The main project of my dissertation, (Mechanisms of P2X-mediated ethanol consumption) is finally published! Check it out here!

This project started over 4 years ago when my advisor basically dared me to coexpress P2X4 and NMDA receptors in oocytes. I made a lot of mistakes along the way (tried to cut a plasmid with the wrong restriction enzyme, tried to do slic with a t7 ligase instead of a t7 polymerase, etc) but I also learned a lot, which made me the scientist I am today.

I also have to give props to the team! This paper would not have been possible without the help of my coauthors. You all have gifts coming soon!

Article summary

P2X are ion channels activated by ATP. Multiple subtypes exist, such as P2X1-7. They play a role in a lot of different CNS functions and pathologies, namely alcohol use disorder (AUD) which is what my advisor’s lab studies. The problem is we don’t know exactly how this occurs. This is very important for drug development/approval: if you want to give a drug to humans, you better have a good target (i.e. know what’s going on and why.) A more established target for AUD is a subset of glutamate (Glu) receptors: NMDA receptors (NMDARs). However, while we know how this receptor contributes to the development of AUD, this receptor is harder to drug: you either make a glutamate or glycine mimetic (think of the off-target potential), target the pore (but make it reversible!), or find an allosteric pocket on the ectodomain (off-target potential; see Ivermectin as an example.) To date, all NMDAR drug candidates for AUD have failed in human clinical trials.

Where do P2X come in? Studies have shown they can affect the activity of other ion channels. In fact, studies have suggested that P2X could affect NMDAR activity, although the exact mechanism (and even effect; potentiation or inhibition?) is controversial. That was the goal of the project; to determine the role of P2X in NMDAR modulation.

First things first, we had to make sure that P2X and NMDARs could function when coexpressed in our model system, xenopus laevis oocytes. We did this a few ways: 1) we measured the P2X or NMDAR response to its respective agonist when expressed alone, and in combination with the other receptor, and 2) we expressed only P2X or NMDAR and measured the response to its agonist, the other receptor’s agonist, and the response to both agonist. In the end, we found that coexpressing these receptors doesn’t affect the separate agonist responses, and that each receptor is unaffected by the agonist of the other receptor (i.e. ATP doesn’t affect NMDARs, and Glu doesn’t affect P2X.)

With our setup validated, we started activating both receptors at the same time. We figured that, if there was no interaction/modulation, then applying Glu+ATP should produce the same response as the sum of individual Glu and ATP responses. However, when coactivating P2X4 and NMDARs, the responses we get are much lower (more inhibited) than we expect. The same occurs with P2X2 and NMDARs. In fact, P2X2-NMDAR interactions seem reciprocal; activating either P2X2 or NMDARs first, then coactivating the other receptor produces very small responses regardless of the order.

P2X4 is known to be the most calcium permeable P2X, which is theorized to affect NMDAR activity. Since our buffer contained calcium, we decided to redo the P2X4-NMDAR experiments and replace calcium with barium. Again we see inhibitory responses when P2X4 and NMDARs are coactivated, suggesting that the interaction is independent of calcium. However, things are more complicated than that; we saw that NMDAR responses were higher after coactivation than when we started the experiment, yet in the absence of calcium, the NMDAR responses were lower after coactivation. How could you see opposite effects after coactivation, but the same effect (inhibitory responses) during coactivation?

We did another set of studies (which our team calls pre-post experiments) where we get a stable NMDAR response (pre), then activate P2X4, then activate NMDARs again (post), to look at the duration of the interaction. If you look at the Glu responses after P2X4 activation, it seems like you can get either long-lasting NMDAR inhibition (in the absence of calcium) or almost no NMDAR inhibition at all (in the presence of calcium) at several time points. Considering our earlier results, this suggests that while the inhibitory interaction is independent of calcium, calcium entry via P2X4 plays a distinct role in NMDAR modulation (or could “mask” inhibition.)

So how does the inhibitory interaction occur? Well interactions like these usually rely on the receptor carboxy-terminal tail (CT.) The fact that P2X2 and P2X4 both interact with NMDARs suggested that maybe they do this via a similar mechanism (P2X have a similar structure, although the CTs have yet to be crystalized.) To test this hypothesis, we coexpressed the P2X2 CT in oocytes already expressing P2X4 and NMDARs, or coexpressed the P2X4 CT in oocytes already expressing P2X2 and NMDARs. As a positive control, we expressed P2X2 CT in P2X2-NMDAR oocytes. When we performed the coactivation experiments, we no longer observed inhibited responses when a P2X CT was present. Its not like expressing a P2X2 CT messes up the ability of the oocyte to express P2X4, since the responses to ATP in each case were not significantly different. In other words, the P2X CT plays a role in the inhibitory interaction with NMDARs.

Once we narrowed it down to the P2X CT, we decided to start mutating P2X4, since its much smaller than P2X2, and was previously shown to interact with GABAA receptors. Looking at the P2X4 CT, we figured that maybe the P2X4 internalization motif (Y378-L381) could be important for the NMDAR interaction (this is not the case with the GABAA interaction) so we decided to truncate P2X4 before the motif (377-TR), or replace the motif with a Flag epitope (P2X4-FLAGIN.) As a positive control, we deleted everything after the internalization motif (382-TR.) It turns out, none of these P2X4 mutants are capable of the inhibitory interaction with NMDARs, suggesting that the P2X4-NMDAR interaction relies on residues 378-388. That’s cool, but what about P2X2-NMDAR interactions? We didn’t have time to truncate P2X2, but we did have a peptide consisting of the 11 CT P2X4 residues (11C) so I figured we could inject 11C into oocytes expressing either P2X2 or P2X4 and NMDARs, to confirm the importance these residues, and again, 11C was able to disrupt the inhibitory interactions.

So big picture, what does this mean? Well, when compared to GABA and glutamate receptors, P2X don’t seem to contribute much to synaptic transmission, but that doesn’t mean they don’t do anything. Even if P2X activation doesn’t produce current, it has an effect: NMDAR modulation. You can also see potentiation of NMDARs via P2X-mediated calcium influx, although it seems this phenomenon is distinct from the P2X-NMDAR inhibitory interaction. An obvious caveat: oocytes aren’t neurons, so we don’t have the whole picture. P2X2 and P2X4 have different localizations in neurons, so maybe these interactions serve distinct physiological functions that can’t be easily distinguished in oocytes.

Regardless, our studies demonstrate that P2X modulation of NMDARs is not either/or situation (inhibition and potentiation can occur), but more complicated (P2X interactions vs calcium influx). Classic biological conclusion. But its an important step towards understanding the role of P2X in alcohol consumption. Remember that our lab has shown that 1) P2X4 KO mice drink more ethanol, 2) P2X4 positive allosteric modulators (IVM, Mox) reduce alcohol consumption in mice, and  3) P2X4 regulate dopaminergic (DA) neuron activity in the VTA. Regulating both NMDARs and DA neurons makes P2X4 a very promising target not just for AUD, but for neuroinflammatory pathologies, as well as stroke.