Author Response: Human Trabecular Meshwork and Hepatocyte Growth Factor

We appreciate the opportunity to respond to the Letter to the Editor submitted by Dr. Aparna Rao¹ regarding our recent publication, “Hepatocyte Growth Factor Confers Trabecular Meshwork Stem Cell Resilience and Paracrine Protection of Trabecular Meshwork Cells in Glaucoma.”² We thank Dr. Rao for her interest in our work and thoughtful comments. We appreciate the careful reading of our work and the constructive suggestions regarding interpretation, experimental design, and future directions. We are pleased that the letter recognizes the potential relevance of trabecular meshwork stem cell (TMSC)–derived factors for glaucoma research and cell-free therapeutic development.

Dr. Rao raises an important point regarding cell passage. In our study, human trabecular meshwork stem cells and trabecular meshwork (TM) cells between passages 3 and 6 were used. These passages were selected to remain within the range recommended for primary human trabecular meshwork cell culture and characterization.³ We agree that primary trabecular meshwork cells may undergo phenotypic changes with extended culture and that passage-related variation should be considered carefully, particularly in studies modeling chronic glaucomatous stress. In our experiments, cells within the same experimental set were handled under comparable culture conditions, and comparisons were made using cells within this limited passage range. All the TM cells used in the experiments were characterized by responsiveness to dexamethasone treatment^2,3 and all the TMSCs used in the experiments were confirmed by OCT4 expression.^2,4 Nevertheless, we agree that future studies designed specifically to compare lower versus higher passages, ideally using passage-matched cells from the same donors, would be valuable for determining whether passage number modifies the magnitude of TGF-β2 responsiveness or the protective effects of TMSC-derived secretome and hepatocyte growth factor (HGF).

The issue of donor age is also well taken. The donors used in this study ranged from 43 to 69 years old, which reflects the adult donor population available for primary human trabecular meshwork cell and trabecular meshwork stem cell studies. Because trabecular meshwork cellularity, stem cell abundance, oxidative stress tolerance, and extracellular matrix responses may be influenced by donor age, age-related biological variation is an inherent consideration in studies using primary human tissues. Even among donors of the same or similar age, substantial interindividual biological variation exists in primary human TM and TMSC cultures. Therefore donor age is only one contributor to variability, and age-matched comparisons alone may not fully eliminate donor-dependent differences. This published study was not powered to perform age-stratified analyses, and donor age was not included as an independent variable in the statistical models. We agree that future studies with larger donor numbers should examine whether age modifies TMSC resilience, HGF expression, or responsiveness to TGF-β2.

We also appreciate the comment regarding the transcriptomic analyses. Our in-house RNA-seq comparison between TMSCs and TM cells was performed using cells from three independent donors, and the publicly available GSE202372 dataset contained five biological replicates for comparing TM cells with or without TGF-β2 treatment. As stated in our article, these analyses were intended to identify candidate pathways and molecular mediators, rather than to establish definitive transcriptomic biomarkers across all donors or culture conditions. We agree that differences between independently generated datasets, including donor characteristics, cell culture conditions, sequencing platform, and analytical pipelines, may introduce variability. Therefore the pathway-level results should be interpreted as hypothesis-generating and supportive of the functional experiments rather than as independent proof of mechanism. Future transcriptomic and proteomic studies with larger donor cohorts and matched experimental conditions will help refine the generalizability of these findings.

Dr. Rao correctly notes that cytokine membrane arrays are semiquantitative screening tools. In our study, the cytokine array was used to compare TMSC-derived secretome and TM-derived secretome and to identify secreted factors enriched in TMSC-derived secretome. HGF emerged as a prominent candidate and was further supported by transcriptomic data and functional experiments using recombinant HGF and pharmacological inhibition of the HGF receptor pathway. We agree that independent quantitative validation, such as ELISA or targeted proteomics, would further strengthen the assessment of HGF concentration in TMSC-derived secretome. We also agree that HGF immunodepletion studies would be particularly useful to determine the relative contribution of HGF compared with other secreted factors. Indeed, our observation that TMSC-derived secretome produced stronger protection than recombinant HGF alone supports the interpretation that HGF is an important but partial mediator of the TMSC secretome protective effect.

The concentration of recombinant HGF used in our experiments was selected based on cell viability assays and was intended to test whether HGF could functionally reproduce key protective effects of TMSC-derived secretome under TGF-β2 stress. We agree that recombinant HGF treatment does not necessarily match the endogenous concentration or presentation of HGF within the complete TMSC secretome. Secretome-derived HGF may act together with additional cytokines, growth factors, extracellular vesicle-associated molecules, or matrix-regulatory proteins. Therefore direct comparison between recombinant HGF and TMSC-derived secretome should be interpreted mechanistically with caution. Our conclusion that HGF is a “key, but partial, mediator” was intended to reflect this distinction.

The letter also asks about the use of SGX-523 as an HGF receptor inhibitor. SGX-523 was selected as a pharmacologic inhibitor of MET signaling, and the concentration used was determined based on TM cell viability assays. The ability of SGX-523 to blunt the protective effects of both recombinant HGF and TMSC-derived secretome supports the involvement of HGF/MET signaling. However, we agree that direct assessment of MET phosphorylation would provide stronger evidence of on-target pathway engagement in trabecular meshwork cells. Future studies examining phospho-MET, downstream signaling intermediates, and complementary genetic approaches will be important for confirming pathway specificity.

We agree with Dr. Rao's comment regarding CTGF. In our study, NAC reduced oxidative stress and lipid peroxidation but did not fully suppress CTGF induction, suggesting that CTGF may be regulated in part through ROS-independent TGF-β2/Smad signaling. Similarly, neither recombinant HGF nor TMSC-derived secretome completely abolished CTGF upregulation. This partial refractoriness likely reflects the complexity of TGF-β2 signaling in trabecular meshwork cells, where oxidative stress, mitochondrial dysfunction, Smad-dependent transcription, cytoskeletal remodeling, and extracellular matrix regulation may proceed through overlapping but nonidentical pathways. The incomplete suppression of CTGF further supports the interpretation that no single factor fully reverses TGF-β2–induced trabecular meshwork dysfunction.

Finally, we appreciate the emphasis on translational caution. Although our data support the potential of TMSC-derived secretome and HGF as cell-free modulators of trabecular meshwork stress responses, we agree that these findings remain preclinical. As noted in our article, future studies using perfused human anterior segment cultures and in vivo ocular hypertension models are necessary to determine whether the cellular protection observed in vitro translates into improved outflow facility, intraocular pressure regulation, and ocular safety. Dose, delivery method, duration of effect, and potential off-target responses will also require systematic investigation.

In summary, we thank Dr. Rao for highlighting important methodological and translational considerations. We agree that future work should include larger donor cohorts, cell passage- and donor age-related analyses, quantitative validation of secreted HGF, MET pathway activation studies, HGF immunodepletion experiments, and ex vivo and in vivo functional validation. These studies will further clarify the role of HGF and other TMSC-derived factors in trabecular meshwork protection and will help determine the therapeutic potential of TMSC-derived secretome for glaucoma.