fig. 5), but statistically insignificant at 10 mm Hg. The data were consistent with those observed in normal eyes reported in our previous study, ¹ which are also plotted in Figure 5 for comparison, except that at 10 mm Hg, the correlation was statistically significant for normal eyes (P < 0.05, R ² = 0.66, n = 7) but insignificant for glaucomatous eyes (P > 0.05, R ² = 0.04, n = 7). Presumably, this difference between normal and glaucomatous eyes could be due to the low outflow rate at 10 mm Hg in glaucomatous eyes. This outflow rate was very close to the lower limit of our perfusion system that could be controlled accurately. Thus, it was subject to a greater error. 

Figure 5

 

Correlation between outflow facility C and logarithmically transformed variance in outflow resistance log₁₀(VAR). An increased C led to a statistically significant decrease in log₁₀(VAR) at pressures of 20 mm Hg (P < 0.005), 30 mm Hg (p < 0.01), and 40 mm Hg (P < 0.001), but not at 10 mm Hg in glaucomatous eyes (n = 8). The data for normal eyes reported in our previous study ¹ are also plotted here for comparison.

TM cellularity was positively correlated to C in glaucomatous eyes (see Fig. 6) and the correlations were statistically significant at 30 mm Hg and 40 mm Hg (P < 0.05, n = 8), but insignificant at 10 and 20 mm Hg. The correlations could not be compared directly with those for normal eyes, as TM cellularity was not assessed in our previous study. ¹  

Figure 6

 

Correlation between outflow facility C and TM cellularity in glaucomatous eyes. C was higher in eyes with higher TM cellularity; and the correlation was statistically significant for C measured at 30 and 40 mm Hg (P < 0.05, n = 8).

Outlier Data Analysis

One “normal” eye reported previously ¹ had outflow impairment and its TM had a lower axial-tensile modulus. Its outflow facilities (0.14, 0.15, 0.12, and 0.08 μL/min/mm Hg measured at 10, 20, 30, and 40 mm Hg, respectively) were either within or below the interquartile ranges (IQR) of outflow facilities observed in glaucomatous eyes (see Table 3); and the Young's modulus of this TM (E = 1.31 MPa) was lower than the minimal E observed in glaucomatous eyes (see Table 4). If this “normal” TM was excluded in data analysis, the outflow facilities at all pressures in the glaucomatous eyes were then significantly lower than those in the normal eyes (n = 6) (P < 0.05, Mann-Whitney U test), and the difference in E between normal TM (59.9 ± 11.8 MPa, n = 6) and glaucomatous TM (12.5 MPa & 1.4, n = 8) became statistically significant (P < 0.01, Mann-Whitney U test). 

The largest cross-sectional area of glaucomatous TM is 0.058 mm², which is larger than the third quartile plus 1.5*IQR (see Fig. 4). As a result, it could be considered statistically as an outlier and be excluded from data analysis. After exclusion of this outlier, the Young's modulus became insignificantly dependent on the cross-sectional area of TM (R ² = 0.46, P > 0.05, n = 7). 

A Case Study of Glaucomatous Eye

We used left eyes (OS) with glaucoma for all experiments reported above. Additionally, we investigated a right eye (OD) with glaucoma undergone a filtration surgery, which was not included in group studies described above. Although specifications for surgery were not provided in the donor information, an Express Shunt was found in the superior quadrant of the eye (see Fig. 7). The donor was a Caucasian male who died of respiratory failure. Similar to the other glaucomatous eyes, this eye was pseudophakic and medicated with glaucoma drops (Alphagan, Azopt, and Travatan). We perfused the eye within 44 hours PMT, and observed that C was normal (0.23 to 0.29 μL/min/mm Hg) with perfusion pressure ranging from 10 to 40 mm Hg. E of its TM was 2.5 MPa, which was approximately one-fifth of the average E of glaucomatous TM and smaller than the E of all TMs in the left eyes without the shunt implant (see Table 4). 

Figure 7

 

Image of Express Shunt in a right eye with glaucoma. It was found in the superior quadrant of the eye after removal of the uveal tissue.

Discussion

This study marks the first time that the circumferential tensile stiffness of human glaucomatous TM is evaluated for determination of its relationship to outflow function. To compare the data measured with glaucomatous eyes in this study with those for normal eyes reported in the previous study, ¹ we designed the current study based on the following criteria: to receive glaucomatous eyes from the same eye banks, to follow the same experimental procedures, and to use the same instrument as those used in the normal eye study. ¹ To further reduce data variability caused by unexpected procedures, all dissections and measurements were performed by the same individual (LJC), the microstrain analyzer was calibrated with a 97.6-g weight before each measurement, and all data were analyzed using the same methods. As a result, the observed differences in E and C between normal and glaucomatous eyes were caused mainly by differences in tissue structures. 

The major finding in this study was that compared with normal eyes, ¹ the TM in glaucomatous eyes had a lower circumferential tensile modulus. Unlike normal eyes, ¹ glaucomatous eyes showed insignificant correlations between E of TM and C, and between E and Var(R). Regression analysis showed that higher cellularity of TM was correlated with higher outflow facility. One glaucomatous eye, which had undergone filtration surgery previously and thereby excluded from further analysis, had normal outflow facility and a TM with low E; presumably, due to reduction in outflow through the TM, as most of the flow bypassed the TM via the shunt. 

Stiffness of Glaucomatous TM and Its Effects on Outflow Function

The circumferential tensile stiffness of glaucomatous TM was observed in the present study to be less than that of normal TM. ¹ Interestingly, the local compressive stiffness of the JCT/inner wall of SC measured previously with a microindention method was found to be higher in glaucomatous eyes than in normal eyes. ² To explain this apparent discrepancy in the observations, we need to consider how stiffnesses measured with different methods are related to tissue structures. The total area occupied by JCT and inner wall of SC in a cross-section of TM tissue is tiny compared with that occupied by uveal and corneoscleral meshworks. The circumferential Young's modulus of TM measured in the current and our previous ¹ studies is determined mainly by the stiffness of uveal and corneoscleral meshworks, whereas microindentation stiffness reported by Last et al. ² depended on the stiffness of JCT and inner wall of SC. Therefore, the apparent discrepancy in E observed between local compressive stiffness and tensile stiffness of whole TM indicates that pathological changes in the eye caused by glaucoma can lead to a reduction in the axial-tensile stiffness of uveal and corneoscleral meshworks and an increase in compressive stiffness of JCT/inner wall. 

TM with lower axial-tensile modulus observed in glaucomatous eyes may be less capable of preventing SC collapse or TM herniation into collector channels under elevated IOP, ¹⁹ which may lead to an increase in outflow resistance with IOP elevation. It is also known that flow resistance is highly sensitive to tissue deformation, ^20–23 which in turn depends on pressure gradient in perfused tissues. In eyes with low C due to high flow resistance in TM, the percent drop of pressure across TM is relatively larger than that in eyes with high C. The larger pressure drop can lead to larger TM deformation. Thus, the sensitivity of outflow resistance to change in perfusion pressure is expected to be higher in eyes with low C. To demonstrate this phenomenon, we plotted C measured at a given pressure versus the logarithmically transformed variance of outflow resistance, log₁₀(VAR), measured at four different pressures. Inverse correlations were observed for both glaucomatous and normal eyes ¹ (see Fig. 5), suggesting that it is more likely to observe a larger Var(R) in an eye if its C measured at a given pressure (e.g., 20 mm Hg) is relatively lower than other eyes, assuming that other tissue properties and experimental conditions are similar for all eyes. 

The circumferential tensile stiffness of TM was inversely correlated with cross-sectional area of the tissue when all data were considered in the analysis, although the correlation became insignificant when the outlier data were removed (see Fig. 4). This correlation might be attributable to the heterogeneity in local tissue stiffness, ^2,24 and potentially segmental variation in TM tissue structures. ²⁵ Glaucomatous TM is known to have decreased cellularity in corneoscleral meshworks. ⁸ A recent study has also shown fusion of trabecular beams, due to the loss of TM cells, in primary angle closure glaucoma. ²⁴ The beam fusion may also happen in POAG since both diseases involve cell loss in the corneoscleral meshwork, ⁸ which may increase stiffness heterogeneity in the corneoscleral meshwork, and cause the dependence of stiffness on cross-sectional area. Further studies are necessary to determine how the tensile stiffness of TM is altered due to tissue structural changes. 

Outflow facility decreased more with pressure elevation in glaucomatous eyes than normal eyes (Fig. 3B). It has been shown in enucleated eyes that this outflow facility reduction with pressure elevation is likely to be due to collapse of SC. ^19,26 In severe cases, it may cause herniation of TM into collector channels and occlusion of the ostia. ^19,26 Herniation and SC collapse could be more prevalent in glaucomatous eyes because in these eyes, the pressure gradient across TM is relatively large, due to low outflow facility, ²⁷ and the circumferential tensile stiffness of TM is relatively small (see Fig. 3A). In fact, the data reported in Figure 3B strongly suggested that SC collapse/herniation had happened during the perfusion of glaucomatous eyes, as the mathematical model developed by Johnson and Kamm ²⁷ had predicted that outflow facility would decrease minimally with increasing IOP (see the data for normal eyes in Fig. 3B) unless the height of SC was reduced by more than 90%. 

Cellularity in Glaucomatous TM

To our knowledge, this is the first time that TM cellularity and outflow facility of eye have shown significant correlations. This new finding could be attributed to a new approach to determining cellularity. In previous studies, the cellularity was assessed based on cell count in histological sections prepared with meridional or radial cuts of TM, ^8,9,28 whereas in the current study, the number of cells was counted in a volume of TM tissue. As a result, the data are less sensitive to cell density variation within a TM. Similar to the previous studies, ⁹ we found that the cellularity in TM was relatively lower in older donors, even though this difference was statistically insignificant. In addition to aging, TM cellularity can be decreased by diseases, exposure to cytotoxic agents, and lack of nutrient supply in underperfused regions, ²⁹ which may contribute to nonuniform stiffness distribution along the TM because of regional variation in outflow. ^2,25  

Potential Mechanisms of Stiffness Change in Glaucomatous TM

As mentioned previously, TM cells may play an important role in regulating tissue stiffness through control of ECM turnover, ⁷ which is mediated by matrix metalloproteinases (MMPs) and tissue inhibitor metalloproteinases (TIMPs). ^7,30 Specifically, it has been shown that MMPs can increase outflow facility in an organ perfusion model, ³¹ and TIMP-mediated inhibition of endogenous MMPs may reduce outflow facility. Synthesis of MMPs can be stimulated by IOP elevation, ³² mechanical stretch, ^33–38 and glaucoma drugs. ^39,40 Donor eyes used in the present study had been treated with various glaucoma drugs (see Table 2), which may alter MMP and TIMP syntheses in ocular tissues. ³⁹ Specifically, latanoprost, brimonidine, and nipradilol have been shown to increase MMP-3 and decrease TIMP-3 expressions. ³⁹ Latanoprost can also increase MMP-1, MMP-2, and MMP-3 levels in monkey ciliary muscle. ⁴⁰ Mechanical stretch of TM cells has been shown to alter expression of many ECM-related genes, ³³ including an MMP-2 gene whose expression is increased in stretched cells. ^34,35,37 Increased levels of MMPs in the anterior segment of the eye could accelerate ECM degradation and initiate ECM remodeling. 

MMPs and TIMPs alone are insufficient to maintain the homeostasis of ECM in TM, as it also requires synthesis of new ECM by TM cells. However, during the progression of glaucoma ⁸ and with aging, the number of TM cells is decreased, ^9,28 especially in the corneoscleral meshwork. ⁸ Therefore, the low tensile stiffness of TM observed in glaucomatous eyes might be caused by a combined effect of high MMPs levels in aqueous humor and low cellularity in TM, which can lead to a significant decrease in ECM density in the uveal and corneoscleral meshworks. Future studies with a larger sample size are needed to confirm the findings. 

Two more factors, which are not directly related to glaucoma, may contribute to the differences in the data compared in this study between normal and glaucomatous tissue samples. The first is cataract surgery. In this study, 7 of the 8 glaucomatous donors were pseudophakic, whereas all the healthy donors were phakic. ¹ Although the direct effects of cataract surgery on the trabecular outflow pathway are unclear, it has been well characterized that the surgery decreases IOP ⁴¹ and may increase outflow facility. ⁴² The removal of crystalline lens may affect molecules that are released into the aqueous humor, potentially affecting the homeostasis in the outflow pathway. Additionally, the ciliary muscle tone may change after cataract surgery, which affects the dynamic structure of TM via the scleral spur. Last, the phacoemulsification process may cause changes in the TM, potentially by eliciting immune responses. The second factor is donor age. The glaucomatous eyes came from donors that on average were 20 years older than the healthy donors. Although there was no significant correlation between age and stiffness for either group, TM of older patients may have different responses to mechanical stress or pharmacological agents that can alter ECM remodeling, potentially due to increase in senescent cells. ⁴³ Further studies must be performed to better elucidate mechanisms that can cause differences in the Young's modulus between normal and glaucomatous TM. 

Conclusions

The circumferential Young's modulus of the glaucomatous TM was lower and more heterogeneous than that of normal TM determined in a previous study. Additionally, a positive correlation was observed between outflow facility measured at 30 and 40 mm Hg and TM cellularity; and higher outflow facility was associated with less variance in outflow resistance in response to pressure elevation. These data imply that the reduction in outflow facility observed in glaucoma patients is related to reduction in both circumferential tensile stiffness and cellularity of TM. 

Acknowledgments

We thank Joseph Izatt for the use of OCT, Hansford Hendargo for the use of OCT analysis software, Jin Liang for assistance in cellularity measurement, and Jianyong Huang for assistance in confocal microscopy. Additionally, we thank Ivantis for the gift of the donor eyes. 

Disclosure: L.J. Camras, Ivantis (F); W.D. Stamer, None; D. Epstein, None; P. Gonzalez, None; F. Yuan, None