✵ more spoons please Read the one-page version →
How wonderful that we have met with a paradox. Now we have some hope of making progress.
Niels Bohr

Low on spoons? Start here.

Read the one-page version →

I dove into the research on what autism looks like in the body as a lifeline to recover my health and cognitive function before it deteriorated too far. When my health and cognition started declining, it was familiar because I had seen others wear my symptoms growing up, so I acted. It didn’t matter how I had to emotionally manipulate myself down this rabbit hole. I was Alice and this information was my wonderland for four years. My experience in watching two women around me growing up told me that the medications wouldn’t work and that I would lose my ability to connect with people even more than I already had.

I had already been working with doctors, and the medications were just building.

I started out looking for areas where a true majority of us had something in common in a sort of rebellion to what I was seeing in the studies, because I kept seeing 30% described as a majority, and that surprised me. Findings with percentages of less than 50% were being applied to us as a broad majority, but that percentage, to me, felt like a very weak starting point. It also surprises me that a condition with known environmental, metabolic, brain, and mitochondrial sources gets studied as one thing without separating those or even looking for what they have in overlap. I have yet to see a study ask what biological patterns show up across enough autistic people that they’re worth investigating as shared vulnerabilities. This was what I explored. I didn’t need an explanation of autism. I needed the greatest possible overlap so I could stabilize and recover.

For my own search, I wanted 80% or more, ideally 100%. I left room for extreme outliers, but I wanted to know what most of us shared, because if most of us shared it, addressing it had a better chance of helping me. I could also feel myself losing the ability to look at myself objectively, so chasing individual symptoms felt like it would come up short.

I wanted to know what biological systems might be struggling, why they were struggling, and what nutritional support could help them function better, because those were the areas I could influence. Not to mention I never doubted my brain. Rumination, anxiety, etc., aside, my brain had still salvaged my nervous system 100 times over, and I could viscerally feel that it was simply being overwhelmed and exhausted. I was losing the ability to mitigate my existence logically.

My goal was to support these areas, full stop. Healing was the only option, because this wasn’t living. After enough time, I began to see a side of autism that was a collection of overlapping biological problems.

After a year-ish of looking, these are the patterns I identified in the literature that a large percentage of us shared: low folate, B6 conversion issues, B2 conversion issues, low sulfur, gut dysbiosis, high oxidative stress, oxalates, glutamate excitotoxicity, and a downregulated enzyme called MOCOS. Addressing those individually was enough to get my cognition to a point where I could move forward with trying to understand more complicated connections, which meant diving more into LPS and CDR and fiber. In short, that worked really well, and I wanted to keep going. Understanding those connections was the most efficient way to help myself over the long term, and that is what I’m presenting here.

The symptoms that cleared for me

  • My feet stopped turning blue from circulation issues
  • My panic attacks stopped
  • Anxiety is at a low level and getting lower
  • My sensory is lower and easier to recover from
  • My capacity to handle stress is building again
  • I don’t melt down unless my iron gets low
  • I’m growing emotionally again
  • My nervous system has an improved capacity
  • I can connect with my “tribe” again
  • My physical endurance is great, strength lifting is slower but improving
  • I don’t get light headed/POTS-ish
  • MCAS symptoms are gone
  • Focus has improved
  • I’m much more comfortable with being perceived than I was, which means I’m learning how to connect in new ways
  • Sleep is more consistent and I have much more pleasant dreams (I take this as a sign of reduced fight or flight)
  • My autoimmune (Congenital Adrenal Hyperplasia, 3-beta enzyme) has come down enough that I’m in a “normal” range without taking steroids, although steroids might still be necessary to avoid bone density issues. I’m in unknown territory now.
  • My voice doesn’t fluctuate based on my anxiety, so I can speak more comfortably wherever I am. Projecting my voice is easier. Protecting my comfort is easier
  • I move more calmly and react faster, instead of moving fast, talking fast, and reacting too fast, which was more of a problem than a strength
  • My thoughts stopped hurling themselves at my skull and my mind can slow, with awareness. Not a monk, but also not at mania speeds
  • With less anxiety, I can learn new emotional concepts like standing in my own power
  • When my nervous system is calm, it’s easier to forgive people who react poorly to how my brain works while still holding respect for myself or allow myself to be angry, depending on the situation
  • I’m not a slave to whatever catches my attention next. I can manage my interests without being overwhelmed by them

The connections make sense now that I understand more. A lot of these share a common thread of needing ATP, and much of what I focused on was supporting my body’s ability to produce and handle ATP. I did that through nutrition and by addressing the gut dysbiosis. I used the studies to show me the nutrient part, and then I got brutally honest with myself to rearrange the rest of my life.

Along the way, I stopped assuming that studies built on neurotypical bodies tell the whole story about ours. And I found that the most useful question wasn’t how to cure autism. It was: what does it take for us to be healthy and stable? I might never be your cup of tea, and that doesn’t bother me. What did bother me was my nervous system being on fire 24/7, losing the ability to connect with my people, losing the ability to shape my life, my brain not working, and everyone expecting me to just change my behavior to fix it like I had any resources to spare to make that effort.

Glutathione, SAMe, and the RDA

If I take the consistent findings of oxidative stress and low glutathione as true, then I have to ask myself, what else do I need to rethink to get results? What if I wasn’t failing to do what my body needed, but had never been taught what my body needed in the first place? And what have I lost by believing the old framework and what can I still gain?

I never would have imagined that nutrition was such a significant problem, ever. The entire time I was implementing this, I still didn’t actually believe it. I didn’t believe it until I started seeing results. Once I experienced changes, I understood that I had more control over my internal world than I’d been led to believe.

Most of what we’re told about how bodies should work comes from people for whom regulation is normal. Every expectation I was trying to meet was built on that baseline, and I was trying to meet it from inside a dysregulated body. That’s what broke me. It wasn’t a lack of effort. It was life advice that came from someone living a different life.

One of the most consistent findings in autism research is oxidative stress. In plain terms, our bodies tend to carry more oxidants than antioxidants and we struggle to clear metabolic waste. What that waste is can look different for each of us, and the reasons we struggle to clear it can have different genetic components. Lead exposure, for example, can produce symptoms that overlap with autism. I never had lead exposure, but that doesn’t mean my body wasn’t being overwhelmed in other ways. For me, it was trauma and the gut with some genetic components.

The RDA

RDAs are set to prevent deficiency disease in healthy people. They aren’t designed to support optimal function, and the reference population are people with regulated nervous systems. They don’t account for higher nutrient use under chronic oxidative stress, gene variants that slow conversion into active forms (folate to methylfolate, B6 to P5P, B2 to FAD), or absorption losses when the gut is inflamed. RDAs were built from studies of healthy people, and no autistic reference group has ever been part of setting them. So “meeting the RDA” can still mean running short for an autistic body. If you think you eat well, what are you basing that on? Because I realized my idea of eating well was never based on what I needed as a neurodivergent. On a very literal level, historically, what are we basing that on, and does the current information support it as still being relevant?

Protein and conditionally essential amino acids

Some amino acids are “conditionally essential.” The body can usually make enough, but not under stress or illness. They are:

  • Arginine (supports blood pressure regulation and the urea cycle)
  • Cysteine (contains sulfur and helps build proteins and antioxidants)
  • Glutamine (fuels rapidly dividing cells like immune and gut cells)
  • Glycine (plays a major role in collagen and tissue repair)
  • Proline (supports skin healing and cellular regeneration)
  • Serine (involved in metabolism and signaling)
  • Tyrosine (used to make thyroid hormones and melanin)
  • Ornithine (participates in the urea cycle)

Glutathione is built from glutamate, cysteine and glycine, all three of which are conditionally essential. So when protein or these amino acids run short, antioxidant capacity drops.

There’s a catch I didn’t expect: stomach acid runs on ATP. The stomach’s acid pump uses ATP to make acid, so low ATP can mean low stomach acid. Low stomach acid means protein isn’t broken down well and minerals like zinc, magnesium and iron, plus B12, absorb poorly. Eating enough protein doesn’t guarantee absorbing enough. Protein that isn’t digested well moves on to the colon and ferments there, which is also an issue. This is one of those places where it’s easy to assume our bodies work like everyone else’s, just slightly off, when they’re really working from a different baseline.

✵ My framework

My solution: I eat my protein with inulin and resistant starch that I drink before meals at home. Basically, I eat my food with extra fiber and take digestive enzymes if it’s a large meal. My gut likes smaller meals. This is to promote the correct gut bacteria and prevent improper digestion, which has downstream consequences.

Glutathione and SAMe

Glutathione is the main antioxidant inside our cells. It’s what clears oxidant overload. SAMe is the body’s main methyl donor, and it runs methylation, which is involved in neurotransmitters, gene regulation, and creatine production. Both come out of the same sulfur pathway (methionine → SAMe → homocysteine → cysteine → glutathione). The protein you eat carries the sulfur amino acids methionine and cysteine. Those are vital for us in several areas we struggle with, including homocysteine and transsulfuration.

Studies of autistic children going back to James et al. in 2004 found lower levels of both SAMe and glutathione, along with more “used up” glutathione (meaning it wasn’t fresh; it had been used and recycled, showing a heavier workload for glutathione in the body). That pathway runs on folate, P5P and B2, which makes separate studies on problems with those vitamins more interesting. The two enzymes that keep glutathione working need minerals and vitamins too: glutathione peroxidase needs selenium, and glutathione reductase, which recycles used glutathione back to its fresh form, needs B2 (as FAD).

What SAMe needs to work:

  • Methionine and ATP. SAMe is made by combining methionine with ATP (enzyme: methionine adenosyltransferase). It’s built from a full spoon.
  • Magnesium and potassium. That enzyme needs both.
  • B12 and methylfolate, or betaine. After SAMe donates its methyl group, it becomes homocysteine. Homocysteine has to be recycled back into methionine, either by methionine synthase (which needs B12 plus methylfolate) or by BHMT (which uses betaine and needs zinc).
  • B6 (P5P). If homocysteine goes down the transsulfuration path toward cysteine and glutathione, both steps need P5P.
  • The purine link. When SAMe is used, what’s left (SAH) splits into homocysteine and adenosine. Adenosine is a purine and a used spoon. If adenosine backs up, SAH backs up, and SAH shuts down methylation. So methylation and the purine pathway share a drain.

Spoons aka ATP aka purines

Purines ask: could understanding the biological processes that shape our internal state help us distinguish who we are from what our bodies are struggling to regulate?

If we have known dysregulations with the Krebs cycle and ATP, is an extreme burnout an outlier, or something that is tracking with what we already know?

When I hear spoons, I think energy. Energy runs on ATP, which is a purine, and purine metabolism is a commonly dysregulated pathway in autism. It isn’t the only finding, but it’s a common one, so much so that one study suggested uric acid as a potential biomarker for ASD (PubMed). Genetic purine disorders have been connected to autism for decades. The first inborn error of purine synthesis ever identified, adenylosuccinate lyase (ADSL) deficiency, was described in 1984 as “an infantile autistic syndrome” (Jaeken & Van den Berghe, 1984).

The energy side shows up too. A 2012 meta-analysis found about 5% of autistic children meet criteria for classic mitochondrial disease, against roughly 0.01% of the general population, and about 30% show blood or urine markers of mitochondrial dysfunction (Rossignol & Frye, 2012). A 2024 meta-analysis of 204 studies found pyruvate, the fuel that feeds the Krebs cycle, elevated in about 41% of autistic individuals and lactate in about 17% (Neurobiology of Disease, 2024).

Understanding purines is probably the core of what pulled me out of burnout and gave me quality of life again because it taught me how to shape my life. Energy molecules require balance. You can’t run full throttle all the time, but their breakdown is in itself a critical antioxidant for the body, so the pathway needs fuel so it can run to its completion.

Burnout, for me, was a downhill spiral of dysregulation that showed no sign of leveling off. I’ve watched people around me with my symptoms decline to the point of seizures and heart damage, and I think it’s worth asking whether burnout is more specifically the early stage of mitochondrial strain on its way to mitochondrial damage. In my experience and that of people around me, trauma seems to be what turns it into an unchecked slide. That keeps me paying attention to purines, because fear is an ATP factory running at full blast. Fear, shame, stress: all of those are byproducts of living in fight or flight, and fight or flight produces ATP. That is something I learned through understanding the elements of this journey.

ATP carries energy in its phosphate bonds. As it’s used, it breaks down step by step: ATP → ADP → AMP → adenosine → inosine → hypoxanthine → xanthine → uric acid. The whole chain needs to work. Low spoons doesn’t default to meaning too few purines.

Burnout might not be running out of energy. It’s also worth asking if the pathway is bottlenecked or short on something somewhere, so the energy isn’t making it through in a usable way. There is ample evidence in autism that the pathway is damaged in a number of us, but not the majority of us. That’s why I didn’t write it off as a lost cause, because it looks like for some of us, the results from support may be capped; and others, support may be very successful.

Folate

Folate asks: what if meeting the nutritional standard doesn’t guarantee that the body’s underlying processes are functioning as they should?

The most important thing I learned about folate is that a folate problem is proposed to look like a purine problem in the body. And that’s an energy problem, which we don’t associate with folate at all.

The body builds new purines from scratch in a ten-step pathway, and folate (as 10-formyl-THF) supplies a carbon at two of those steps. Without enough usable folate, purine building stalls. A 2020 study modeled this in autism. Folate depletion produced nearly the same effect on the metabolic network as blocking any one of six purine-building genes, including ADSL, the same enzyme from the 1984 autism case. The authors propose the result is less ATP, and then less SAMe in neurons, which ties folate, purines, and methylation into one bottleneck (Geryk et al., 2020). It’s a computer model, not a patient study, but it gives a mechanism for why folate problems and purine problems look alike in us.

The other important part in this is that folate needs B12 and it needs serine, and serine is one of the conditional amino acids. Folate is a carrier. It picks up a single carbon and hands it off to build purines, DNA and SAMe. The main place it picks up that carbon is serine. An enzyme called SHMT strips a carbon off serine and loads it onto folate, which turns serine into glycine in the process. That enzyme needs P5P. So folate’s main carbon donor is a conditionally essential amino acid, and the handoff runs on B6. When serine or P5P runs short, folate can be present and still have nothing to carry.

The three forms of folate

  • Folic acid is the synthetic, fully oxidized form used in fortified food and cheap supplements. It isn’t active. The body has to reduce it with an enzyme (DHFR) that works slowly in humans, so some folic acid can circulate unconverted.
  • Folinic acid (leucovorin, or 5-formyl-THF) is already reduced. It skips the slow DHFR step and enters the folate cycle further along. It’s also the form used for cerebral folate deficiency, because it can reach the brain by a different route when the main folate receptor is blocked.
  • Methylfolate (5-MTHF) is the form that circulates in blood. It skips MTHFR, the enzyme people worry about with the genetic variant. But it can only hand off its methyl group through methionine synthase, which needs B12. Without enough B12, methylfolate gets stuck. This is called the methyl trap.

Folate needs B2 too. MTHFR, the enzyme that makes methylfolate, uses B2 (as FAD) as its cofactor. Low functional B2 slows MTHFR whether or not you carry the gene variant.

Deficiency vs. MTHFR in autism

  • Folate receptor autoantibodies block folate from crossing into the brain. Across studies of autistic children, they were found in about 75%, a majority.
  • The MTHFR C677T variant is a real but modest risk factor. In a meta-analysis of 1,672 autistic people and 6,760 controls, carrying the T allele raised the odds about 1.4 times. That association showed up only in countries without folic acid food fortification.

So the folate problem documented in most of us is a delivery problem, not a gene problem.

Cell Danger Response and P2X7

The danger signal asks: what happens when the body’s alarm keeps ringing after the danger has passed?

Cell Danger Response

CDR asks: what happens when the body’s protective response to danger begins to interfere with its ability to return to normal function?

Dr. Robert Naviaux proposed that when cells sense a threat, mitochondria shift from making energy to defending the cell. They make less usable ATP inside and release more ATP outside as a warning. That sets off a cascade that affects nutrient availability and inflammation, and essentially puts the body into fight or flight. It’s meant to be temporary. In autism, it may get stuck on. This was the first theory I found that was describing how I experienced life.

He tested this with suramin, a century-old drug that blocks ATP from reaching those warning receptors. It first reversed autism-like traits in mice. Then a small, double-blind, placebo-controlled trial (SAT-1) gave 10 boys either one dose of suramin or a placebo, five in each group. All five boys who received suramin improved on the main autism-severity measure (ADOS-2), by an average of 1.6 points, while the placebo group didn’t change. Families and clinicians reported temporary improvements in language, social behavior, repetitive behavior and coping skills, though a one-word vocabulary test didn’t change. Two nonverbal children spoke their first sentences about a week later. The effects faded after several weeks (Naviaux et al., 2017; UC San Diego, 2017). My criterion was the majority of us in a study. Five out of five who received it fit my criterion. Suramin wasn’t an available solution, but understanding purines is.

The takeaway for me: quieting the danger signal changed core traits. That suggests the signal itself is affecting us. That’s what Naviaux’s entire theory is. My personal standard is to address my body before taking medication. I view it as making myself whole before adding in pharmaceutical medication. Suramin isn’t available, but even if it was, I don’t want to cover up problems. I want to balance my system, and then see what medications I may need. I hold both of these as truths: that medicine can fix an imbalance in the body, and also that medicine does a very poor job of addressing imbalances before it suggests a pill.

Are we treating the person, or are we treating what the medical system knows how to recognize? Because then you have to question if they’re taking into consideration the entire depth of information available. Suramin did the impossible using the metabolism, not the brain. I’m Level 1. I deteriorated, so what do I need to recover?

In the case of autism, addressing or even questioning imbalances requires rewriting some foundational beliefs about autism. Further down, I discuss a study that opened the door for me on how the environment is impacting us.

Purine receptors

The purine receptors in autism ask: how might altered ATP and adenosine signaling affect the feedback loops that regulate inflammation, neuronal communication, and cellular responses to stress in autism?

Looking at the purine receptors themselves was one of the most fascinating parts of this exploration. I couldn’t dive into it until further in my healing because it was more advanced, and it probably wouldn’t have clicked earlier anyway. I needed the other pieces first to see the connections.

In broad strokes, there are two families of purine receptors:

  • P1 receptors respond to adenosine, the used spoon. There are four: A1, A2A, A2B and A3.
  • P2 receptors respond to ATP, the full spoon. They split into P2X (seven ion-channel receptors, P2X1 through P2X7) and P2Y (eight receptors that signal inside the cell).
  1. ATP
  2. ADP
  3. AMP
  4. adenosine
  5. inosine
  6. hypoxanthine
  7. xanthine
  8. uric acid
The full lifespan of ATP.

They’re on nearly every cell and help regulate everything from cell growth to pain signaling. Here are a couple of studies exploring them in autism:

  • A 2026 systematic review of 23 animal studies found that abnormal purine receptor levels, an off balance between ATP and adenosine, and changes in the enzymes that break ATP down outside cells all contributed to behavioral and synaptic problems, neuroinflammation, and disrupted glial communication in autism models (Guha et al., 2026).
  • A 2023 study combining blood metabolites and gene expression in autistic children found differences in three purine-handling genes (ADA, ADSL, ATIC) and five purine receptor genes: P2X7, P2Y2, P2Y6, P2Y8 and P2Y10 (Purine signaling pathway dysfunction in ASD, 2023).
  • Suramin blocks P2X and P2Y receptors broadly, which is why one drug could quiet the whole alarm system at once.

P2X7

P2X7 asks: what happens when a system designed to detect cellular danger becomes part of the problem when that danger signal persists?

One receptor sits at the center of the danger signal: P2X7. It responds to high levels of ATP outside the cell. When it fires, it triggers inflammation. It also teams up with a channel called pannexin-1, which releases more ATP and makes P2X7 fire again. In mouse research, P2X7 activation was needed to produce autism-like traits after maternal immune activation, and blocking it prevented them. So I get why Naviaux saw these receptors and went, to the ‘tism!

Here’s the part I love: nutrition directly affects P2X7. Magnesium acts as a brake, both by binding with excess ATP and by acting on the receptor itself. Zinc and copper are brakes too. Low magnesium is a real problem, not a minor gap. Deficiency causes inflammation within days and makes the body overreact to immune challenges. And because ATP has to be bound to magnesium to work at all, low magnesium is a double hit: fewer usable spoons inside the cell, and less brake on the alarm outside it.

When I learned this, I was pretty far along in my journey. I had supported the purine pathway in other ways, but I had not done much with magnesium. I paired magnesium with zinc before sleep with excellent results. At that point I didn’t have high anxiety, but there was a clear stability that happened where I was able to focus in on my world much better. It improved my regulation. (High doses of magnesium aren’t safe for everyone, especially with kidney problems, so this is my experience, not a dose recommendation. I also want to note that I take iron and copper during the day. Taking zinc will deplete both.)

Gut microbiome

What happens when a system that is supposed to nourish and protect the body doesn’t function as it should and becomes part of the problem?

I love the art of the microbiome, even though it humbles me. I noticed that gut research in autism often circles back to how the gut might be impacting the brain. I had very little interest in the brain at that point. It wasn’t because the brain isn’t interesting. It’s because the questions they were asking always involved bacteria that I didn’t have a problem with. I tested my own gut, and I didn’t have the bacteria they cared about.

So I asked a different question: if there’s a dysbiosis, what does it have in common that we all share? So many studies find something 40 or 60% of us have and propose it as a reason or contributor. What do all of us have?

A dysbiotic microbiome is a microbial community whose composition or activity is associated with impaired health or disruption of normal functions. The important distinction is that dysbiosis is not defined by one particular bacterium being too high or too low.

This matters in autism because gut bacteria can add or subtract from our well being. They can produce vitamins, like some of the B vitamins, and they can produce metabolites that take resources from our bodies to handle, like aldehydes, p-cresol, and lipopolysaccharides. All three of those have studies showing they are problems in autism. Restricted eating can also reduce gut diversity, and the research doesn’t show gut bacteria causing autism. My point here is about load, not cause. If you remember, we have higher levels of oxidative stress and low glutathione. We also have a leaky gut, so what is supposed to stay in the gut and pass through our stool often doesn’t. What the gut adds to our metabolic load matters.

Lipopolysaccharides

LPS asks: what happens when bacterial signals from the gut activate the body’s defenses, and under what circumstances does that response become a source of physiological stress rather than protection?

I did gut testing on myself, and I have higher Proteobacteria. Those are gram-negative bacteria, and they’re associated with gut inflammation. Gram-negative bacteria share one thing: lipopolysaccharide, or LPS, in their outer wall. One study (Emanuele et al., 2010) found significantly higher endotoxin in 22 adults with severe autism than in 28 controls, and higher endotoxin tracked with lower social scores (PubMed). That’s a group difference, not every person, and it’s profound autism, so I treat it as a direction rather than proof that it applies to me. After diving into LPS, see if its effects on the body match your symptoms.

LPS produces what’s called sickness behavior: fatigue, social withdrawal, sleep changes, pain, low appetite, and loss of pleasure. Those are familiar to a lot of us, with or without being sick (Savage et al., 2019).

With dysbiosis and a leaky gut barrier, more LPS gets into the blood. Some leaking is unavoidable and a part of life, but more than normal is inflammatory.

Immune cells respond to LPS by releasing ATP and nitric oxide. That ties the gut directly back to the danger signal: it produces ATP, which would need to be bound to magnesium to not bind to P2X7. Activated neutrophils, including ones activated by LPS, can also turn sulfite into reactive sulfite-derived free radicals. In mice, an induced sulfite oxidase deficiency led to much more of that radical damage (Ranguelova et al., 2012; Kumar et al., 2018). That ties LPS to sulfite clearance, which I talk about later.

LPS also pulls minerals out of the blood. In healthy human volunteers given endotoxin, serum zinc dropped after each dose. It wasn’t lost in urine. The body moved it out of the blood, likely driven by cytokines (Gaetke et al., 1997). In pregnant rats, LPS lowered serum zinc, magnesium, selenium and manganese (Kirsten et al., 2013). That’s the P2X7 brakes being pulled at the same time the alarm is going off.

Mast cells are part of this loop too. ATP firing P2X7 on mast cells makes them release histamine and other inflammatory signals, and that’s been linked to pain sensitization and migraine. They’re one of the places where extra ATP becomes histamine, pain, and inflammation.

So the picture looks like this thus far: a group of people with struggles handling what ATP breaks down into, a receptor that turns extra ATP into an alarm, and a gut that keeps feeding that alarm.

LPS may also affect nutrition directly. A 2025 hypothesis paper proposed that chronic LPS exposure in autism suppresses the transporters that move vitamins like biotin, thiamine and pantothenic acid into cells, setting up mitochondrial problems. It was based on gene expression in 12 autistic and 12 control brains, and the author calls it an open call for validation, not a finding (Dervishi, 2025).

This was the website that initially introduced me to LPS. XX

✵ My framework

My solution for LPS: short term was chitosan, which has been studied to bind it. Chitosan binds LPS directly in lab studies, and in mice, oral and injected chitosan in liposome form protected against endotoxin shock and raised survival to 83–100% (Davydova/Yermak review, 2013; Vestnik FEB RAS, 2018). Human gut studies haven’t been done, so this is my experience plus animal and lab data. That had a tremendous effect on me, but the effect had a cap. Long term, my work is geared toward influencing gut bacteria through using fiber and pectin to improve butyrate production and short chain fatty acids. Because this is the most important piece with LPS: the LPS will always exist if the dysbiosis isn’t managed. Binding is short-term relief. Research on fixing the dysbiosis is leading more towards more diversity in food, more inulin, and more resistant fiber, especially when I eat protein. The fiber is everything, but it can’t only be one type of fiber.

MOCOS and the molybdenum cofactor

MOCOS asks: could a disruption in a fundamental biochemical pathway help explain why some bodies struggle to process metabolic waste, maintain balance, and support normal cellular function?

The way I look at MoCo is as an overlooked mechanism that, while it is not the entire picture, creates a bottleneck that slows things down, and as an opportunity to see the bigger system differently.

I love the molybdenum cofactor. It made me happy. Its teachings for how the rest of the world impacts autism are beautifully woven into one place, from gut to environment to somatic. I found it early on, saw how dense the information was, went nope, not there yet, and came back to it when my brain worked better a couple of years down the line. I’m so glad I came back to it because it bridged so many of my questions, and I’m grateful to it for that.

When I started looking through studies on autism, I was tired and had no idea what to expect. The literature was showing every area of the body is impacted. I was trying to stick to areas I might be able to influence, but I was seeing one study claiming to cure a kid with antifungals, another questioning the gut bacteria clostridia’s effect on the brain, another saying it’s aldehyde toxicity. Then there’s the candida pipeline, studies showing sulfur deficiencies, and I’m looking at all of this seeing nothing cohesive until I understood this enzyme.

This enzyme covers the sulfite to sulfate conversion, aldehyde handling, and part of the purine pathway. Clostridia, candida, and fungus all emit aldehydes as a byproduct of their existence, and then you have a sulfur deficiency if sulfite isn’t getting converted, not to mention its impact on purines. The molybdenum cofactor seems to have its toe dipped into everything without being the entire problem itself, so I make sure I support it.

First, to define MoCo and MOCOS. The molybdenum cofactor (MoCo) is built from GTP, which is itself a purine, plus molybdenum. It powers four enzymes: sulfite oxidase (SUOX), xanthine dehydrogenase (XDH), aldehyde oxidase (AOX1), and mARC. MOCOS is the enzyme that adds sulfur to the “cofactor” so that XDH and AOX1 can work. Sulfite oxidase does not need MOCOS, but it does require the molybdenum cofactor to have enough nutrients to be working well. If you have ever looked into getting rid of candida and see recommendations for taking more molybdenum while you do, this enzyme is why.

MOCOS was found to be underexpressed in stem cells from 9 of 11 autistic adults in a 2016 study of olfactory stem cells, and lowering it caused synaptic problems and higher sensitivity to oxidative stress (Féron et al., 2016). 3 out of 11 had issues with AOX1, an aldehyde gene. 1 out of 11 had a MOCOS and AOX1, and 1 out of 11 had none. A 2024 follow-up preprint on cells from 8 of the most severely affected adults in that cohort found impaired glycolysis, an ATP deficiency, and reduced cell respiration (Féron et al., 2024 preprint). That’s MOCOS tied straight back to spoons. The 2016 study included two “Asperger’s” level participants. Unfortunately, the follow-up study excluded them.

Together, the enzymes MoCo is responsible for cover three kinds of cleanup:

  • Purines: Xanthine dehydrogenase finishes ATP breakdown by turning xanthine into uric acid, which is also an antioxidant. When it’s slow, spent spoons back up and you lose an antioxidant, which we need.
  • Aldehydes: Aldehyde oxidase clears aldehydes from our body. Aldehydes are a byproduct of our metabolism, oxidative damage, food, and chemicals. Autism studies show raised levels. One review goes so far as to argue that every competing theory of autism is consistent with aldehyde toxicity, meaning autism symptoms mimic aldehyde toxicity symptoms (Jurnak, 2015, Nutrition and Metabolic Insights). The same paper lists the deficiencies it expects aldehyde load to cause: sulfur antioxidants, B1, B6, folate, zinc, possibly magnesium, and retinoic acid. If you react to shampoos, smells, perfumes, etc., that can be an aldehyde issue.
  • Sulfites: Sulfite oxidase turns toxic sulfite into usable sulfate. Sulfite is a neurotoxin that has to be converted in the body. We need sulfate for detox, gut lining, connective tissue, and clearing serotonin. Autism research shows low sulfate levels overall and high sulfate loss in urine. Overload can look like headaches, nausea, racing heart, and food or preservative reactions, but it’s never been tested whether we are sulfite toxic in our bodies. All I’ve found is a mathematical estimate of what we would be based on sulfate loss in the urine. I believe I was sulfite toxic in the worst of my burnout as I had a low level nausea that just hung around.
  1. ATP
  2. ADP
  3. AMP
  4. adenosine
  5. inosine
  6. hypoxanthine
  7. xanthine
  8. uric acid
Xanthine dehydrogenase runs the last two steps.

MoCo is expansive

  • Socially and emotionally: Most expectations of what someone should be capable of are built on a regulated baseline. By regulated, I mean that stress can be added and the body will return to calm afterward. It can recoup. But we don’t seem to be able to recoup well. The emotions linger. Social interactions are also designed around you presenting your regulated self and interacting from a regulated state. That builds a sense of self. Masking is not interacting from who you are. It has an energy cost, so it’s a literal spoon drain. People pleasing and performing are a spoon drain. Requests made of you without adequate support are a spoon drain. Trauma is a spoon drain. All of these require ATP, and ATP is a finite resource in us.
  • Environmentally: Many pollutants emit aldehydes, which are a known problem in autism. Many food preservatives are sulfites, which we potentially struggle with. Fried food is a huge aldehyde burden. All of this adds to our oxidative stress. A body prone to oxidative stress reacts sooner to pollutants, metals, mold, and chemicals. Any load that overwhelms clearance counts. Registering harm earlier isn’t a weakness. It’s information.
  • Nutritionally: Food and RDAs are designed for average needs. With higher demand and lower clearance, we feel gaps and additives first. What makes us sick early may be quietly affecting everyone else with time. Preservatives, by definition, impact our gut biome. If they are restricting the growth of mold, they are also restricting our gut bacteria, and studies show that. In a group of people who already have limited diversity, this has more of an impact.

✵ Side quests

Here are two side quests I went on when I started noticing that the MOCOS downregulation had its toe in everything. One is how it feeds into current theories of autism, and the other is studies showing what these toxins block in the body, mirrored against studies of those areas being a problem in autism.

Because the MOCOS study was so small, I went down these rabbit holes to figure out how much energy I was going to put into supporting these areas, because it was a big ask.

Side quest 1

Seeing how MoCo interacted with current theories

Also known as seeing if the shoe fits, since the study was so small, for funsies.

This is not every idea out there, but it was enough for me to take MOCOS seriously considering how small of a study it was.

Oxidative stress

The findings of high oxidative stress in autism as a theory. During inflammation, xanthine oxidase (XO) generates superoxide, sulfite oxidase (SUOX) converts toxic sulfite to sulfate, and aldehyde oxidase (AOX1) helps clear reactive aldehydes. Reduced MoCo availability can therefore increase reactive species while impairing detoxification, placing oxidative stress at the center of many downstream abnormalities.

BH4

Oxidative stress oxidizes BH4 to BH2, uncoupling the four BH4-dependent conversions (dopamine, serotonin, melatonin) and flipping nitric oxide synthase into a superoxide source. BH4 comes from GTP in the purine de novo pathway, so purine disruption hits it from one side and oxidative stress from the other. LPS adds a third route through peroxynitrite. MoCo disruption increases oxidative stress through the xanthine pathway. Oxidative stress can come from multiple directions, but this is one of the large ones.

Microglia

Microglial activation runs on purinergic (ATP and adenosine) signaling and produces oxidative and aldehyde load, the same currencies the MoCo enzymes handle. So you need this enzyme working well for microglia activation. Sustained XDH-linked purinergic tone keeps microglia primed, though this one is more associative than mechanistically locked.

Excitatory / inhibitory imbalance

Two routes converge: SUOX-limited sulfite forms S-sulfocysteine, an NMDA agonist driving excitotoxicity and calpain-mediated gephyrin destruction (loss of the GABA and glycine brakes); and PLP shortage slows glutamate-to-GABA conversion (that enzyme is B6 dependent). Gephyrin's dual role links the MoCo build itself to inhibitory-synapse scaffolding. Gephyrin is the single protein where the cofactor system and E/I balance meet.

Aldehyde toxicity

AOX1, which needs MOCOS sulfuration, is a primary clearance route for reactive aldehydes, so reduced AOX1 leaves aldehydes to damage proteins, membranes, and DNA. This is one of the two airtight MOCOS-downstream branches, alongside purines. There is a fantastic paper talking about the role of aldehydes in autism and claiming it fits every known theory up to the publishing of that paper. A large number of the genes needed to handle aldehydes in the body are also thought to be connected with autism.

Mast cells, MCAS, and POTS

Sulfite triggers non-IgE degranulation through NADPH-oxidase-driven ROS, so a MoCo bottleneck that reduces the body's ability to convert sulfites matters here. LPS has its own TLR4 route. Enzymes at the intersection: DBH (aldehyde-inhibited, copper-dependent); renin (acetaldehyde triggers mast-cell release); tryptase (sulfite and oxalate both trigger); COX-2 (sulfite-driven prostaglandin release). MCAS/POTS is a possible downstream manifestation, not the organizing principle.

Essential fatty acids

Membrane PUFAs are especially vulnerable to oxidative lipid peroxidation, producing reactive aldehydes. Reduced AOX1 activity may increase the persistence of some aldehydes, linking membrane integrity indirectly to MoCo-dependent aldehyde metabolism.

Glyphosate

Stephanie Seneff's work connecting glyphosate (Roundup) to autism has received substantial criticism. She proposed that glyphosate takes the place of glycine and contributes to autism, leaning on glycine's role in the brain, but when I dug more, I found a different potential connection that interested me. One component of her proposed mechanism involves PIN1. PIN1 regulates gephyrin conformation at inhibitory synapses, and gephyrin is also the enzyme responsible for the final steps of molybdenum cofactor (MoCo) biosynthesis. Whether PIN1 also regulates gephyrin's MoCo function has not been investigated. The known biological links exist, but the proposed PIN1-MoCo connection has not yet been experimentally tested. I'm not saying it is the cause of autism or the MOCOS downregulation. I'm saying it could mimic a side of what is known to be happening already. There are multiple genetics and multiple paths to Rome in autism.

PANS, PANDAS, and post-viral

Infection and immune activation load all three MoCo queues at once. Xanthine oxidase is upregulated during inflammation, so purine substrate rises while the enzyme handling it also becomes a ROS source. Cell stress and immune cell turnover push ATP out as a danger signal, and immune cells responding to endotoxin generate sulfite. A system already running at its clearance ceiling has no reserve for that surge, which may be part of why the onset is abrupt and the resolution is slow. The overlap here is capacity, not cause. PANS and PANDAS involve anti-neuronal antibodies and basal ganglia targeting that this framework does not explain. What it may explain is why some of us flare harder and recover slower from the same infectious hit. The BH4 entry above compounds this, since PANDAS work implicates dopamine receptor antibodies and BH4 oxidation stalls dopamine synthesis from the other direction.

Candida

Candida produces acetaldehyde directly through its ethanol metabolism, placing it upstream of AOX1 and the aldehyde queue rather than merely alongside it. It also produces arabinose, which forms adducts on lysine residues, the same binding chemistry that makes aldehydes damaging. The loop runs both directions: purines feed candida, stress raises purine turnover, candida raises aldehyde load, and aldehyde load competes for the same sulfurated cofactor that clears purines. That is the mechanism behind the antifungal case reports easing autism features without any of those studies naming a cofactor. It was an interesting moment when I realized that the same medication that blocks purine receptors also clears candida. XX (an earlier draft says "because sugar breaks down into purines"; fructose speeds ATP breakdown to uric acid, which isn't the same as sugar becoming purines)

Clostridia

Clostridia has been proposed to be a factor in autism. Clostridia loads two MoCo queues at the same time. On the aldehyde side, clostridial fermentation generates acetaldehyde and related reactive aldehydes directly in the gut lumen, feeding the same AOX1 clearance route that candida burdens and that MOCOS downregulation limits. On the sulfur side, clostridial phenolic metabolites, p-cresol and HPHPA among them, are cleared by sulfation, and every molecule consumes sulfate, the exact pool a limited SUOX is already failing to replenish. So one overgrowth drains sulfate from the demand side while sulfite conversion underdelivers on the supply side, and simultaneously adds to the aldehyde queue that competes for the same sulfurated cofactor. It also competes for the sulfation capacity measured in the 90s paracetamol studies, which means the phenol findings and the clostridia findings may be one finding seen from two angles. Worth distinguishing sulfate-reducing gram-negatives like Desulfovibrio, which sit on the LPS side as well. There are a lot of possible questions here.

Side quest 2

What these toxins block, and where that shows up in autism

Sulfites, aldehydes, MCAS and POTS overlap, purines, oxalates, androgens.

The purpose of this section is to put two things side by side for each enzyme: what a specific toxin does to it, and whether there's evidence of that enzyme being dysregulated in autism. Enzymes can be slowed by genetic mutations, but they can also be slowed by nutrient deficiencies or by toxins interfering with them. The result is the same.

Sulfites

Pyridoxal kinase
Glutamate dehydrogenase (GDH)
Pyruvate dehydrogenase complex (PDH) and α-ketoglutarate dehydrogenase (KGDH)
Enzymes with critical disulfide bonds

Any human enzyme that depends on a critical disulfide bond, or a cysteine in its oxidized form, is susceptible to direct inactivation by sulfite. Three examples:

Aldehydes

Aldehyde toxicity is addressed well in one review. It describes the aldehydes the body makes internally and argues that, in looking for genetic links between autism and alcoholism, it became apparent that “all theories of ASD are consistent with aldehyde toxicity.” The deficiencies it associates with excess aldehydes are sulfur-containing antioxidants, thiamine, B6, folate, zinc, possibly magnesium, and retinoic acid (vitamin A).

There's also a tentative link between candida and autism, either as an aggravation or a risk factor. When you know that fungi like candida and aspergillus produce aldehydes as a byproduct, case studies that claim to cure autism with antifungals start to make more sense as symptom improvement. I don't think “complete recovery” is a good term, but I do think there was symptom reduction.

How aldehydes block enzymes: Reactive aldehydes come from metabolism (acetaldehyde, or 4-HNE from lipid peroxidation) and from the environment (acrolein, formaldehyde). They can attach permanently to proteins at lysine, arginine, cysteine and histidine. When that happens at an enzyme's active site or another important spot, the enzyme stops working.

Dopamine β-hydroxylase (DBH)
Glucose-6-phosphate dehydrogenase (G6PD)
Pancreatic ribonuclease (RNase A)
Angiotensin-converting enzyme (ACE)
Glutathione reductase (GR)
Arylamine N-acetyltransferase 1 (NAT1)
Aspartate aminotransferase (AST, cytosolic form)
Pyruvate dehydrogenase and α-ketoglutarate dehydrogenase (PDH and KGDH)

Enzymes with MCAS and POTS overlap

Dopamine β-hydroxylase (DBH): aldehydes
Renin: acetaldehyde

  • What happens: Acetaldehyde triggers mast cells to degranulate and release renin. In a human mast cell line, this happened in a dose-dependent way. Local renin release can activate the renin–angiotensin system, which regulates blood pressure. Caveat: the concentrations used (30–300 mM) were very high.

  • Autism evidence: [None listed yet]

  • Source: Mast cell degranulation and renin release induced by acetaldehyde (ResearchGate figure)

Tryptase (mast cell protease): sulfite and oxalate
Cyclooxygenase-2 (COX-2): sulfite

Purines

In human physiology, “purine toxicity” usually means purine metabolites building up enough to inhibit enzymes directly. Unlike sulfites or aldehydes, purines are normal molecules in the cell. They become a problem when they're dysregulated, for example when a genetic enzyme deficiency lets deoxynucleotides or uric acid build up.

Purines got considerably more interesting for me when I learned that purinergic signaling is part of how the fight or flight response starts, because anyone who's autistic knows how high our anxiety can get. Purines can also come from food, which is what's usually associated with gout, but I was more interested in how much the body produces itself. Genetic purine dysregulation has been associated with autism for decades. On a personal level, I questioned how much my daily anxiety was producing purines that were damaging my nervous system, which created more anxiety, which created more damage.

The friends I've had who were people pleasers, and I include myself here, struggled the most with burnout, some of them to a very extreme level. People pleasing is a fear response. The friends who couldn't have cared less what anyone thought, or who practiced Buddhism, have done pretty well. My answer to the purine problem was meditation and leaning into Buddhism. Do I know for sure that it remedies a purine issue? No. But dealing with being a people pleaser and meditating only brings benefit, from my perspective.

Purines get converted into uric acid, which at healthy levels is neuroprotective. High fructose corn syrup also gets turned into uric acid in the body. So my solution also includes staying away from high fructose corn syrup, which adds a burden we don't handle well.

Purine issues show up as high or low uric acid, depending partly on how well the molybdenum cofactor enzyme is working. XDH converts purines into uric acid. High uric acid can mean too many purines are being broken down. Low uric acid can mean the enzyme isn't keeping up. In autism, uric acid shows up abnormal in both directions, and low uric acid is common. Because uric acid is neuroprotective, low isn't ideal for protecting the brain.

Ribonucleotide reductase (RNR)
S-adenosylhomocysteine hydrolase (SAH hydrolase)

Oxalates

Oxalates are a signaling molecule between bacteria and fungi, and they're present in every microbiome, from the forest floor to our gut. Since autism includes gut dysbiosis, most of us have an oxalate problem to some degree. Oxalates matter because they intersect with the sulfur pathway, and sulfur is the third most abundant mineral in the body. They bind minerals, especially calcium, which can create a deficiency even when your diet has enough calcium.

Oxalates can be produced internally and taken in from the diet. B6 is critical here because the AGXT enzyme depends on it. AGXT converts glyoxylate into glycine, which keeps glyoxylate from turning into oxalate. Genetic problems with AGXT are rare and cause a condition called hyperoxaluria, which is known to lead to oxalate kidney stones. Oxalate levels in autistic children have been found to be significantly higher than in controls.

Where to learn more: The TLO group on Facebook is the most in-depth place I've found.

To support myself here, I supported my B6 levels. I also removed the highest oxalate foods from my daily diet, like spinach, potatoes, celery, nuts and carrots. Those were my biggest offenders, but others may be different. If I have potatoes on Thanksgiving, that's not an issue for me personally, but it could be for some people. Overall, I don't crave or miss these foods, which surprised me. When I changed my diet in a way that improved my quality of life, I became self-regulating about maintaining it, because any regression didn't feel good.

Androgens and the autism connection

This one interests me because in 2015 I was diagnosed with Congenital Adrenal Hyperplasia (3β-hydroxysteroid dehydrogenase) through an ACTH stimulation test, but it was always called a “partial” enzyme deficiency. Since I started figuring out what nutrition worked for my autism, my bloodwork has improved every year. Unfortunately, that has meant more expenses, because my endocrinologist now wants to run tests like renin, believing the original diagnosis was wrong. I don't think it was wrong. Like I pointed out earlier, enzymes can be inhibited by genetic mutations, but also by nutrient deficiencies or by toxins interfering with them. The result is the same. The enzymes below are part of the Congenital Adrenal Hyperplasia family.

Autism evidence for the whole family: Androgen levels are altered in autism, and children with CAH show differences in autistic traits.

CYP21A2 (21-hydroxylase)

  • What it does: Makes cortisol and aldosterone. It's the most common enzyme involved in CAH.

  • Toxin evidence: [Needs a CYP21A2-specific source. The study below is about CYP1A2, a liver drug-metabolizing enzyme, not CYP21A2: in sulfite oxidase-deficient rats, CYP1A2 activity dropped about 28%. Keep it as a liver detox finding or move it.]

  • Source: Alteration of drug metabolizing enzymes in sulphite oxidase deficiency (PMC)

CYP11B1 (11β-hydroxylase)
CYP17A1 (17α-hydroxylase/17,20-lyase)
HSD3B2 (3β-hydroxysteroid dehydrogenase type II), my CAH enzyme
StAR (steroidogenic acute regulatory protein)

Skip the side quests and keep going to B6 (P5P) →

B6 (P5P)

In autism, could the problem sometimes be not how much vitamin B6 is available, but how effectively the body converts and uses it?

Some studies have found altered circulating vitamin B6 profiles in autistic children. Researchers have proposed that impaired conversion to the active cofactor PLP could contribute, but the specific enzymatic defect has not been established as a general feature of autism. I looked into P5P specifically because it was an overlap between glutamate excitotoxicity, oxalates, and aldehydes.

Glutamate excitotoxicity

Glutamate is the brain’s main “go” signal, and GABA is its main “calm down” signal. The enzyme that turns glutamate into GABA, glutamate decarboxylase (GAD), only works when it has sufficient P5P. No P5P, no conversion: glutamate stays high and GABA stays low.

This shows up in autism on both ends:

  • Less of the converting enzyme. Both forms of GAD (GAD65 and GAD67) were significantly reduced in the cerebellum and parietal cortex of autistic brains (Blatt & Fatemi, 2011 review).
  • More glutamate. A meta-analysis of 12 studies with 880 participants found higher blood glutamate in autism, though results varied a lot between studies (Zheng et al., 2016).

Oxalates

Oxalates ask: what if understanding oxalates could teach us something bigger about how individual differences in the gut and metabolism shape our response to food?

Oxalates were a huge dilemma for me. I was constantly asking myself if something like this applied. What helped me find peace in my solution was understanding them better. Oxalates are fascinating to look at ecologically because they are part of the chemistry connecting organisms to minerals, microbes, and their environment. They change with an ecosystem’s dysbiosis, and autism has the gut dysbiosis.

The big idea is that oxalate behaves differently depending on the ecosystem around it. In a forest, it can help organisms manage minerals and weather rock. In marine sediments, it can participate in microbial carbon cycling and mineral transformations. In both settings, its effects depend on what produces it, what consumes it, and which minerals are available.

For reference, oxalate can be produced inside the body or eaten in certain foods. It binds minerals and becomes a damaging crystal. Some fungi, like Aspergillus, produce oxalate, and some gut bacteria, like Oxalobacter formigenes, break it down, so the balance shifts when the biome does. Autism does not have Oxalobacter formigenes. In autism, a 2012 study found significantly higher oxalate in the blood and urine of autistic children than in controls (Konstantynowicz et al., 2012, European Journal of Paediatric Neurology).

The balance for me came in learning how to reduce high oxalate foods without adding the stress that creates purines. I am not true low oxalate, but I do not eat high oxalate, because even with a more diverse diet, I still repeat the same meals. If I base my meals around high oxalate foods, I could easily end up high oxalate. That means I removed spinach, potatoes, celery, etc. from my diet. Spinach got replaced with bok choy, potatoes are replaced with squash, and so on. If I have celery in chicken soup, it’s okay. I will never make a celery or spinach smoothie. I do not eat flour from nuts or nut milk, like almond milk. I will eat a cookie with walnuts in it occasionally. It’s a balance of inclusion and exclusion.

B6 comes into play here too. A liver enzyme called AGXT (alanine-glyoxylate aminotransferase) keeps glyoxylate from being turned into oxalate, and it needs P5P to work efficiently. Genetic disorders of this gene (primary hyperoxaluria type 1) are rare, but when they happen, oxalates can become extremely damaging. So my solution to address oxalate was to replace any high oxalate foods I ate with low to medium oxalate food and make sure I had enough B6 and that my B6 was working.

As a note, animals on high oxalate pasture will die from a calcium deficiency, which showed me that it can damage the body. Those of us with big tempers might look more into oxalates. A calcium deficiency creates anger issues, and oxalates bind calcium. When I looked around in my own life, the autistic friends I had who had more aggressive histories also had signs of oxalate issues, like interstitial bladder issues and kidney stones.

What B6 needs to work

B6 from food has to be converted into P5P before the body can use it. Two enzymes do that:

  • Pyridoxal kinase needs ATP, bound to magnesium.
  • PNPO needs B2 (as FMN).

This is where autism research gets interesting. Unsupplemented autistic children had 75% higher total B6 in their blood than controls. The authors read that as B6 piling up unconverted, consistent with earlier reports of low pyridoxal kinase activity and low P5P in autism (Adams et al., 2006). So a “high B6” lab result can hide a P5P shortage. Taking P5P directly skips the conversion step, and making sure magnesium, B2 and ATP are available supports it.

P5P also runs the enzyme that makes serotonin and dopamine (AADC), and one study found that enzyme’s activity significantly reduced in autistic subjects (Footitt, 2013). Magnesium is the other half of the glutamate story: it sits in the NMDA glutamate receptor at rest and acts as a brake on overexcitation.

B2

B2 asks: what if the problem isn’t simply getting enough nutrients, but whether the body’s biochemical machinery has the resources and capacity to use them effectively? And how would we know? Because for burnout specifically, could differences in riboflavin availability or metabolism help explain some of the difficulties in energy production, antioxidant recycling, and other enzyme-dependent processes that are observed in us?

B2 sits underneath almost every pathway in this framework. It’s the hidden partner to folate, P5P, glutathione, the molybdenum enzymes, and ATP production. You could take only B2 and feel nothing, and never understand its value without understanding the framework.

And there’s a whole study in autism saying that every participant had a functional deficiency in B2, meaning the vitamin was there but not able to be used by the body. It looked at Organic Acids Test results from 600 autistic people aged 2 to 30 and reported functional B2 deficiency markers in every one (“Functional Vitamin B2 Deficiency in Autism,” Journal of Pediatrics, Perinatology and Child Health). That study deserves a follow-up, because it’s the one on my list with the most potential for a bias. There was no control group, the organic acid markers it used are indirect and can rise for other reasons, and the journal’s peer review is weak.

B2 is necessary for B6 and B12 to work also. We know B6 is an issue in autism, but not why. B12 is vital for glutathione. In postmortem brain tissue, the active forms of B12 (methylcobalamin and adenosylcobalamin) were more than three times lower in autistic subjects than in age-matched controls, and lower methyl-B12 went with lower methionine synthase activity and higher homocysteine (Zhang et al., 2016). The lead researcher noted that blood B12 may not give the full picture of what’s happening in the brain.

What B2 needs to work

Riboflavin from food has to be converted into its working forms, FMN and then FAD (Riboflavin review, PMC):

  • Riboflavin kinase turns riboflavin into FMN. It needs zinc and ATP.
  • FAD synthetase turns FMN into FAD. It needs magnesium and ATP.
  • Thyroid hormone turns both enzymes up (Riboflavin metabolism review, PMC). That’s where iodine and selenium come in: iodine is needed to make thyroid hormone, and selenium-dependent enzymes convert it into its active form. Without enough thyroid hormone, the FAD-dependent enzyme glutathione reductase drops to levels seen in riboflavin deficiency.
  • Molybdenum is B2’s partner rather than its activator. Xanthine dehydrogenase and aldehyde oxidase each need both the molybdenum cofactor and FAD to work, so one without the other stalls both enzymes.

If it seems like I didn’t find a solution, I didn’t. That was never my goal. My goal was to support my system so I could heal. If it feels like this is a lot to do, it is. But autism is not the common cold, so I view it as necessary. It was a slow process to implement all of this, but as I progressed, it became easier. It’s very hard to be unhappy and autistic. Doing this made it much easier to be autistic because I wasn’t fighting my metabolism anymore.

By nutrient

By nutrient asks: what does each nutrient need to work, and what does it hand back to the rest of the system?
  • Zinc and copper: Zinc holds the gut lining’s tight junctions together, brakes P2X7, activates B2, and runs BHMT, which recycles homocysteine. In 1,967 autistic children, 29.7% were zinc deficient by hair testing, rising to 43.5% of boys and 52.5% of girls aged 0–3. That study compared against reference ranges, not a matched control group (Yasuda et al., 2011). LPS lowers serum zinc in humans (Gaetke et al., 1997). Copper is needed to build the molybdenum cofactor and is another brake on P2X7. Taking zinc depletes copper and iron, so I take them at different times.
  • Iron: I melt down when my iron gets low. Zinc competes with it, and LPS lowers plasma iron alongside zinc in animal studies.
  • Molybdenum: It’s needed for the molybdenum cofactor and usually isn’t in supplements. In one study, autistic children showed signs of functional molybdenum deficiency (excess urinary sulfite), and molybdenum supplementation normalized sulfite in about 36% of them (Waring & Klovrza, 2000). Molybdenum works side by side with B2 in xanthine dehydrogenase and aldehyde oxidase.
  • Selenium: Glutathione peroxidase needs it, and so does thyroid hormone activation, which turns up the enzymes that activate B2. LPS lowered selenium in rats.
  • Iodine: Needed to make thyroid hormone, which turns up B2 activation.
  • Magnesium: Binds ATP so it’s usable and doesn’t aggravate P2X7. It’s also the brake on the NMDA glutamate receptor, it’s needed to activate both B6 and B2, and the enzyme that makes SAMe needs it. LPS lowered magnesium in rats.
  • Folate: Needed for two steps of purine building and to run methylation. It gets its carbon from serine. See the Folate section for the three forms.
  • B12: Needed by methionine synthase to recycle homocysteine and release methylfolate. Active B12 was more than three times lower in autistic brain tissue (Zhang et al., 2016).
  • B2: Activates B6 (PNPO), runs MTHFR to make methylfolate, recycles glutathione, and partners with molybdenum. See the B2 section.
  • B6 (P5P): Turns glutamate into GABA, keeps glyoxylate from becoming oxalate (AGXT), runs transsulfuration toward glutathione, and runs SHMT, which loads folate with carbon from serine.
  • B3 (as NAD+): Aldehyde dehydrogenases use it to clear aldehydes. That’s the other main aldehyde route besides aldehyde oxidase.
  • B1, biotin and pantothenic acid: These are the vitamins whose transporters may be suppressed by chronic LPS (Dervishi, 2025, hypothesis). All three feed mitochondrial energy production.
  • Fiber: Fiber binds toxins in the gut so they can be excreted. Fermentable fiber like inulin and resistant starch doesn’t increase protein absorption, but it changes what happens to protein that reaches the colon: bacteria ferment the fiber instead, and protein fermentation byproducts drop. One of those byproducts, p-cresol, is elevated in the urine of young autistic children (Altieri et al., 2011; Persico & Napolioni review).
  • Serine: Folate’s main carbon donor, and conditionally essential.
  • Sulfur and molybdenum: Sulfur amino acids feed glutathione and SAMe, and sulfite oxidase needs molybdenum to turn sulfite into usable sulfate.
  • Creatine: Making creatine uses about 40% of SAMe’s methyl groups (Brosnan et al., 2011), and the brain uses creatine to buffer ATP. While healing, I used creatine so my body could spend its SAMe and ATP elsewhere.
  • Glycine, glutamine, methionine and betaine: Glycine and glutamine are building blocks for both glutathione and purines. Methionine starts the SAMe pathway. Betaine recycles homocysteine back into methionine (BHMT, which needs zinc).

Removing burdens

Removing burdens asks: what can I take off my system so the support I add has room to work?
  • Chitosan and activated charcoal: These binders work in the gut and catch things before they’re absorbed, which lightens the load on detox. They also bind nutrients and medications, so I take them apart from food, supplements and meds.
  • Breathwork: It shifts the nervous system out of threat mode, supports oxygen delivery for ATP, and lowers the stress demand that drains spoons.
  • Social expectations: Dropping masking and unrealistic standards turned out to be a physical intervention, not just an emotional one. It removes a constant energy cost. The more regulated I got, the simpler I wanted my life to be so I could enjoy the regulation and not lose it. I started wanting to shift to focusing on those around me instead of chasing dopamine. Regulation wants to maintain itself. I don’t chase dopamine. I rate each activity on whether it benefits me and adds to my existence, in a much more methodical way now.

How I put it all together

This world was not designed for me, so I will redesign my world until I figure out what works.

Parts of the loop I impacted, as I understand it:

  1. Gut dysbiosis lets LPS in, which primes inflammation.
  2. A bottlenecked ATP pathway and a stuck danger signal push ATP outside cells. P2X7 fires, pannexin-1 amplifies it, and mast cells join in. Low magnesium or zinc makes this worse.
  3. Inflammation raises oxidative stress, which uses up glutathione.
  4. Rebuilding glutathione takes folate, P5P, B2, protein, glycine and sulfur. Absorbing those depends on stomach acid, which depends on ATP.
  5. Clearing the waste takes molybdenum and MOCOS.
  6. When clearance backs up, cells stay stressed, ATP stays bottlenecked, and the loop repeats.

What I did:

  • Supported supply: nutrients, magnesium, zinc and amino acids.
  • Supported clearance: molybdenum, sulfate and antioxidants.
  • Lowered load: gut, binders, breathwork and expectations.
  • Watched which symptoms cleared as feedback.

What changed when the loop quieted down: less mast cell reactivity, better blood flow, more ATP available for my systems to actually run, and a steadier mind. My brain improved because my body had what it needed.

Research already shows these areas improve symptoms

What if much of what we need is already in the research, and it just hasn’t been asked about in a way that helps us?

When I started this, I didn’t have time to wait for a solution to come down the pipeline. What surprised me was how much of what I needed was already in the research. It just wasn’t being asked about in a way that would have helped me.

Most autism research asks how to bring us closer to a typical baseline. I’d like to add another question alongside it: what does it take for an autistic body to be comfortable, and what around us works against that? Looking through that lens, a pattern shows up. When you change what’s happening in the body, traits we were told live only in the brain start to move.

  • The gut: In a microbiota transfer study, children with autism and gut problems kept improving for two years after treatment. It was small and still needs a controlled trial, but the direction is hard to ignore.
  • Folate: When folate is blocked from entering the brain, a form that takes a different route has improved communication, irritability, coordination and seizures.
  • Folate and purines: In a metabolic model, a folate deficiency produced nearly the same effect as a blocked purine pathway (Geryk et al., 2020; see Folate).
  • The danger signal: One dose of a drug that quiets the cell’s alarm brought temporary gains in core autism measures in all five boys who received it.
  • Inflammation, sleep, and blood flow: Inflammation pulls brain chemistry away from serotonin and melatonin, disrupts sleep, and is tied to reduced blood flow to the brain. Treatments aimed at those areas have shown some relief.
  • The barriers in between: Autism research shows changes in both the gut barrier and the blood-brain barrier. Those are the physical route from body to brain.

None of this is anyone’s fault. These are just the conditions most of us live in. A more sensitive body registers them sooner. To me that isn’t a flaw. It’s information on how society’s development isn’t working for us.

Sources

67 sources
  1. Purine and purinergic receptors (Burnstock)
  2. Pannexin-1 and P2X7 feed-forward model
  3. Pannexin-1 mediates IL-1β release by P2X7
  4. Zinc/copper inhibition of P2X7
  5. Copper, zinc, magnesium modulation of P2X7
  6. Magnesium sulfate inhibits inflammation through P2X7
  7. Magnesium and the inflammatory response
  8. Suramin SAT1, UC San Diego
  9. SAT1 results
  10. Horváth et al. 2019, P2X7 and autism-like behavior
  11. Systematic review: ASD and gut microbiota (2023)
  12. Sulfite-derived radicals in LPS-challenged mice
  13. P2X7 induces degranulation in human mast cells
  14. P2X7 in meningeal mast cells and migraine
  15. Sickness behavior, Savage et al. 2019
  16. Low-grade endotoxemia in severe autism
  17. Serotonin sulfation deficit in ASD
  18. 4-HNE in autism
  19. Brain MDA in autism
  20. Curran et al. 2007, fever and ASD
  21. Kang et al. 2019, MTT two-year follow-up
  22. Leucovorin meta-analysis
  23. Kynurenine pathway and ASD
  24. Cerebral hypoperfusion in ASD
  25. Fiorentino et al. 2016, BBB and gut barrier
  26. Antibiotics and the gut for years
  27. Emulsifiers and gut barrier
  28. Stress and intestinal barrier
  29. James et al. 2004, SAMe and glutathione in autistic children
  30. Jaeken & Van den Berghe, 1984. An infantile autistic syndrome characterised by the presence of succinylpurines in body fluids. Lancet
  31. Rossignol & Frye, 2012. Mitochondrial dysfunction in autism spectrum disorders: a systematic review and meta-analysis. Molecular Psychiatry
  32. Biomarkers of mitochondrial dysfunction in autism spectrum disorder: a systematic review and meta-analysis, 2024. Neurobiology of Disease
  33. Geryk et al., 2020. The key role of purine metabolism in the folate-dependent phenotype of autism spectrum disorders: an in silico analysis. Metabolites
  34. UC San Diego, 2017. SAT-1 suramin trial results
  35. Guha et al., 2026. Autism spectrum disorders and purinergic signaling: a systematic review of emerging insights from preclinical studies. Journal of Molecular Neuroscience
  36. Purine signaling pathway dysfunction in autism spectrum disorders: evidence from multiple omics data, 2023
  37. Dervishi, 2025. A systems hypothesis of lipopolysaccharide-induced vitamin transport suppression and metabolic reprogramming in autism spectrum disorders. Metabolites
  38. Féron et al., 2016. Olfactory stem cells reveal MOCOS as a new player in autism spectrum disorders. Molecular Psychiatry
  39. Féron et al., 2024 preprint. Autism spectrum disorders: an impaired glycolysis induces an ATP deficiency and a reduced cell respiration. bioRxiv
  40. Blatt & Fatemi, 2011. Alterations in GABAergic biomarkers in the autism brain
  41. Zheng et al., 2016. Blood glutamate levels in autism spectrum disorder: a systematic review and meta-analysis. PLOS ONE
  42. Adams et al., 2006. Abnormally high plasma levels of vitamin B6 in children with autism not taking supplements
  43. Footitt, 2013. Vitamin B6 and serotonin metabolism in neurological disorders of childhood (thesis)
  44. Konstantynowicz et al., 2012. A potential pathogenic role of oxalate in autism. European Journal of Paediatric Neurology 16(5):485–91
  45. Riboflavin: human physiology and health review
  46. Riboflavin metabolism and thyroid regulation review
  47. Riboflavin enzyme roles (Principles of Nutritional Assessment)
  48. Yasuda et al., 2011. Infantile zinc deficiency: association with autism spectrum disorders. Scientific Reports
  49. Resistant starch decreases protein fermentation in vitro
  50. Altieri et al., 2011. Urinary p-cresol is elevated in small children with severe autism spectrum disorder
  51. Persico & Napolioni, 2013. Urinary p-cresol in autism spectrum disorder (review)
  52. Naviaux et al., 2017. Low-dose suramin in autism spectrum disorder: a small, phase I/II, randomized clinical trial. Annals of Clinical and Translational Neurology
  53. Emanuele et al., 2010. Low-grade endotoxemia in patients with severe autism. Neuroscience Letters
  54. Ranguelova et al., 2012. Formation of reactive sulfite-derived free radicals by the activation of human neutrophils. Free Radical Biology & Medicine
  55. Kumar et al., 2018. Sulfite-induced protein radical formation in LPS aerosol-challenged mice. Redox Biology
  56. Gaetke et al., 1997. Effects of endotoxin on zinc metabolism in human volunteers. American Journal of Physiology
  57. Kirsten et al., 2013. Lipopolysaccharide exposure reduces maternal serum Zn, Mg, Se and Mn concentrations in rats. Brain, Behavior, and Immunity
  58. LPS lowers plasma zinc and iron in rats. Journal of Thermal Biology, 2000
  59. Yermak/Davydova et al., 2013. Marine compounds with therapeutic potential in gram-negative sepsis (chitosan and LPS). Marine Drugs
  60. Chitosan and its acyl derivatives in liposomes protect mice against endotoxin shock. Vestnik FEB RAS, 2018
  61. Jurnak, 2015. The pivotal role of aldehyde toxicity in autism spectrum disorder. Nutrition and Metabolic Insights
  62. Functional Vitamin B2 Deficiency in Autism. Journal of Pediatrics, Perinatology and Child Health
  63. Zhang et al., 2016. Decreased brain levels of vitamin B12 in aging, autism and schizophrenia. PLOS One
  64. Waring & Klovrza, 2000. Sulphur metabolism in autism (molybdenum result, as cited in urinary elements study)
  65. Pu et al., 2013. Association between MTHFR gene polymorphisms and the risk of autism spectrum disorders: a meta-analysis. Autism Research
  66. Frye/Rossignol. Folate receptor alpha autoimmunity and cerebral folate deficiency in autism. Journal of Pediatric Biochemistry
  67. Anuroj, 2026. Cerebral folate deficiency and the role of leucovorin. Drugs in Context