The trabecular meshwork represents a location of great interest for glaucoma research, and non-invasive methods are needed to decellularize it and convert it into a scaffold for cellular transplantation. As porcine eyes share several important features with human eyes ^{1– 3} , can be used to mimic surgical interventions ^{4, 5} and different glaucoma types, including secondary open angle ⁶ and angle closure glaucomas ⁷ , they have been widely used as ex vivo models in glaucoma research. In contrast to human donor eyes, they are also ubiquitously available and remain responsive to stimuli and drugs when freshly harvested. This allows the exploration of anatomical functions ^{3, 8} and the effects of biologicals ⁹ , drugs ¹⁰ , and surgical interventions ¹¹ . Corresponding outflow changes can be studied with high spatial and temporal resolution ^{5, 12, 13} .

Ab interno trabeculectomy of the nasal circumference increases outflow in whole porcine eyes as well as anterior segment cultures ⁵ . Outflow is changed mostly at the site of ablation but also enhanced circumferentially ⁵ . In contrast to ab interno trabeculectomy, freeze-thaw decellularization is a nonspecific and site-agnostic method to remove all living cells from eye cultures to ready them for the transplantation and study of cells of interest ^{14, 15} . In perfused porcine anterior segments, this is accompanied by a significant reduction in intraocular pressure (IOP) ¹⁶ . Freeze-thawed scaffolds may be easier to generate than artificial three-dimensional trabecular meshwork structures ^{17– 19} .

Here, we hypothesized that subjecting porcine eyes to freeze-thaw decellularization would result in faster outflow throughout the entire circumference and alter the outflow pattern due to a loss of endothelial integrity of the distal outflow tract. We opted to use this cytoablation technique as we have shown in previous studies that it provides a complete ablation of the trabecular meshwork effectively while keeping its cytoskeletal structure intact ¹⁶ .

A total of 28 eyes were included in the analysis consisting of 16 right eyes and 12 left eyes. There were six eyes in each of A _CO, W _CO, and W _FT, and ten eyes in A _FT. Within the control groups, A _CO and W _CO, there were no significant differences between the quadrant filling times (all p > 0.05) ²⁰ . Additionally, both dyes showed similar filling times within each quadrant ( Table 1, all p > 0.05).

Figure 1 shows the filling times of anterior segments pre- and post-freeze thaw (A _FT). In these eyes, quadrant TS tended to have a faster average filling time than the other three quadrants, both pre (270 ± 136 s) and post freeze-thaw (441 ± 371 s), but without reaching statistical significance (p > 0.05). Moreover, freeze-thaw did not cause a significant change in the filling times of any quadrant (p > 0.05).

Figure 2 depicts the filling times before and after freeze-thaw in W _FT eyes. There was no significant difference in the filling times between quadrants (p = 0.401). TS showed the shortest filling time before freeze thaw with an average of 162 ± 101 s. The filling time of this quadrant significantly increased after freeze-thaw (355 ± 175 s, p = 0.002). None of the other W _FT quadrants showed a significant change in filling times after freeze-thawing.

Figure 3 depicts the canalogram of an A _FT eye. In this eye, the post-freeze-thaw canalogram demonstrated additional temporal outflow channels that were not visible before (TI) or only partially filling (TS).

Figure 4 shows canalograms of W _FT eye pre- and post- freeze-thaw. Visually, both images show a similar filling pattern after 20 minutes of perfusion. In this case, fluorescence was captured mainly in the nasal episcleral veins.

Histological analyses demonstrated unchanged trabecular beams but without intact trabecular meshwork cells ( Figure 5). Compared to control eyes, there were no noticeable changes to the tissue surrounding the angular aqueous plexus. The lumina of this plexus, including larger, Schlemm’s canal like-segments remained intact.

The post-hoc power analysis of the filling times pre- and post- freeze-thaw for each group is presented in Table 2. In general, results revealed a low power for most of the quadrants. Within the experimental groups, the TS quadrant comparison had the highest power with values of 0.5 and 0.75 in A _FT and W _FT, respectively. All other quadrants in both groups had a power below 0.3.

We have developed a freeze-thaw protocol to decellularize anterior segments and to obtain a three-dimensional matrix to transplant modified trabecular meshwork (TM) cells or other cells under study ^{16, 21} . This approach might generate ex vivo models for genetically altered, glaucomatous TM cells or primary cells harvested in excisional ab interno trabeculectomy ^{21– 23} . Surprisingly, freeze-thawing did not alter the vascular drainage spaces qualitatively in a significant way. Rather, dyes remained mostly intravascularly within the time observed without any noticeable differences on average as detectable with these methods. Fluorescein has a molecular size of 376 Da and sulforhodamine 101 acid chloride a size of 625 Da, small enough to pass through the TM relatively quickly, yet large enough to be confined to the intravascular space for minutes ^{24– 26} . Histologic analyses also revealed a conserved TM structure ^{16, 21} . It is not unreasonable to assume here that the extracellular matrix also remains unchanged as we have previously studied the effects of freeze-thawing on this structure ¹⁶ . This preserved scaffold along with the unchanged outflow vessels suggests freeze-thawing as a technique to produce realistic seeding matrices for trabecular meshwork transplantation.

Because freeze-thaw had previously been shown to cause an IOP drop of about 30% in porcine eyes ¹⁶ , we expected to see faster outflow throughout the 360 degrees of angular aqueous plexus of both anterior segments and whole eyes. Surprisingly, this was also not the case. The only quadrant in which there was a statistically significant difference was the temporal superior quadrant of whole eyes, and we observed an increase rather than a decrease in filling time. It is possible that freeze-thawing causes relatively small changes per trabecular meshwork area and that our study missed those due to the large standard deviation. However, these small changes might have a large cumulative impact and lower IOP ¹⁶ because they occur over the entire circumference. For instance, flow in TI and NS had a lower average flow time after freeze-thaw compared to before, but without reaching statistical significance. Moreover, due to their moderate molecular size, movement of fluorescein and sulforhodamine through an intact trabecular meshwork might not be fundamentally different from movement through remnants of disrupted trabecular meshwork cells or their debris ²⁷ . Additionally, our methods might miss diffuse fluid movements directly through the sclera that is akin to uveoscleral outflow.

We observed a longer post freeze-thaw filling time in the temporal superior quadrant of whole eyes. It is likely that intraocular tissues that are present in whole eyes, but are missing in anterior segments ²⁸ , were disrupted by the freeze-thawing cycles. Their debris might have caused the delayed filling ²⁷ , in particular iris pigment ^{6, 29, 30} . Pigment can lead to a slow increase of IOP over hours to days ⁶ while larger debris from cell death can have a more immediate effect on the trabecular meshwork ²⁷ . However, previous research has shown successful repopulatization after using a saponin detergent for decellularization ³¹ . Therefore, we assume a similar outcome after a less invasive cell depletion by our freeze-thaw method.

One limitation is that we did not measure IOP because a goal of this study was to assess freeze-thawing in anterior segment cultures in comparison to whole eyes; however, the latter cannot easily be connected to a pressure transducer. Another limitation is that the eyes were not incubated for three days as we have done previously ¹³ . Due to the compression mount, outflow in anterior segments perfusion cultures is initially impaired, but this effect vanishes within less than three days of culture ¹⁶ . Finally, the 28 eyes used in this study were only able to provide a moderately low testing power to detect an outflow difference, providing further evidence that any flow difference caused must not be large.

In conclusion, we found no major outflow differences caused by freeze-thaw treatment of anterior segments or whole eyes. These results validated freeze-thawing as a method to generate a three-dimensional seeding matrix without losing distal outflow tract vessels.