Page 45 - Read Online

Plössl et al. J Transl Genet Genom 2022;6:46-62 https://dx.doi.org/10.20517/jtgg.2021.39 Page 58

phenotypic effects in vitro. In contrast, complex diseases, such as AMD, are defined by a multifaceted
genetic repertoire composed of several genetic variants that individually only display weak effects on a trait
or phenotype. Additionally, it is mostly unknown how and to what extent environmental factors interact
with the given genetic predisposition. Not surprisingly, such a situation is inherently difficult to model.
Various animal models and cell model systems have been implemented to study AMD pathogenesis,
although they all have limitations and none of them fully mimics the complexity of AMD, specifically
regarding the elaborate genetic profile of susceptibility to disease [32-34] . Since the RPE is central to AMD
[35]
pathogenesis (reviewed in ), numerous human RPE cell culture models, including the ARPE-19 cell
line [36,37] and primary (fetal) RPE , have been used to investigate the molecular pathology of AMD. In
[38]
recent years, iPSC-RPE cells have emerged as a state-of-the-art model system [39-42] . Several studies have
compared hiPSC-RPE cell lines derived from AMD patients to those derived from healthy controls. Those
studies reported patient-derived cells to display disease phenotypes associated with AMD such as increased
susceptibility to oxidative stress, higher ROS levels upon oxidative stress induction, upregulation of
complement genes, and mitochondrial dysfunction [39,41-45] . However, the majority of these studies did not
consider the complexity of AMD genetics and regularly focused on a single genetic variant, e.g., the risk-
increasing variant CFH Y402H , or refrained from genotyping their donors at all . To our knowledge, the
[41]
[46]
iPSC-RPE repository established in this study is a first collection of iPSC-RPE lines in which donors have
been selected solely on the basis of their genetic risk profile summarized by a GRS considering risk-
increasing as well as protective genetic variants rather than their phenotype. By choosing our donors from
the two extreme ends of the genetic AMD risk spectrum, the differences in their genetic risk to develop
AMD are comparable with the differences usually seen in monogenic disorders (i.e., the odds ratio between
[47]
HR and LR can reach values as high as 20) .

A distinctive condition believed to play an important role in AMD pathobiology is oxidative stress or, more
[48]
precisely, an impaired oxidative stress defense in the RPE (reviewed in ). With increasing age, a reduced
antioxidative capacity of the postmitotic RPE becomes less efficient in neutralizing accumulation of ROS,
subsequently resulting in cell degeneration and the initiation of programmed cell death [15,49-51] . NRF2 is a
master antioxidant transcription factor in many cell types and NRF2-mediated stress response is known to
play a protective role in diseases such as cancer, cardiovascular disease, and Alzheimer’s disease (reviewed
[19]
in [14,52,53] ). Further, Nrf2 deficiency in mice has been shown to lead to an AMD-like phenotype . We
therefore chose to examine the response of iPSC-RPE cells with distinct genetic AMD risk profiles to
oxidative stress induced by SI via the NRF2 pathway. Of note, in our system, there was no measurable
difference in oxidative stress response between HR and LR cell lines. This was also true in experiments after
a prolonged SI treatment for 72 h. In unstressed aging mouse RPE, Nrf2-downstream genes Hmox1, Nqo1,
and Gclm (glutamate-cysteine ligase regulatory subunit) were found to be upregulated compared to RPE of
young mice. This observation was suggested to be caused by an increase in basal oxidative stress with age
counteracted by adaptive upregulation of the antioxidant transcripts. Additionally, aged mouse RPE showed
impaired induction of the Nrf2 pathway upon oxidative stress with SI . Translating these findings to our
[15]
iPSC-RPE cell model, differences in the handling of oxidative stress in LR or HR cell lines may only emerge
if the cells are cultivated for a longer period of time, as other AMD-associated changes such as drusen
[54]
formation are known to take several weeks to develop . Additionally, as reviewed in , there are many
[48]
ways to induce oxidative stress in vitro, and, depending on the stressor chosen, cells exhibit different
response mechanisms. For example, the chemical stressor paraquat (PQ) was applied to induce a prolonged
oxidative stress in iPSC-RPE, ESC-RPE (RPE derived from embryonic stem cells), and ARPE-19 cells. In
this study, the expression of NRF2 response genes showed clear dynamics, with HMOX1 and NQO1 being
the main NRF2 effectors in the early stress phases and NQO1 and GCLC being significantly increased at
week 3 . PQ reacts with oxygen to generate superoxide radicals as well as hydrogen peroxide and hydroxyl
[49]
radicals. It also promotes superoxide radical production in mitochondria, which represents a major source