Circumferential Trabecular Meshwork Cell Density in the Human Eye

Abstract

The cells residing in the trabecular meshwork (TM) fulfill important roles in the maintenance of the tissue and the regulation of intraocular pressure (IOP). Here we examine (i) TM cell distribution along the circumference of the human eye, (ii) differences in TM cell density between regions of high and low outflow, and (iii) whether TM cell distribution in eyes from donors with primary open angle glaucoma (POAG) differs from that of normal eyes. Toward this end, the TM cell density from 12 radial segments around the circumference of the TM of human donor eyes (n=6) with and without POAG was determined using histochemical methods. Areas of high, median, and low outflow were mapped in a different set of human donor eyes that were perfused in organ culture, and TM cell densities in these areas were determined in normal (n=11) and POAG eyes (n=6).

Our analysis of 1,380 tissue sections taken from the first set of six eyes shows that the average TM cell density of these six eyes ranges from 15.5 to 23.7 cells/100 μm and is negatively correlated to the maximum IOP recorded for each donor eye (R²=0.91). Considerable differences in TM cell density exist among sections taken from the same segment of an individual eye (average standard deviation=2.35 cells/100 μm). Less variability is observed among the segment averages across the eye’s circumference (average standard deviation = 1.03 cells/100 μm). Variations in cell density are similar between normal and POAG eyes and are not correlated with the anatomic position of examined segments (p=0.745). The analysis of the second set of eyes shows that TM regions of high outflow display a TM cell density similar to regions of median or low outflow in both normal and POAG eyes. Together these findings demonstrate that (i) statistically significant differences in TM cell density exist along the circumference of each eye (ii) TM cellularity is not correlated with segmental flow and (iii) eyes with POAG, while displaying reduced TM cellularity, do not exhibit higher TM cell variability than normal eyes. Finally, statistical analysis of sections and segments indicates that measurements from 12 sections taken from 2 segments provide a reliable and cost-effective estimate of a human eye’s TM cell density.

1. Introduction

The aqueous humor inside the eye is continuously replaced and drained through the trabecular meshwork (TM), a specialized tissue of the eye that controls the outflow rate (1). In a healthy eye aqueous humor outflow rates are regulated to maintain a modest intraocular pressure (IOP) that ensures proper inflation of the globe but is also tolerated by the cells of the neural retina. The number of TM cells decreases with age and is particularly low in individuals with Primary Open Angle Glaucoma (POAG) or other types of open angle glaucoma (2–7). The reason for the observed age-related loss of TM cells remains a topic of active investigation, but appears to involve genetic factors, cell stress, and senescence of the TM (8–10).

The functional consequences of reduced TM cellularity have not been established, but may involve insufficient turnover and maintenance of the TM. This, in turn, may be the underlying cause for the biomechanical changes observed in aging eyes or those with glaucoma, including stiffening of the TM (11–13), which may result in decreased ability of the eye to regulate aqueous humor outflow (14). The functional importance of sustaining a sufficient number of TM cells to maintain intraocular pressure (IOP) is further underscored by the ability of the TM of older eyes or those with glaucoma to regain some function when the number of TM cells is increased. Early efforts to achieve this have relied on the use of laser trabeculoplasty to cause a proliferative response in the TM resulting in a reduction of IOP in nonhuman primates (15) and organ cultured human eyes (16, 17). More recent studies have demonstrated that transplantation of TM cells derived from induced pluripotent stem cells results in increased TM cell density, increased aqueous humor outflow facility, and lower IOP in human (18, 19) and mouse eyes (20, 21).

Examination of the TM in histologic tissue sections of human eyes often reveals considerable differences in the density of TM cells, even when multiple sections are taken from the same eye. It is conceivable that the distribution of TM cells differs between segments of the eye or exhibits regional differences, e.g. nasal vs. temporal. Such anatomical differences have been reported by some investigators (4, 22), while others did not observe them (3, 23).

These discrepancies may be due to the limited number of sections evaluated. Furthermore, no studies have been reported that examine the distribution of TM cell density in eyes with POAG. This disease is frequently accompanied with an elevation of IOP, suggesting TM dysfunction. However, many POAG patients present with IOP within the normal population range (≤21 mmHg) and presumably normal TM function. In this study we carried out extensive sampling of tissue section derived from six human eyes from older tissue donors including eyes with POAG and clinically documented elevated IOP.

2. Materials and Methods

2.1. Donors:

Morphometric Study:

Six human donor eyes were obtained from the Iowa Lions Eye Bank (Iowa City, IA) with written consent of the next of kin and preserved within 6 hours post-mortem (average=5.2 hours). Donors ranged in age from 60–84 years of age and eyes from both male and female donors were included (Table 1). Review of clinical records indicated that three of the donors were free of glaucoma and never presented with intraocular pressure (IOP) above 21 mmHg. Three additional eyes were chosen from individuals with a clinical diagnosis of primary open angle glaucoma (POAG). None of the eyes had undergone anterior segment surgeries. It must be noted that all POAG donors had received IOP lowering medical treatment. Eyes with other ocular disease (e.g. age-related macular degeneration or diabetic retinopathy) were excluded.

Table 1:

Characteristics of donor eyes used for the morphometric study.

Donor	PM Time (Hours)	Age (Years)	Sex	Glaucoma	Max IOP (mmHg)	Glaucoma Medications
1	6:13	67	M	No	17	None
2	5:29	60	F	No	15	None
3	2:40	64	M	No	21	None
4	5:11	84	F	Yes	19	Timolol
5	5:44	67	F	Yes	23	Cosopt, Travoprost
6	5:37	81	F	Yes	24	Brimonidine, Latanoprost

Glaucoma medication and maximum IOP are listed as reported in patients’ charts. PM Time: post mortem interval to tissue fixation

Ocular Perfusion study:

A separate set of 17 eyes from 10 donors was used for the perfusion study (Table 2). The use of these eyes was approved by the Oregon Health & Science University Institutional Review Board and experiments were conducted in accordance with the tenets of the Declaration of Helsinki for the use of human tissue. Eyes, including information regarding the glaucoma status, were obtained from Lions VisionGift, Portland, OR, USA. Normal donors were = 82.8 +/− 6 years of age and glaucoma donors were = 82 +/− 8 years of age. The length of time from death to the establishment of stationary culture was less than 48 hours.

Table 2:

Characteristics of donor eyes used for the perfusion culture.

Donor	Eye used	Age (Years)	Sex	Glaucoma
1	OS and OD	84	F	No
2	OS and OD	78	M	No
3	OS and OD	93	M	No
4	OD	79	F	No
5	OS and OD	77	F	No
6	OS and OD	86	M	No
7	OD	76	M	Yes
8	OS and OD	93	M	Yes
9	OS and OD	85	M	Yes
10	OS	74	F	Yes

2.2. Sample preparation (morphometric study):

The anterior segment was dissected and eyes were fixed in 4% paraformaldehyde. After 4 hours in fixative tissue was transferred to storage buffer (1% paraformaldehyde in phosphate buffered saline). Prior to experimentation, lenses were removed and the orientation of the anterior segment was determined using the retinal arcades, fovea and optic nerve head of the posterior segment as guides. The anterior segment was then dissected into 12 approximately evenly sized pie-shaped segments (Figure 1A) in such a manner that segment #1 always represents the TM located superiorly, segment #4 the temporal quadrant, segment #7 the inferior quadrant, and segment #10 the nasal quadrant.

Figure 1A:

Schematic representation of the dissection scheme used in this study. The donor eyes were divided into twelve equivalent segments (1–12). From each of these 20 tissue sections were obtained for analysis. Segment 1 always represents the superior portion of the eye, whereas segment 10 always represents the nasal portion. Here, a left eye is depicted. In right eyes the nasal quadrant is located on the opposite side and segments would be numbered in a counter-clockwise fashion. (B) Radial section of the TM of a normal human eye displaying mildly reduced TM cellularity. Arrow indicates TM cell nuclei, whereas the arrowhead indicates pigmented debris. The outlined area denotes the analyzed region. TM= sclera, CL=corneal limbus, IR=iris, SC=Schlemm’s Canal, Scalebar = 50 μm

Each wedge was processed on a STP 120 spin tissue processor (Thermo Fisher Scientific, Waltham, MA) and embedded in paraffin using a Histostar embedding module (Thermo Fisher Scientific). 20 semi-consecutive 5 μm thick radial sections were obtained from each block using a Microm HM310 microtome (Thermo Fisher Scientific). The sections were stained using Surgipath Harris Hematoxylin (Leica, Buffalo Grove, IL) and Eosin-Y Alcoholic (Anatech Ltd, Battle Creek, MI). Photomicrographs were obtained from each section using a Zeiss Observer Z1 microscope (White Plains, NY) at 20x magnification.

In this study, we defined ‘trabecular meshwork’ as tissue located directly uveal to Schlemm’s canal. In these radial sections Schlemm’s canal was assumed to be oriented in an anterior to posterior orientation and all nuclei in the region between the visible anterior and posterior margin of Schlemm’s were counted (Figure 1B). The width of Schlemm’s varies throughout the eye and cell counts were normalized to a unit length representing the number of cells per 100 μm radial length of Schlemm’s canal. Nuclei of iris cells or Schlemm’s canal endothelial cells were not included. It was assumed that each nucleus represents one cell and that no cell will have more than one nucleus. Cells were counted manually using the ImageJ plugin Cell Counter.

2.4. Statistical Evaluation:

Averages and standard deviations were calculated using Prism 8 (GraphPad Software, San Diego, CA). Unless otherwise noted, cell count values are given as Average +/− Standard Deviation normalized to 100 μm length of Schlemm’s canal.

Strategies for estimating the TM cell density of human eyes and for selecting sample sizes of eyes in comparative studies were determined from estimates in the mixed effects model (24, 25) described in equation 1,

$$Y_{gijk}=\mu +{\delta }_g+{\alpha }_{i(g)}+{\beta }_{j(g,i)}+{\epsilon }_{k(g,i,j)}$$ (1)

The model includes fixed group effects δ_g (g=2 groups: normal/glaucoma) and three variance components: $Var(\alpha )={\sigma }_{\alpha }^2$ for eye, $Var(\beta )={\sigma }_{\beta }^2$ for segment, and $Var(\epsilon )={\sigma }_{\epsilon }^2$ for section. The parentheses in the subscripts of the model components reflect the fact that factors are nested within each other: g =1,2 for the two groups; i =1,2,3 for the 3 eyes in each group; j =1,2,…,12 for the 12 segments in each eye; and k = 1,2,…,20 for the 20 sections in each segment. Restricted maximum likelihood estimates of the parameters in equation 1 were calculated using the PROC MIXED package within the SAS software, version 14.1 (Cary, NC, USA). A computer program, written in R Statistical Software code, used these estimates to determine the numbers of segments and sections needed to attain the specified estimation precision. The statistical significance of the difference in cell distribution between normal and glaucoma eyes relies on the robust “sandwich” estimator of the variance-covariance matrix of the two fixed-group effects, as described in (25).

2.5. Labeling and separation of segmental flow regions (perfusion study):

The anterior portion of each eye was initially placed into serum-free stationary organ culture for 5–7 days to facilitate postmortem recovery (26). After stationary culture, the tissue was perfused with serum-free Dulbecco’s Modified Eagle’s Medium (a 1:1 mixture of high and low glucose DMEM) containing 1% Penicillin/Streptomycin/Fungizone, at constant pressure of 8.8 mmHg, which is similar to normal physiological rate and pressures (minus episcleral venous pressure) in vivo (27). All eyes were maintained at 8.8 mmHg and perfusion flow rates, ranging from 1–9 μl/min, were recorded for the duration of the perfusion time.

After starting perfusion baseline flow rates are allowed to stabilize (approximately 48 hours), and the outflow facility of each eye is determined to ensure the successful establishment of the organ culture. The average outflow facility of the normal eyes was slightly higher than the average outflow facility of the glaucomatous eyes, but this difference was not statistically significant (Appendix A). Fluorescently-labeled amine-modified 200nm Fluospheres (red) (Thermo Fisher Scientific, Waltham, MA, USA) are used to map outflow regions. They were diluted 1:500 into phosphate buffered saline, vortexed vigorously, and 200μl of that mixture was injected as a bolus directly into the organ culture and perfused for 1 hour. This time was adjusted respective to flow rates to ensure that an equal volume (approximately 100 μl) of this mixture was perfused for each eye (28). Tissues were removed from organ culture and imaged en face using a Leica DM500 microscope (Leica Microsystems, Buffalo Grove, IL, USA). Relative fluorescence intensity was mapped across flow regions of the TM and plotted against the circumferential distance using Image J software as documented previously (28, 29). This output consists of approximately 10,000 data points around the circumference of the TM for each eye and the total RFU is used to normalize the intensity on a per eye basis. High flow (HF) regions are defined as those displaying the top 1/3 of normalized RFU intensity, (LF) Low flow regions are the bottom 1/3 of RFU intensity, and intermediate flow (MF) regions are those displaying RFU values in between.

HF, MF, and LF flow regions were separated by cutting tissues into radial wedges between the different regions. Frontal sections were then cut with a single-edged razor blade perpendicular to the ocular surface, resulting in a section tangential to the corneoscleral limbus that bisects Schlemm’s canal as described previously (29).

2.6. Imaging, cell counting, and statistical analyses (perfusion study):

Tissue wedges were placed on 0.17 mm Delta T cover glass bottom culture dishes from Bioptechs Inc. (Butler, PA) in Slowfade Gold antifade reagent with DAPI (Invitrogen), and imaged by confocal microscopy using an Olympus FV1000 microscope. Optical sections were acquired using sequential scanning in separate laser channels. Image acquisition settings and number of optical sections in a stack were kept constant when comparing images.

DAPI-stained nuclei were counted in each image (representing 100 μm × 150 μm) and averaged for each region from each eye. Overall cell densities in the groups of normal and glaucomatous eyes were evaluate using two-tailed, unpaired t-test. Differences in cell numbers between the same regions comparing normal to glaucomatous tissues were measured for statistical significance using Brown-Forsythe and Welsh ANOVA with Dunnett’s multiple comparisons tests. Levene’s test was used to determine if normal and glaucomatous samples have equal variances.

3. Results

In this study, we obtained 20 sections from each of the 12 segments (240 sections/eye) of three normal eyes and three eyes with clinically documented POAG (Table 1). One exception is eye 3 for which the tissue in segments 1, 2, and 3 appeared to be disrupted, perhaps during sample preparation, and was unsuitable for analysis. Radial sections were prepared in a semi-consecutive fashion, i.e. sections were cut sequentially, but those displaying artifacts were discarded. On each section the number of cells in the TM overlying Schlemm’s canal was manually counted. Since the width of Schlemm’s varies throughout the eye, cell counts were then normalized to a unit length (cells per 100 μm, Figure 1B). Among the three normal eyes few morphological differences in the appearance of the TM were observed. In contrast, the morphologic appearance of the TM varied noticeably between the POAG eyes, likely reflecting different disease states among the donors. While eye 4 displayed relatively normal TM morphology, eye 6 revealed classic hallmarks of glaucomatous degeneration, including thickened and partially fused TM beams, decreased intra-trabecular spaces, and visibly reduced TM cellularity.

We then determined the overall TM cell density of each eye, based upon all sections. In the normal eyes these ranged from 20.6 ±2.41 cells/100 μm (eye 3), to 23.2 ±2.62 cells/100 μm (eye 1), and 23.7 ±2.66 cells/100 μm (eye 2) and in POAG eyes from 15.5 ±2.66 cells/100 μm (eye 6) to 18.2 ±2.53 cells/100 μm (eye 5) and 21.7 ±2.54 cells/100 μm (eye 4) (Figure 2, Table 3). The eyes of POAG donors contained fewer TM cells overall than those of normal eyes (18.5±3.10 cells/100 vs. 22.5±1.65 cells/100 μm, p=0.074) indicating an average cell loss of 17.7% in glaucoma eyes (Table 3). The overall variability in TM cell density (SD Total) did not significantly differ between the six eyes (p=0.745, Levene test) or between normal and glaucoma eyes (p= 0.887, Levene test).

Figure 2:

Cellularity in each segment of the study eyes. A, B, C: Normal eyes, D, E, F: POAG eyes. Values are given as cells/100 μm. Section measurements are represented by dots. The height of the bar represents their average. Error bars denote standard deviation.

Table 3:

Trabecular Meshwork (TM) densities in the six donor eyes used in the morphometric study.

Eye	POAG	Average	Maximum	Minimum	SD Total	SD due to Segment	SD due to Section
1	No	23.2	39.7	16.5	2.62	1.14	2.30
2	No	23.7	35.9	16.9	2.66	1.29	2.29
3	No	20.6	27.9	15.6	2.41	0.94	2.23
4	Yes	21.7	31.2	14.7	2.54	1.26	2.25
5	Yes	18.2	25.3	11.3	2.53	0.74	2.44
6	Yes	15.5	23.8	9.2	2.66	0.81	2.57

Measurements are expressed as number of cells per 100 μm. SD Total represents the standard deviation of section measurements around their eye average. SD due to Segment denotes the SD of the 12 segment averages. SD due to Section averages the SDs of the 12 segments, with each segment SD measuring the variability of its 20 section measurements.

There is considerable variability across the section measurements that are taken on each eye. There are two sources for this variability: (i) variability due to segments and (ii) variability due to sections within each segment. Of these, variability due to sections is the larger component and ranges from 2.23 cells/100 μm (eye 3) to 2.57 cells/100 μm (eye 6). The variability due to segment is the smaller component and ranges from 0.74 cells/100 μm (eye 5) to 1.29 cells/100 μm (eye 2). Neither the variability due to segment nor the variability due to section differs between normal and glaucoma eye (p= 0.385 and 0.199, respectively).

Although differences among the segment means are considerably smaller than differences among section measurements within the same segment, these findings do not imply that the TM density across the 12 segments is uniform. Results of one-way ANOVAs, testing the equality of the 12 segment averages within each studied eye, indicate statistically significant differences (eye 1: p= 0.000, eye 2: p=0.000, eye 3: p=0.001 eye 4: p=0.000, eye 5: p=0.065 and eye 6: p=0.038).

We further evaluated whether the anatomical locations of segments with relative high or low cell densities are conserved among all eyes, e.g. whether nasal TM always has a higher cell density than temporal TM. In order to directly compare findings from eyes with divergent overall cellular densities, segmental data was converted to represent a percentage of that eye’s overall average TM cell density. Comparison of data from both normal (Figure 3A) and POAG eyes (Figure 3B) did not indicate a consistent pattern among the eyes suggesting that the observed differences among segments of each eye are random.

Figure 3.

Normalized cellularity of (A) normal and (B) POAG eyes. The height of each bar reflects the proportional difference in cellularity of the segment relative to the overall cell density average of the eye.

Our morphometric study includes eyes with a considerable range of maximum recorded IOP, allowing us to investigate the correlation between TM cellularity and IOP. Among the eyes examined here, TM cellularity was highly correlated with the maximum recorded IOP of the donor, irrespective of the presence of POAG (R²= 0.91, Figure 4).

Figure 4:

(A) TM cell density (cells/100 μm) in the six examined eyes correlated with the highest IOP clinically observed in the donor (R²=0.91).

Functional analyses have demonstrated that outflow of aqueous humor through the TM is not uniform throughout the circumference of the eye and that regions of high and low flow exist in all eyes (30, 31). One explanation for segmental differences in outflow could be that regions with high outflow rates contain many TM cells or conversely that regions with low flow rates are sparsely populated. We investigated this question using a second set of eyes maintained in perfusion culture consisting of eleven normal and six glaucomatous eyes (Table 2). This established technique can be employed to study outflow dynamics ex vivo (27) and has been used extensively in our previously published work (29, 32, 33). High (HF), intermediate (MF), and low (LF) flow regions can be mapped following infusion of fluorescently labeled Fluospheres (Figure 5A). Using these criteria we mapped flow regions in 11 normal eyes (from 6 donors) and 6 glaucomatous eyes (from 4 donors). We then determined the average number of DAPI-stained nuclei in the HF, MF, and LF regions of each eye by confocal microscopy (Figure 5B, Table 4). Eyes obtained from normal donors contain on average 19.15 cells per image (ranging from 10 to 35, SD=5.23 cells/image). POAG eyes contained significantly fewer cells with an average of 11.06 cells per image (ranging from 3 to 25 cells, SD=5.06 cells/image, and p<0.0001). However, the SDs are similar between both groups, supporting our morphometric finding that variability in TM density does not differ between normal and POAG eyes (p=0.858, Levene test). A significant reduction in cellularity in POAG eyes is also apparent when HF regions are compared between normal and glaucomatous tissues (p= 0.030) (Figure 5B). This is also true for the comparison between MF (p=0.015) and LF (p = 0.009) regions of normal and glaucomatous tissues. However, within each group of donor eyes differences in the cell numbers between the HF, MF, and LF regions are not statistically significant. Finally, our data also indicate that variability in cell density does not differ between HF, MF, and LF regions (p>0.999, Levene test). These findings demonstrate that TM cell density and segmental flow are not correlated in either normal or in glaucomatous eyes.

Figure 5:

Cell number variability in normal (n=11) and glaucomatous (n=6) segmental regions. (A) En face fluorescence images identifying high (HF), intermediate (MF), and low flow (LF) regions of the TM following perfusion with fluorescent microbeads in a single human donor eye. (B) Average number of DAPI-stained nuclei in segmental flow regions from normal (“Norm”) and glaucomatous (“Gl”) eyes. A statistically significant reduction of cellularity is observed when the same flow regions are compared between normal to glaucomatous tissues. “*” p = 0.030, “**” p = 0.015, “***” p = 0.009 (ANOVA with Dunnett’s multiple comparisons test).

Table 4:

Trabecular Meshwork (TM) cell densities in the donor eyes used in the perfusion study.

	Average	Maximum	Minimum	Range	SD
Normals -Overall	19.15	35.0	10.0	25	5.23
POAG-Overall	11.06	25.0	3.0	22	5.06
Normals-HF	20.55	28.0	15.0	13	4.03
POAG-HF	9.75	13.0	7.0	6.0	2.5
Normals -MF	17.14	22.0	11.0	11	3.98
POAG-MF	11.25	16.0	5.0	11	4.65
Normal -LF	20.50	35.0	10.0	25	7.07
POAG-LF	11.69	25.0	4.0	21	6.92

Measurements are expressed as number of cells visible per microscopy field. SD Total represents the standard deviation of measurements.

One goal of this study was to determine the number of sections and segments that should be analyzed to achieve an accurate estimate of the TM cell density in human eyes. Studies frequently aim to achieve a standard deviation for the measurement error that is within 5 % (or ≤ 1 cell/100 μm) of the target, defined here as the average TM cellularity across all six eyes (roughly 20 cells/100 μm). Several sampling schemes with different combinations of the number of segments and sections can be used to achieve equally accurate results. The least labor intensive and consequently most cost effective of these requires using 12 sections obtained from 2 segments. A technical report detailing these considerations is included as appendix B.

Furthermore, many research projects compare the mean TM cell density between two groups of samples, as we did here with normal and glaucomatous eyes. This introduces differences between individual donor eyes as a third, considerable, source of variation that can be incorporated into the mixed-effects model (34). Based upon measurements obtained in this study, an experiment designed to detect a 10% reduction in the cell density of normal eyes (here 22.5 cells/100 μm from Table 3) with 80 percent power requires n=9 eyes of the experimental group.

Discussion

In this study we evaluated the circumferential distribution of TM cells in human eyes with and without elevated IOP and POAG in order to determine whether regional differences in cell density exist within the tissue. Toward this end, we analyzed 1,380 tissue sections representing 12 areas from 6 human donor eyes. Our findings indicate that in the human TM cell density varies noticeably between tissue sections, especially in eyes with higher IOP, and to a lesser degree throughout the circumference of the eye. However, sizable differences in cellular density among sections do exist, amounting to 10–15% of the average TM cell density of the eye. While fluctuations in cell density are observed in all eyes, the anatomical location of regions with relatively high or low density varies among eyes, indicating that these differences are not a consistent anatomic feature of the TM.

Several large glaucoma trials have demonstrated that maximum IOP is an important clinical measure with strong predictive value of glaucoma progression in patients. In the “Los Angeles Latino Eye Study” maximum IOP was found to be the most consistent IOP measure for predicting glaucoma risk (35). Likewise, the “Collaborative Initial Glaucoma Treatment Study” found maximum IOP to be more predictive of visual field loss than mean IOP, particularly for glaucoma patients treated medically (36). Among the eyes examined here we observed a strong inverse correlation between TM cell density and the highest recorded IOP of the eye donor, supportive of the notion that reduced TM cell density hampers the ability of the tissue to regulate IOP.

Aqueous humor outflow varies considerably among various areas of the TM and regions of ‘high flow’ and ‘low flow’ have been described (29, 37, 38). The molecular mechanisms behind these findings are being investigated, and it has also not yet been firmly established whether the respective regions are static or change throughout life. It has been estimated that only about one third of the TM is actively involved in aqueous humor outflow at any given time (39). It is conceivable that high segmental outflow is associated with regions of higher cellularity, either due to a direct effect of the cells on TM contraction and relaxation, or indirectly, since areas of lower cell density may not be able to adequately maintain TM homeostasis. However, our analysis of TM cell density in high, intermediate, and low flow regions did not find noticeable differences between these areas in either normal or glaucomatous eyes. Consequently, we did not find evidence to support the hypothesis that high flow TM regions are characterized by higher TM cell density or conversely that low flow TM regions lack TM cells. Outflow facility, which is a functional measure of TM cell activity and viability in tissues, was determined for all perfused eyes. We observed a slightly lower facility in glaucomatous compared to normal eyes, but this difference was not statistically significant. No correlation was found between TM cell density and outflow facility.

The decrease in TM cellularity in POAG eyes has been extensively examined and our morphometric study was not intended to repeat these findings (2, 4, 23). Rather, POAG eyes were included in order to gain a broader IOP range among the samples, as well as to gain an indication whether TM distribution is inherently altered in the disease. Elevated IOP and POAG are not synonymous and eyes were chosen to include both a normal eye with a highest recorded IOP at the top of the normal range (eye 3, max. IOP=21 mmHg) and a POAG eye with unremarkable pressure (eye 4, max. IOP=19 mmHg). The density and distribution of TM cells in eye 4 is very similar to those observed in the normal eyes, suggesting that TM function was normal despite the donor’s POAG. We did observe a slight increase in TM cell variability among sections taken from eyes with higher IOP but not among segments, even in eyes with substantially reduced cellularity. Similar data were also obtained in perfusion-cultured eyes where variation in TM density was similar between normal and POAG eyes, despite clearly reduced overall cell densities in the latter group. This indicates that the health of the TM cells in POAG does not dramatically decline in distinct regions which could result in sharp, but local, reductions in aqueous humor outflow. Together these findings lead us to speculate that the factors leading to TM cell loss in POAG act uniformly throughout the eye rather than leading to a dramatic decrease of TM density in restricted regions.

We intentionally generated a large number of segments and sections so that we could determine cost-effective sampling strategies and the resulting required sample sizes for group comparisons of human eyes. The reported sampling strategies for estimating TM cellularity varied considerably across laboratories. Many laboratories evaluated 3 sections from 4 segments, for a total of 12 sections (4, 22). Others evaluated as few as 3 sections from 2 segments, for a total of 6 sections (40) or as many as 12 sections from 2 segments, for a total of 24 sections (18). The rationale for choosing these sampling protocols appears to reflect the investigators’ intuitive assessment of measurement variability among segments and sections. Our data demonstrate that the differences between sections from the same segment are a larger source of error than those between segment averages. As a consequence, it is more effective to generate a sufficient number of sections rather than sampling many segments. Furthermore, the absence of conserved regional variation in the distribution of cells throughout the TM indicates that the site from which the sample is obtained is of lesser concern. This facilitates sample preparation and permits the use of donor tissue in which the orientation of the eye can no longer be determined.

While we generated ample data to support our findings of TM cell distribution, the relatively small number of eyes used in our morphometric study is a limitation of this study. Diseased tissue frequently presents with a wide range of phenotypes and it is possible that some eyes with POAG exhibit changes not observed here. Furthermore, unlike the normal donors, all POAG donors in the morphometric study received IOP lowering medications, were female and on average slightly older than the donors of the normal eyes. While differences between normal and POAG eyes were not the focus of our study, expanding the analysis to additional eyes may provide further insights. The design of such studies would be aided by the sampling schemes established herein.

Conclusions

In this study we evaluated intra eye differences in the distribution of TM cell within the tissue. We found noticeably higher variation in TM cell density among semi-consecutive tissue sections than among segment averages. While TM cell densities within each eye varied between 10 and 15% the anatomic location of higher or density segments is not conserved among eyes. Finally, we did not detect differences in cell density between regions of high and low outflow, even in eyes with POAG. These data show that TM cellularity is not correlated with segmental flow and indicates that mechanisms other than cell density are instrumental in the regulation of local aqueous humor outflow.

• We analyzed the circumferential density of trabecular meshwork cells in human eyes

• TM regions of high or low aqueous humor outflow exhibit similar cell densities

• The location of areas of relative high and low TM cell density are not conserved among eyes

• Eyes with POAG exhibit similar variability in cell density than control eyes