A dual-layered microfluidic conventional aqueous humor outflow model reveals cellular interaction via TGF- β signaling in glaucoma

Introduction:: Glaucoma is a progressive neurodegenerative disease with no cure, affecting over 60 million individuals worldwide and leaving 8 million blind [1]. Current therapy is supportive and focused on the reduction of intraocular pressure (IOP), a primary risk factor for glaucoma progression. IOP homeostasis is mainly regulated by the conventional outflow route, comprised of the trabecular meshwork (TM) and Schlemm’s canal (SC), located adjacent to the iridocorneal angle [2]. In glaucoma, TM and SC experience complex changes, including extracellular matrix (ECM) deposition, contractile TM cytoskeleton, and lower SC pore density, which ultimately lead to decreased drainage and increased IOP [3-6].

While TM and SC collaborate closely to facilitate and regulate aqueous humor drainage, their cellular interaction mechanism is poorly understood. This is partially due to difficulties in studying TM-SC crosstalk in traditional glaucoma models. Though the mouse tissue-specific gene deletion model provides some clues about TM-SC interactions [7], it is costly and time-consuming, which makes it hard to expand to other potential pathways and future drug screening applications. The majority of in vitro glaucoma models focused on only TM [8-10] or SC [4, 11, 12], without integration of both cell types.

Here, we developed a dual-layered microfluidic model that integrates TM and SC in a biomimetic manner. Using this model, we showed that TM-SC interaction plays a crucial role in steroid-induced glaucoma. Next, we identified TGF-β and VEGFC as the key ligands in the interaction. We further validate our findings in vivo glaucoma mouse model.


Materials and Methods:: 1. Glaucoma-on-a-chip

The device was fabricated as previously described [13]. To achieve dual layer channel, an M-shaped SU-8 channel was fabricated to support needles with two different sizes (Figure 1A, C). The polydimethylsiloxane (PDMS) device was assembled by bonding a PDMS gasket on top of a glass coverslip. Rat tail collagen 1 (Corning) was prepared at 2.5 mg/ml and was injected into the PDMS device to surround two 400-μm acupuncture needles (Hwato). After collagen polymerizes, 400-μm needles were withdrawn, and a 250-μm needle was inserted to the right-sided channel along with primary human trabecular meshwork cell (HTMC) suspension in 3.0 mg/mL collagen (Fig 1B). After collagen cell suspension polymerizes in the gap between 400-um channel and 250um needle, the 250-μm needle was withdrawn and primary human dermal lymphatic endothelial cell (LEC) or primary human Schlemm Canal endothelial cell (HSC) were seeded into the channels.

Previous data demonstrate that the steroid-induced model faithfully recapitulates the pathophysiology of glaucoma [14]. Hence, we use dexamethasone to trigger glaucomatous changes in the devices. Drainage experiments were performed by filling the left-sided microchannel (empty) with media and measuring the media volume that drained into the right-sided microchannel (TM+SC).

2. In vivo mouse model

The steroid-induced glaucoma mouse model is described previously [15]. Briefly, 20uL 10 mg/mL dexamethasone-21-acetate suspension was periocular injected to mouse eye weekly. The IOP was measured with a TonoLab rebound tonometer (Colonial Medical Supply, Franconia, NH) another day. Mouse eye was dissected and whole mount stained according to previous protocol [16].


Results, Conclusions, and Discussions:: 1. TM-SC interaction is essential for modeling steroid-induced glaucoma

Our device achieves adjacent coculture of TM and SC (Fig 1D), which mimics Schlemm’s Canal and Juxtacanalicular Trabecular Meshwork in vivo [17]. Based on this model, the coculture model showed a significant drainage decrease with dexamethasone treatment, while the drainage score didn’t change significantly in monoculture of LEC or TM (Fig 2E). Moreover, we observed thickened junction width of the endothelium in glaucoma coculture model (Fig 2 A-D, F). Such a result indicates that cellular interaction of TM and endothelium play a key role in glaucoma.

2. TGF-β is one of the cellular interaction legends between TM and SC.

With the machine learning tool NicheNet [18], we predicted potential ligands in TM-SC crosstalk (Fig 3A) based on single-cell RNA-seq data [7]. Among them, we are curious whether TGF-β is important in TM-SC interaction. By blocking TGF-β pathway with inhibitors SB431542 and LY2109761, we showed that drainage score is raised significantly. In addition, we have also observed decreased VE-Cad junction width with TGF-β inhibition. Our result reveals that TGF-β is vital in TM-SC interaction when glaucoma develops.

3. In vivo validation of TGF-β role in TM-SC interaction

To verify that TM and SC also interact via TGF-β in vivo and further confirm that our model reproduces in vivo glaucoma condition, we intraperitoneally injected SB431542 in the steroid-induced glaucoma mouse model. In Fig 4A, SB431542 treatment significantly decreases IOP in glaucoma mice and reaches a similar IOP with the control group after 7 days administered. Besides, the VE-Cad junction width changes were also in line with in vitro experiments.

4. VEGFC is the downstream factor of the TGF-β in TM-SC interaction.

We then explored the downstream factor of TGF-β. Compared with other angiogenesis factors, VEGFC expression is dramatically decreased by dexamethasone and recovered by TGF-β inhibition (Fig 5A). Such data is quantitatively validated by ELISA assay (Fig 5B). Additionally, we treated MAZ51, a VEGFR3 inhibitor, in our glaucoma-on-a-chip model. Fig 5C showed that MAZ51 eliminates the curing effect of TGF-β inhibition, indicating that VEGFC is the downstream factor of TGF-β in TM-SC interaction.


Acknowledgements (Optional): : We wish to thank Dr. Dan Stamer and Dr. Young-Kwon Hong for providing cell lines. We thank Dr. Tae Joon Kwak, Aria R. Yslas for helpful discussions. Funding: NSF CNF (NNCI-2025233).


References (Optional): : 1. Quigley HA, Broman ATJBjoo: The number of people with glaucoma worldwide in 2010 and 2020. 2006, 90(3):262-267.

2. van der Merwe EL, Kidson SHJEER: The three-dimensional organisation of the post-trabecular aqueous outflow pathway and limbal vasculature in the mouse. 2014, 125:226-235.

3. Vranka JA, Kelley MJ, Acott TS, Keller KEJEer: Extracellular matrix in the trabecular meshwork: intraocular pressure regulation and dysregulation in glaucoma. 2015, 133:112-125.

4. Dautriche CN, Szymanski D, Kerr M, Torrejon KY, Bergkvist M, Xie Y, Danias J, Stamer W, Sharfstein STJB: A biomimetic Schlemm's canal inner wall: A model to study outflow physiology, glaucoma pathology and high-throughput drug screening. 2015, 65:86-92.

5. Wang J, Liu X, Zhong YJIjoo: Rho/Rho-associated kinase pathway in glaucoma. 2013, 43(5):1357-1367.

6. Andrés-Guerrero V, García-Feijoo J, Konstas AGJAit: Targeting Schlemm’s canal in the medical therapy of glaucoma: current and future considerations. 2017, 34:1049-1069.

7. Thomson BR, Liu P, Onay T, Du J, Tompson SW, Misener S, Purohit RR, Young TL, Jin J, Quaggin SEJNc: Cellular crosstalk regulates the aqueous humor outflow pathway and provides new targets for glaucoma therapies. 2021, 12(1):6072.

8. Zhao Y, Kim B, Grzybowski D, Weber P, Roberts CJIO, Science V: Investigation of microtopography regulated human trabecular meshwork culture for glaucoma treatment. 2009, 50(13):4880-4880.

9. Torrejon KY, Papke EL, Halman JR, Stolwijk J, Dautriche CN, Bergkvist M, Danias J, Sharfstein ST, Xie YJB, bioengineering: Bioengineered glaucomatous 3D human trabecular meshwork as an in vitro disease model. 2016, 113(6):1357-1368.

10. Russell P, Gasiorowski JZ, Nealy PF, Murphy CJJIo, science v: Response of human trabecular meshwork cells to topographic cues on the nanoscale level. 2008, 49(2):629-635.

11. Pedrigi RM, Simon D, Reed A, Stamer WD, Overby DRJEer: A model of giant vacuole dynamics in human Schlemm’s canal endothelial cells. 2011, 92(1):57-66.

12. Tian YI, Zhang X, Torrejon K, Danias J, Gindina S, Nayyar A, Du Y, Xie YJAb: A bioengineering approach to Schlemm's canal-like stem cell differentiation for in vitro glaucoma drug screening. 2020, 105:203-213.

13. Nguyen D-HT, Stapleton SC, Yang MT, Cha SS, Choi CK, Galie PA, Chen CSJPotNAoS: Biomimetic model to reconstitute angiogenic sprouting morphogenesis in vitro. 2013, 110(17):6712-6717.

14. Torrejon KY, Papke EL, Halman JR, Bergkvist M, Danias J, Sharfstein ST, Xie YJSR: TGFβ2-induced outflow alterations in a bioengineered trabecular meshwork are offset by a rho-associated kinase inhibitor. 2016, 6(1):38319.

15. Patel GC, Phan TN, Maddineni P, Kasetti RB, Millar JC, Clark AF, Zode GSJTAjop: Dexamethasone-induced ocular hypertension in mice: Effects of myocilin and route of administration. 2017, 187(4):713-723.

16. Thomson BR, Quaggin SEJLM, Protocols: Morphological analysis of Schlemm’s canal in mice. 2018:153-160.

17. Lu R, Soden PA, Lee EJM: Tissue-engineered models for glaucoma research. 2020, 11(6):612.

18. Browaeys R, Saelens W, Saeys YJNm: NicheNet: modeling intercellular communication by linking ligands to target genes. 2020, 17(2):159-162.