Epigenetic silencing of selected hypothalamic neuropeptides in narcolepsy with cataplexy, 2023, Seifinejad et al

[jnmaciuch]

Epigenetic silencing of selected hypothalamic neuropeptides in narcolepsy with cataplexy

Spoiler: Authors

Seifinejad, Ali; Ramosaj, Mergim; Shan, Ling; Li, Sha; Possovre, Marie-Laure; Pfister, Corinne; Fronczek, Rolf; Garrett-Sinha, Lee A.; Frieser, David; Honda, Makoto; Arribat, Yoan; Grepper, Dogan; Amati, Francesca; Picot, Marie; Agnoletto, Andrea; Iseli, Christian; Chartrel, Nicolas; Liblau, Roland; Lammers, Gert J.; Vassalli, Anne; Tafti, Mehdi

Abstract
Narcolepsy with cataplexy is a sleep disorder caused by deficiency in the hypothalamic neuropeptide hypocretin/orexin (HCRT), unanimously believed to result from autoimmune destruction of hypocretin-producing neurons. HCRT deficiency can also occur in secondary forms of narcolepsy and be only temporary, suggesting it can occur without irreversible neuronal loss.

The recent discovery that narcolepsy patients also show loss of hypothalamic (corticotropin-releasing hormone) CRH-producing neurons suggests that other mechanisms than cell-specific autoimmune attack, are involved. Here, we identify the HCRT cell-colocalized neuropeptide QRFP as the best marker of HCRT neurons. We show that if HCRT neurons are ablated in mice, in addition to Hcrt, Qrfp transcript is also lost in the lateral hypothalamus, while in mice where only the Hcrt gene is inactivated Qrfp is unchanged. Similarly, postmortem hypothalamic tissues of narcolepsy patients show preserved QRFP expression, suggesting the neurons are present but fail to actively produce HCRT.

We show that the promoter of the HCRT gene of patients exhibits hypermethylation at a methylation-sensitive and evolutionary-conserved PAX5:ETS1 transcription factor-binding site, suggesting the gene is subject to transcriptional silencing. We show also that in addition to HCRT, CRH and Dynorphin ( PDYN ) gene promoters, exhibit hypermethylation in the hypothalamus of patients. Altogether, we propose that HCRT , PDYN , and CRH are epigenetically silenced by a hypothalamic assault (inflammation) in narcolepsy patients, without concurrent cell death. Since methylation is reversible, our findings open the prospect of reversing or curing narcolepsy.

[Paragraph breaks added]

Web | DOI | PMC | PDF | Proceedings of the National Academy of Sciences

 

[jnmaciuch]

Linked by @ME/CFS Science Blog on this thread discussing loss of CRH neurons. Posting as its own thread since I had some thoughts on how these findings might be useful to ME/CFS

 

[jnmaciuch]

Explain like I'm brain foggy:
(as always, focusing on a few of the most important findings)

Type 1 narcolepsy is defined by a loss of hypocretin/orexin production (I'll use just hypocretin for the rest of this summary). The cause is unknown, though a strong genetic association with HLA and spikes in T1 narcolepsy cases after a specific influenza vaccine suggest an immune basis. Since there is no evidence of T cell infiltration or other signs that the HCRT (hypocretin-producing) neurons are actually being killed by the immune system, this paper proposed an alternative hypothesis: that an initial immune response might cause hypocretin production to get turned off in neurons epigenetically, without killing those neurons.

First, the authors looked at another gene known to be selectively expressed in HCRT neurons (Qrfp) and found it to be expressed even when hypocretin itself was undetectable. This supported the idea that those cells were still alive, just failing to produce hypocretin.

They next did an assay that looks for a specific epigenetic mark--methylation, which is a sign of epigenetic repression. They found strong evidence of methylation at the promoter region for the hypocretin gene (the place where transcription of the gene starts), suggesting that there was epigenetic modifcation causing the gene to not get transcribed.

Since another study found CRH neurons to be depleted in type 1 narcolepsy autopsies, they checked the CRH promoter and found similar hypermethylation as the hypocretin promoter.

Finally, they looked at the DNA sequence around those methylation marks to try to figure out what transcriptional regulator might be responsible for the difference. They found consensus sequences (basically stretches of DNA that specifically bind to a known transcription factor, a bit like a lock/key) matching the PAX5:ETS1 complex, which is known to induce methylation.

It is unknown what caused those changes via PAX5:ETS1. But since a similar signature was found between two different types of neurons, the authors suggest that this was evidence of a widespread respones to an earlier immune signal, rather than the immune system specifically attacking HCRT neurons as has been previously proposed.

 

[jnmaciuch]

A study on cerebrospinal fluid in T1N found no evidence of differential cytokines, also supporting the idea that there is no easily detectable active T cell response in the brain in T1N (though there's always the possibility of other evidence that hasn't been found yet).

The question at the heart of this thread's study is:
Is there a signaling loop that maintains epigenetic suppression of hypocretin and CRH in T1N that we haven't found yet, or is it possible for a one-time immune response to cause an epigenetic shift that can be maintained for years/decades?

 

[jnmaciuch]

The next question is if we can use the specific neurohormones that seem to be differentially affected in T1N vs. ME/CFS to narrow in on possible culprits.

Per Shan et al. (the T1N autopsy study), hypocretin and CRH were found to be depleted in T1N but vasopressin, oxytocin, tyrosine hydroxylase and thyrotropin-releasing hormone neurons were normal.

In ME/CFS, we have one piece of evidence that CRH is depleted and another hinting that vasopressin is depleted. Both of these studies might end up being artifacts. Though the vasopressin finding does align well with the prevalence of orthostatic intolerance issues and some member's reports of increased urination. To my knowledge there's no evidence of abnormal hypocretin in ME/CFS (at least to the level of T1N, though happy to be corrected if I'm wrong).

Just for the sake of theorizing, co-repression of CRH and vasopressin would make a lot of sense, since both seem to be dependent on very similar signaling pathways (review). Specifically, both are reliant on cAMP/CREB signaling.

So the relevant question for ME/CFS would be figuring out what would have to differ in an epigenetic modification-inducing immune response to end up with hypocretin + CRH depletion in narcolepsy but CRH + vasopressin depletion in ME/CFS? It could be that similar signals are involved, but in different proportions leading to different transcriptional outcomes (perhaps skewed by a certain HLA haplotype in narcolepsy to some extent?)

 

[jnmaciuch]

Selfishly checking connections with my own interests, it seems that IFN-a has been found to actively repress both hypocretin production and cAMP signaling in neurons (some evidence in supplementals). So I'm not sure it provides answers if we're trying to find something that represses CRH + vasopressin through cAMP but not hypocretin.

Though there are some major caveats since 1) cytokines typically don't signal in isolation, 2) these studies are looking at active repression rather than long-term epigenetic regulation, and 3) these neurons are responsive to many many different inputs and have their own internal feedback mechanisms, so we might not be able to predict what happens to hypothalmic neurons from limited in-vitro studies anyways.

That's it for me. Happy to hear any thoughts from others!

 

[DHagen]

To my knowledge there's no evidence of abnormal hypocretin in ME/CFS (at least to the level of T1N, though happy to be corrected if I'm wrong).

I am sure you are aware (and there are potential issues with each to be sure), but I am just linking a couple of relatively recent threads on papers addressing hypocretin/orexin "dysfunction" in the event that there's something to be tied together here.

Disruption of the hypothalamic orexin system links SARS-CoV-2 infection to persistent cortical neuronal pathology, 2026, Yoon et al.​

An integrative review on the orexin system and hypothalamic dysfunction in [ME/CFS]: implications for precision medicine, 2025, López-Amador​

 

[jnmaciuch]

I am sure you are aware (and there are potential issues with each to be sure), but I am just linking a couple of relatively recent threads on papers addressing hypocretin/orexin "dysfunction" in the event that there's something to be tied together here.

Disruption of the hypothalamic orexin system links SARS-CoV-2 infection to persistent cortical neuronal pathology, 2026, Yoon et al.​

An integrative review on the orexin system and hypothalamic dysfunction in [ME/CFS]: implications for precision medicine, 2025, López-Amador​

Thanks for adding the links. The first thread already has some good comments. Since we already know that cytokines like IFNa can suppress orexin in the short term, I think their findings during acute infection are no surprise (and this was not a signature that persisted long term in their mouse model, so idk the relevance to long COVID).

The second link I hadn't seen before. It has one citation for "low orexin" in ME/CFS and it's this study, which measures orexin in rats under chronic stress. I'm not sure if that was a mistake or if the author is implying that rats under chronic stress are a model of ME/CFS. If that's the only citation the author could find for a theory entirely based around orexin, makes me pretty confident that there are no documented orexin abnormalities in ME/CFS.

[Edit: it looks like the published version removed that rat study citation, which was in the Qeios version. Instead it cites Myhill et al. which found no difference between healthy and ME/CFS and calls those results "inconclusive"]

 

[bobbler]

The next question is if we can use the specific neurohormones that seem to be differentially affected in T1N vs. ME/CFS to narrow in on possible culprits.

Per Shan et al. (the T1N autopsy study), hypocretin and CRH were found to be depleted in T1N but vasopressin, oxytocin, tyrosine hydroxylase and thyrotropin-releasing hormone neurons were normal.

In ME/CFS, we have one piece of evidence that CRH is depleted and another hinting that vasopressin is depleted. Both of these studies might end up being artifacts. Though the vasopressin finding does align well with the prevalence of orthostatic intolerance issues and some member's reports of increased urination. To my knowledge there's no evidence of abnormal hypocretin in ME/CFS (at least to the level of T1N, though happy to be corrected if I'm wrong).

Just for the sake of theorizing, co-repression of CRH and vasopressin would make a lot of sense, since both seem to be dependent on very similar signaling pathways (review). Specifically, both are reliant on cAMP/CREB signaling.

So the relevant question for ME/CFS would be figuring out what would have to differ in an epigenetic modification-inducing immune response to end up with hypocretin + CRH depletion in narcolepsy but CRH + vasopressin depletion in ME/CFS? It could be that similar signals are involved, but in different proportions leading to different transcriptional outcomes (perhaps skewed by a certain HLA haplotype in narcolepsy to some extent?)

Sounds sensible to me. On both is there the possibility if it not being constantly ‘low’ necessarily but unable to turn it on enough when demand requires /runs out vs other people’s max in response to ‘stimulus’

There’s definitely something very specific in PEM for vasopressin tho (36hrs after an appointment ) that even compared to not being very normal is needing to wee every 30mins and being thirsty. And I get akin to restless leg and specific leg pains with it (interesting low potassium comes up - would only mildly low or just messed about with temporarily electrolytes?) And tbf even as a kid I noticed that when doing running training I’d be taking a drink bottle 4tikes the size to others so it’s not like things are normal to begin with .

Whereas cortisol related stuff the effect is more immediate when I’m actually breaching it normally by taking too much for too long. Due to meds and different situation I wouldn’t know if these days I’d still get the ‘can’t wake up at all’ which I assume is perhaps cortisol related somewhat but could not be at all, and I think comes after the night needing to wee with legs hurting now but when at work the not waking was the first I knew of it as keeping track of PEM wasn’t really in my control and mental /sensiry load was huge from doing 25yhings at once whilst people queued to ask me questions and the phone rang some days (which also seems to contribute)

I might be using tropey lay person assumptions here but when it’s comes to adrenaline isn’t that also where narcolepsy differs as there are lots where that (and eg excitement) can cause them to fall asleep, where the me/cfs nightmare with getting so exhausted adrenaline happens or just hasn’t tailed off from the day or pushing through is I’ve been exhausted but unable to sleep for days in the past, certain barring odd hours and definitely when I’d need to be somewhere at a certain time the next day (job, flight etc). I mean it’s got me home whilst my body has been threatening to collapse and my determination to not have to lie down somewhere I wouldn’t get rest anyway and be stuck etc but then it’s often days of being a level exhaustion but in exertion people couldn’t imagine. And yet no nodding off to the tv even for minutes if respite (weird given when I am resting and not having done that I’m often able to nod off)

On the other hand I have the common in exhaustion not being able to wake for many hours and sleeping for days really with the odd weirdly awake hour but almost coma out of it. This is very much cumulative too. But I know lots I’ve spoken to with me/cfs don’t have that? I don’t know if this is like narcolepsy as it’s often combined with deep sleep where that talks of due to broken unrefreshing sleep.