3.1. Functional expression of P2X7Rs in hiPSC-derived neurons
In order to assess the functional expression of P2X7Rs and if targeting
of these receptors reduces hyperexcitability in a human brain-relevant
cellular model, we differentiated neurons from two hiPSC lines (Kesavan
et al., 2023).
As a first approach, immunocytochemistry was performed on cells at 18-22
days after plating NSCs using two different P2X7R antibodies and the
neuronal marker β-III tubulin. Of note, while the β-III tubulin antibody
produced a more diffuse staining, the P2X7R antibodies showed the
expected punctate staining in the neuronal soma and along the processes,
demonstrating the presence of P2X7Rs on hiPSC-derived neuronal soma and
processes (Figure 1a-c ).
To complement the immunocytochemistry findings, we next sought to assess
whether these human neurons respond to the application of
P2X7R-stimulating agonists. Previous studies have shown that the
activation of the P2X7R requires high (mM range) concentrations of ATP.
The P2X7R is, however, 10-30 times more sensitive to the ATP analog and
non-selective P2X7R agonist, 2′,3′-O-(4-benzoylbenzoyl)-ATP (BzATP)
(Donnelly-Roberts et al., 2009). To assess the functional expression of
P2X7Rs, hiPSC line 1-derived neurons grown on coverslips without glial
monolayer were loaded with 2 μM Cal-520 AM calcium-sensitive fluorescent
dye to eliminate unspecific fluorescence from glia. The application of
the P2X7R agonist BzATP (300 μM) resulted in a discernable increase in
calcium fluorescence in hiPSc line 1-derived neurons (Figure 2a ). Fluorescence rose quickly in response to a pulse of 300 μM
BzATP for 5 s signifying a persistent rise in
[Ca2+]i during constant agonist exposure. In
contrast, when the cells were pre-incubated with the P2X7R antagonist
AFC-5128 (Fischer et al., 2016; Kesavan et al., 2023) (30 nM) pulse
ejection of 300 μM BzATP and 30 nM AFC-5128, changes in
[Ca2+]i were strongly reduced
(BzATP 0.0819 ± 0.0060, BzATP and AFC-5128 0.0127 ± 0.0050; p < 0.0001). The peak amplitude of the individual ΔF/F0 traces
upon BzATP application was reduced when stimulated in the presence of
AFC-5128 (BzATP 0.2086 ± 0.0362, BzATP and AFC-5128 0.0975 ± 0.024;p < 0.0001) (Figure 2b ). Similar results were
evident when the area under ΔF/F0 curve (AUC) of individual ΔF/F0 traces
were compared (BzATP 2.170 ± 0.2276, BzATP and AFC-5128 0.5178 ± 0.1307;p < 0.0001) (Figure 2c ), suggesting
BzATP-evoked responses are mediated mainly via the P2X7R.
To further confirm the functional expression of P2X7Rs, we repeated the
BzATP application in the presence of another highly-specific P2X7R
antagonist, JNJ-47965567. Similar to AFC-5128, BzATP-evoked
[Ca2+]i signals were also reduced
in the presence of a second P2X7R antagonist JNJ-47965567 (BzATP 0.1089
± 0.0134, BzATP and JNJ-47965567 0.0209 ± 0.0004; p <
0.000) (Figure 2d ). The peak amplitude of the
individual ΔF/F0 traces upon BzATP application was reduced when
stimulated in the presence of JNJ-47965567 (BzATP 0.6528 ± 0.0428, BzATP
and JNJ-47965567 0.0460 ± 0.0064, mean ± SEM; p <
0.0001) (Figure 2e ). As before, similar results were
evident when the area under ΔF/F0 curve of individual ΔF/F0 traces were
compared (BzATP 2.808 ± 0.1811, BzATP and JNJ-47965567 0.0777 ± 0.0114;p < 0.0001) (Figure 2f ).
We further confirmed the expression of P2X7Rs using hiPSC line 2-derived
neurons. Bz-ATP evoked Ca2+ influx was also blocked by
AFC-5128 (BzATP 0.0592 ± 0.0038, BzATP and AFC-5128 0.0035 ± 0.0011)
(Figure 2g ). The peak amplitude of the individual
ΔF/F0 traces upon BzATP application was reduced when stimulated in the
presence of AFC-5128 (BzATP 0.1359 ± 0.0233, BzATP and AFC-5128 0.0674 ±
0.0086; p < 0.0111) (Figure 2h) .
Similar results were obtained when the area under ΔF/F0 curve of
individual ΔF/F0 traces were compared (BzATP 1.367 ± 0.2385, BzATP and
JNJ-47965567 0.4862 ± 0.0849; p = 0.0007) (Figure 2i ).
BzATP-evoked Ca2+ influx was also blocked by
JNJ-47965567 (BzATP 0.0241 ± 0.0035, BzATP and JNJ-47965567 0.0018 ±
0.0008; p < 0.0001) (Figure 2j ). The
peak amplitude of the individual ΔF/F0 traces upon BzATP application was
reduced when stimulated in the presence of JNJ-47965567 (BzATP 0.2110 ±
0.0299, BzATP and JNJ-47965567 0.0631 ± 0.0115; p = 0.0002)
(Figure 2k ). A comparison of area under ΔF/F0 curve of
individual ΔF/F0 traces reveals a significant reduction in the presence
of JNJ-47965567 (BzATP 1.584 ± 0.2567, BzATP and JNJ-47965567 0.5119 ±
0.1265; p = 0.0002) (Figure 2l ). These results
demonstrate the functional expression of P2X7Rs in two lines of
hiPSC-derived neurons.
3.2. Antagonism of P2X7Rs alleviates PTX-induced chronic
epileptiform-like events under inflammatory conditions
Pharmacological targeting of P2X7Rs have repeatedly been proven
effective in rodent models of epilepsy (Engel et al., 2012;
Jimenez-Pacheco et al., 2016; Jimenez-Pacheco et al., 2013; Mamad et
al., 2023). The effects of P2X7R antagonism on hyperexcitability in a
human neuronal network has, however, not been reported to date. For
this, we first developed an in vitro model consisting of hiPSC-derived
neurons differentiated on human primary astrocytes. When co-cultured
with astrocytes, functional maturation of hiPSC-derived neurons has been
shown to be improved (Hedegaard et al., 2020; Johnson et al., 2007).Here, network perturbations leading to epileptiform-like bursts
were induced via the application of PTX (100 µM), a non-competitive
antagonist of inhibitory γ-aminobutyric acid (GABAA)
receptors (Figure 3a, b). Loose patch-clamp experiments were performed
to assess network excitability. As depicted by exemplary traces and
quantifications, loose patch-clamp experiments showed that treatment
with PTX for 10 min increased the burst frequency from 0.0558 ± 0.0146
Hz at baseline to 0.0925 ± 0.0149 Hz; p = 0.0358) post-treatment
(Figure 3c). Similarly, calcium imaging experiments using Cal-520
loaded neurons also revealed PTX induced an increase in network activity
and synchronisation, hallmarks of seizures (Figure 3d,
e ).
We next assessed the effect of P2X7R antagonism in this model
(Figure 4a). Comparison of burst parameters during PTX application vs.
co-application of PTX and AFC-5128 revealed no significant difference in
burst frequency ([F(2,11) = 1.79; p = 0.2053)]: baseline
0.0800 ± 0.0277, PTX 0.1421 ± 0.044, PTX and AFC-5128 0.0995 ± 0.0227
Hz), number of spikes in burst (([F(2,11) = 1.79; p =
0.2053)]: baseline 19.65 ± 2.62, PTX 19.17 ± 3.551, PTX and AFC-5128
24.94 ± 4.708; burst duration ([F(2,11) = 1.327; p =
0.2857)]: baseline 1025 ± 202.9, PTX 1209 ± 220.7, PTX and AFC-5128
1367 ± 299.5 ms) and interburst interval ([F(2,11) = 1.365; p =
0.2725)]: baseline 20701 ± 5737, PTX 12334 ± 2778, PTX and AFC-5128
17696 ± 4797 ms) (Figure 4b-e). Similar results were observed with the
P2X7R antagonist JNJ-47965567. Comparison of burst parameters during PTX
application vs. co-application of PTX and JNJ-47965567 revealed no
significant difference in burst frequency ([F(2,6) = 0.1479; p=0.4297)]: baseline 0.0477 ± 0.01258, PTX 0.0692 ± 0.01705, PTX and
JNJ-47965567 0.06347 ± 0.02547 Hz), number of spikes in burst ([F(2,6)
= 1.696; p = 0.2322)]: baseline 31.56 ± 8.213, PTX 59.16 ±
22.14, PTX and JNJ-47965567 26.93.3 ± 9.223), burst duration (baseline
1179 ± 272.6, PTX 1562 ± 281.1, PTX and JNJ-47965567 1230 ± 526.2 ms)
and interburst interval ([F(2,6) = 1.606; p = 0.1916)]:
baseline 15.03± 4.294, PTX 10.19 ± 2.126, PTX and JNJ-47965567
19.90± 7.103 s) (Figure 4f-i).
The above findings indicate that P2X7R antagonism is not
sufficient to prevent acute epileptiform-like activity evoked by PTX in
this human model. We hypothesized this may because of the absence of any
background inflammatory tone. To test this idea, we adapted the model to
more closely reflect chronic inflammatory conditions by co-treating the
PTX-exposed hiPSC-derived neuronal networks with a cocktail of
pro-inflammatory agents (TNF-α, IL-1α, IL-1β and C1q) for 7 to 12 days
(Figure 4j). Previous data have shown that treatment with these agents
leads to a neuroinflammatory phenotype in hiPSC-derived neuronal
networks (Hyvarinen et al., 2019; Stoberl et al., 2023). Loose
patch-clamp recordings were performed in the presence of PTX alone or
PTX and AFC-5128 from the neuronal cultures treated with PTX and
pro-inflammatory agents. In contrast to results from our acute settings,
burst frequency was significantly reduced in the presence of AFC-5128
(PTX 0.1012 ± 0.0156, PTX and AFC-5128 0.0662 ± 0.0159 Hz;p = 0.0342) (Figure 4k) illustrating that P2X7R antagonism
alleviates hyperexcitability in a chronic model of epileptiform-like
events under inflammatory conditions. Moreover, the mean number of
spikes in burst (PTX 19.20 ± 2.771, PTX and AFC-5128 14.95 ± 2.542 Hz;p = 0.1688) and burst duration (PTX 1404 ± 274.4, PTX and
AFC-5128 1264 ± 209.5 ms; p = 0.1688) were slightly reduced in
the presence of AFC-5128, however, this did not reach statistical
significance (Figure 4l, m). Finally, mean inter-burst-intervals (IBI)
were also increased following AFC-5128 application (PTX 10.61 ± 2.370,
PTX and AFC-5128 40.87 ± 16.28 s; p = 0.0479) (Figure 4n).
Next, we used histograms and kernel density plots to delineate
IBI subpopulations Figure 4o-q). After the addition of AFC-5128, a
distinct peak appeared (indicated by arrow in Figure 4p) at longer IBIs
indicating AFC-5128 increases successive intervals between bursts.
Furthermore, the shape and scale parameters of the distribution of the
IBIs were measured by fitting a gamma distribution to the histogram
(Figure 4r, s and Table 1). We found that the mean shape parameter of
the gamma distribution of IBI was 0.794 in PTX-treated group
and 0.482 in PTX and AFC-5128 co-treated group. Similarly, the mean
scale parameter from the gamma distribution was 8.722 in PTX-treated
group but increased to 31.048 in PTX and AFC-5128 co-treated group
demonstrating increase in longer intervals between bursts when AFC-5128
was co-applied. Comparison of the cumulative frequency distributions of
PTX or PTX and AFC-5128 treated groups also revealed increase in IBI
after AFC-5128 application (Figure 4t; p = 0.0198). These results
provide additional confirmation that AFC-5128 increases the IBI in
chronic inflammatory model of epileptiform-like events.