Transforming Growth Factor Beta Switch in Aqueous Humor of Patients With Fuchs' Endothelial Corneal Dystrophy

We have read with great interest the article by Matthaei et al.¹ on the expression of epithelial–mesenchymal transition (EMT)-related cytokines (monocyte chemoattractant protein [MCP]-1/chemokine [CC motif] ligand [CCL]2, transforming growth factor [TGF]-β1, TGF-β2, TGF-β3, basic fibroblast growth factor, tumor necrosis factor [TNF]-α, and interleukin [IL]-1β) in the aqueous humor (AH) of phakic and pseudophakic Fuchs' endothelial corneal dystrophy (FECD) eyes. With this letter, we want to confirm their findings on MCP-1, but contradict their findings on TGF-β2. Additionally, we present new data on inflammatory cytokines in the AH of FECD eyes. 

Matthaei et al.¹ found a significantly elevated concentration of MCP-1, TGF-β1, and TGF-β2 in pseudophakic FECD eyes (FECD_psph) compared with cataract controls (which we will abbreviate as Control_ph) and phakic FECD eyes (FECD_ph), but they observed no differences between FECD_ph and Control_ph. They rightly concluded that there was no primary role for these aqueous cytokines in FECD pathogenesis, but that the upregulation of these EMT-associated proteins in FECD_psph was probably secondary to cataract surgery. 

Within the frame of a study into the molecular pathogenesis of FECD, we have recently conducted a similar experiment. Unlike Matthaei et al.,¹ we also included a group with pseudophakic non-FECD edematous corneas (Edema_psph), and our analysis involved nine molecules in AH: five of them related to EMT (MCP-1, TGF-β1, TGF-β2, TGF-β3, and TNF-α), and four related to inflammation (IL-8/chemokine [CXC motif] ligand [CXCL]8, IL-6, growth-regulated oncogene α [GROα]/CXCL1 and IFN-α2). We compared Control_ph (n = 21) with FECD_ph (n = 5), FECD_psph (n = 19), and Edema_psph (n = 8). The group Edema_psph consisted of both graft failures (n = 2) and pseudophakic bullous keratopathy (n = 6). Female/male ratio was 3:2 and all patients were older than 50 years. Magnetic bead ELISAs were purchased from Merck Millipore (Darmstadt, Germany). Plate readings were done with a suspension array system and software (Bio-Plex 200 System and Bio-Plex Manager, version 4.1.1.456; Bio-Rad Laboratories, Inc., Hercules, CA, USA). Undiluted AH was used in duplicate. 

Similar to Matthaei et al.,¹ we observed a significant upregulation of MCP-1 in FECD_psph compared with FECD_ph (P < 0.05) and Control_ph (P < 0.001; Fig.). In addition, we found MCP-1 to be significantly upregulated in Edema_psph compared with FECD_ph (P < 0.001) and Control_ph (P < 0.001). Monocyte chemoattractant protein 1 did not differ significantly between both phakic groups and between both pseudophakic groups, respectively. This strengthens the conclusion of Matthaei et al.¹ that MCP-1 does not play a primary role in the pathogenesis of FECD, but rather that it reflects a common response after cataract surgery. Like MCP-1, IL-8 was significantly upregulated in Edema_psph compared with both phakic groups: Control_ph (P < 0.001) and FECD_ph (P < 0.01). Interleukin 8 was also significantly upregulated in FECD_psph compared with Control_ph (P < 0.05). This suggests again that IL-8 does not play a primary role in the pathogenesis of FECD, but that the rise in IL-8 reflects a common response to cataract surgery. However, unlike Matthaei et al.,¹ we did not observe an upregulation of TGF-β2 in FECD_psph compared with both phakic groups. Instead, we saw significantly less TGF-β2 in the AH of pseudophakic group “Edema_psph” compared with both phakic groups: Control_ph (P < 0.05) and FECD_ph (P < 0.001). In addition, there was a significant difference in TGF-β2 concentration between both phakic groups (P < 0.05) and between both pseudophakic groups (P < 0.05), respectively, with a higher concentration in patients with FECD. This suggests that TGF-β2 is in fact specific for FECD within each category of phakic or pseudophakic eyes. 

Transforming growth factor β3 was below and TGF-β1 was above the detection limit in the study of Matthaei et al.¹ In our study, it was the reverse. Therefore, we cannot comment on the findings for TGF-β1, but we can give new information on TGF-β3. There was significantly more TGF-β3 in FECD_ph compared with all other groups: Control_ph (P < 0.01), FECD_psph (P < 0.01), and Edema_psph (P < 0.001). 

Both GROα and IL-6 were specifically and significantly upregulated in Edema_psph. Interleukin 6 was upregulated in Edema_psph compared with FECD_ph and FECD_psph (P < 0.05 for both comparisons). But there was no significant difference between Edema_psph and Control_ph, nor between FECD_ph or FECD_psph and Control_ph. Concentrations of GROα did not differ between FECD_ph or FECD_psph and Control_ph either, but GROα was significantly upregulated in Edema_psph compared with all other groups (P < 0.001 for all comparisons). This suggests that the rise in GROα is specific for Edema_psph. 

These findings are in line with microarray expression analysis (MEA) and quantitative (q)PCR data that we have produced on FECD_psph corneal endothelium (CEn) compared with normal donor CEn (De Roo AK, Janssens T, Wouters J, Govaere O, Foets B, van den Oord JJ, unpublished observations, 2016). Quantitative PCR showed a significant upregulation of TGF-β2 and TGF-β3 and a significant downregulation of GROα in FECD_psph. Interleukin 6 and 8 were significantly downregulated in FECD_psph based on MEA, but not significantly changed on qPCR. Despite the increased concentration of MCP-1 in the AH of FECD_psph, the corresponding mRNA was significantly downregulated according to MEA, but not significantly changed according to qPCR. 

Our data on TGF-β2 and TGF-β3 combined with the data on TGF-β1 from Matthaei et al.,¹ suggest that TGF-β2 and TGF-β3 are specific for FECD and that there is a switch toward TGF-β1 after cataract surgery. Notably, all three isoforms of TGF-β (1, 2, and 3) are expressed in the normal cornea, with TGF-β2 being present at the highest concentration and TGF-β3 at the lowest concentration.² Interestingly, TGF-β2 is involved in tissue development² and TGF-β3 is known to have an antiscarring effect (counteracting TGF-β1 signaling).³ On the other hand, TGF-β1 plays a role in wound healing and classical EMT, and has been proposed as a key player in the disease mechanism of pseudophakic bullous keratopathy.² Furthermore, our data suggest MCP-1 and IL-8 to rise in response to cataract surgery. 

In conclusion, the TGF-β family of cytokines likely plays a role in FECD, but additional studies are required to define which members are involved in the pathogenesis, and which arise in response to surgery.1552 

Figure

 

Cytokine concentrations in aqueous humor. Transforming growth factors β2 and β3 are specifically and significantly upregulated in FECD_ph. Monocyte chemoattractant protein 1 and IL-8 are significantly upregulated after cataract surgery. Interleukin 6 and GROα are specifically and significantly upregulated in Edema_psph compared with FECD_ph and FECD_psph. Asterisks in individual graphs indicate the P value: *P < 0.05, **P < 0.01, ***P < 0.001. Error bars: display the mean and 95% confidence interval. Edema indicates non–FECD-related corneal edema and consists of both pseudophakic bullous keratopathy and graft failures. Graft failures (n = 2) are indicated in red. Distinction was made between phakic and pseudophakic eyes. Logarithmic (log) and square root (sqrt) transformations were used. The concentration itself was in pg/mL.

Acknowledgments

The authors thank Nele Berghmans from the Rega Institute (KU Leuven, Belgium) for technical assistance. 

Supported by Perdaens Fonds Eye Research and the Fund for Research in Ophthalmology (FRO, Belgium). AKDR is a PhD fellow of the Research Foundation, Flanders (FWO). The authors alone are responsible for the content and writing of the paper.