Page 43 - Read Online
P. 43

Plössl et al. J Transl Genet Genom 2022;6:46-62  https://dx.doi.org/10.20517/jtgg.2021.39  Page 56

               the SI concentrations consistent over time. Concentrations of 0.125 and 0.25 mM were chosen for these
               experiments based on findings in preliminary tests (data not shown). Initially, we confirmed that both SI
               concentrations had no negative effect on iPSC-RPE cytotoxicity [Supplementary Figure 2A], stable TEER
               values [Supplementary Figure 2B], and monolayer integrity and intact cobblestone morphology, as
               indicated by uniform ZO-1 staining [Supplementary Figure 2C]. Relative cytotoxicity of the SI treatments
               was not increased in comparison to control, but it was increased when cells were treated with 1.5 mM SI for
               72 h [Supplementary Figure 2A]. Treatment with SI resulted in a dose-dependent upregulation of HMOX1
               expression, with statistical significance being reached for the 0.25 mM treatment in all cell lines (Kruskal-
               Wallis test with post hoc Dunn’s multiple comparison test using Benjamini-Hochberg method, P < 0.01)
               [Figure 5A].


               When data were analyzed with regard to differences between HR and LR cell lines, the dose-dependent
               effect was clearly visible in both groups, although cell lines with differing AMD-associated GRS failed to
               show differences in HMOX1 expression after treatment with either of the concentrations (Mann-Whitney U
               -test, P > 0.05) [Figure 5A]. In line with the results from the 24 h treatment, NQO1 expression was impacted
               more strongly by 72 h SI treatment than HMOX1 expression. NQO1 expression also increased dose
               dependently and reached values of 3-10-fold upregulation, showing statistically significant differences
               between control and SI treatment for both concentrations in all cell lines (Kruskal-Wallis test with post hoc
               Dunn’s multiple comparison test using Benjamini-Hochberg method, P < 0.05 for comparison control vs.
                                     -5
               0.125 mM SI, P < 5 × 10  for comparison control vs. 0.25 mM SI). Concordant with data on HMOX1
               expression, NOQ1 expression upon oxidative stress treatment was not differentially expressed in HR and LR
               cell lines (Mann-Whitney U-test, P > 0.05) [Figure 5B]. The qRT-PCR results were confirmed by Western
               blot analysis and a SI dose-dependent increase in HMOX1 and NOQ1 expression was observed at 72 h of SI
               treatment compared to control. After densitometric evaluation of Western blot signal intensities, the
               observed differences failed to reach statistical significance in all cell lines (Kruskal-Wallis test with post hoc
               Dunn’s multiple comparison test using Benjamini-Hochberg method, P < 0.05), even though a clear SI dose-
               dependent trend in increased HMOX1 and NQO1 protein was notable in the Western blot images as well as
               the quantifications. No differences were found when comparing oxidative stress induced protein amounts
               of HMOX1 and NQO1 between HR and LR cell lines (Mann-Whitney U-test, P > 0.05) [Figure 5C and D].


               Taken together, oxidative stress was robustly induced in an acute 24 h setting and an extended duration
               situation using lower SI concentrations for 72 h. NRF2 downstream response genes HMOX1 and NQO1
               were upregulated under the two conditions chosen, although in no situation a genetic influence in HR and
               LR AMD iPSC-RPE lines was apparent. This led us to conclude that the genetic risk to develop AMD has no
               measurable effect on influencing the NRF2-associated first-line oxidative stress defense.

               DISCUSSION
               Our aim was to establish a cellular model system which allows us to combine both genetic and
               environmental risk factors in vitro crucial to elucidate the individual contributions to AMD pathogenesis.
               Consequently, we selected our study participants based on their genetic risk profile for AMD rather than on
               their AMD phenotype alone, resulting in four CNV patients with a very high GRS for AMD and four
               individuals with a very low GRS who had no AMD phenotype. Patient-derived iPSC-RPE cells were
               generated from these individuals and characterized for RPE integrity before they were subjected to analysis
               on NRF2-mediated oxidative stress response mechanisms. We established a short term, acute stress model
               in which cells were treated with SI for 24 h as well as an intermediate stress duration model of 72 h SI
               treatment. While significant upregulation of NRF2 pathway response genes was established under all
               experimental conditions, there was no difference in stress reaction between HR and LR cell lines. This led us
   38   39   40   41   42   43   44   45   46   47   48