I keep stumbling into the same intersection from different directions.
For the last few weeks I’ve been pulling on a thread that started innocently — a question about why some post-COVID patients respond dramatically to extended fasting while others crash. The literature on SARS-CoV-2 and autophagy is dense, and most of it is written by virologists for virologists. But buried inside it is a mechanism that, viewed through the Host Capacity Model (HCM) lens I’ve been developing, may explain a lot more than viral replication kinetics. It may explain why post-COVID gut dysfunction lingers, why mitochondrial symptoms persist long after the virus is gone, and why our usual “stimulate or inhibit autophagy” framing for interventions misses the actual problem.
I want to lay out what I’m seeing. This is exploratory — I’m not making clinical claims. I am making a mechanistic argument that I think deserves more attention than it’s currently getting.
The double-edged sword that most coverage gets half-right
The standard summary goes: autophagy is the body’s cellular cleanup process; it normally clears intracellular pathogens through a specialized form called xenophagy; SARS-CoV-2 has evolved to hijack the machinery and block the cleanup. All of that is accurate, but the framing is too binary. The virus doesn’t simply “block” autophagy. It does something more interesting and, in my view, more dangerous: it lets the early steps of autophagy proceed and then prevents the late steps from finishing.
Specifically, viral proteins ORF3a and NSP6 interfere with autophagosome-lysosome fusion. The cell senses stress, builds autophagosomes around damaged mitochondria, aggregated proteins, and viral debris — and then the system stalls. The autophagosome never delivers its cargo to the lysosome for degradation. The trash gets bagged but never taken out.
At the same time, the virus repurposes components of the autophagic machinery to build double-membrane vesicles that house viral replication. So we’re not looking at suppression — we’re looking at redirection plus terminal blockade. The cell is doing the work but the work never completes.
This is the part that matters: incomplete autophagy is worse than no autophagy. A cell that never initiated cleanup at least isn’t accumulating sealed bags of damaged cargo. A cell that initiates and then stalls is filling up with autophagosomes that contain exactly the material most likely to trigger inflammation when it leaks — damaged mitochondria with exposed mtDNA, aggregated proteins, oxidized lipids. The buildup itself becomes a damage signal.
Where this collides with the HCM framework
If you’ve read anything I’ve written about colonocyte bioenergetics, you know the core argument: chronic gut dysfunction in many of these complex cases is not primarily a microbial problem. It’s an energetic and epigenetic problem. The colonocyte loses its capacity to oxidize butyrate. The mitochondrial pool degrades. NAD⁺ collapses through CD38-driven catabolism. SIRT3 can no longer keep the mitochondrial proteome deacetylated. Iron-sulfur cluster biogenesis falters. SLC5A8 gets epigenetically silenced. The system locks into a state where it can’t recover on its own, even after the original insult is gone.
Autophagy — specifically mitophagy — is the maintenance function that should be preventing this. Every healthy mitochondrion in a colonocyte exists in dynamic equilibrium with the mitophagy system that culls damaged ones. Disable mitophagy and the damaged mitochondria accumulate. ETC efficiency drops. mtROS rises. mtDNA leaks into the cytosol and activates cGAS-STING. CD38 gets induced by the resulting inflammatory tone. NAD⁺ drops. SIRT3 dysfunction follows. SOD2 stays acetylated and inactive. More ROS. More damage. The loop becomes self-sustaining.
This is exactly the cascade I’ve been writing about as the upstream driver of HCM failure — and SARS-CoV-2 sits at the top of it by disabling the one process that would normally prevent the loop from closing.
The virus doesn’t have to persist indefinitely for this to matter. It just has to block mitophagy long enough during acute infection for the loop to become self-sustaining. Once CD38 is elevated and NAD⁺ is depleted and SIRT3 is compromised, the loop continues to run even after viral antigens are cleared. The cell is now stuck in a state where it can’t restore mitochondrial quality on its own, because the substrate it needs to do so is being catabolized faster than it can be made.
The colonocyte makes this worse
There’s a tissue-specific twist that I haven’t seen articulated quite this way. Colonocytes are uniquely dependent on β-oxidation of butyrate for their energy. Butyrate enters primarily through SLC5A8. Once inside, it gets oxidized in mitochondria. So the colonocyte needs:
Functional SLC5A8 to import butyrate.
Functional mitochondria to oxidize butyrate.
Functional mitophagy to maintain that mitochondrial pool.
Now consider what happens when COVID blocks mitophagy in colonocytes:
Damaged mitochondria accumulate; β-oxidation capacity drops.
Butyrate isn’t oxidized as quickly; intracellular butyrate rises.
Butyrate at high intracellular concentrations acts as an HDAC inhibitor.
HDAC inhibition can silence specific genes through altered chromatin context — including, in this tissue, SLC5A8 itself.
SLC5A8 expression drops; butyrate import decreases; the colonocyte loses its preferred fuel.
The cell shifts toward glucose fermentation. Lactate rises. The local luminal environment shifts. Microbial ecology follows.
This is the butyrate paradox I wrote about earlier, but now I see where the trigger could come from. The viral block on mitophagy doesn’t just damage mitochondria — it sets in motion an epigenetic reorganization of the colonocyte that locks in the dysfunction.
If this picture is right, the gut symptoms of Long COVID aren’t a separate phenomenon from the mitochondrial fatigue and brain fog. They’re the same cascade expressed in a tissue with a particular metabolic vulnerability.
Why fasting helps some patients and hurts others
Extended fasting and intermittent fasting both trigger autophagy through mTOR inhibition and AMPK activation. The clinical reports from Long COVID patients are mixed: some describe dramatic improvements in fatigue and cognition; others crash and feel worse.
The incomplete-autophagy model offers a clean explanation. Fasting acts upstream — it tells the cell to initiate more autophagy. But if ORF3a, NSP6, or downstream consequences of their activity are still blocking autophagosome-lysosome fusion, fasting just produces more sealed bags of trash that don’t get processed. The patient feels the early mTOR-related benefits (improved insulin signaling, ketone metabolism, reduced inflammatory tone) and then crashes as the unprocessed autophagosomes accumulate and leak damage signals.
The therapeutic question, then, isn’t “should we stimulate autophagy?” It’s “where is the block, and how do we restore flux?”
What I’m exploring next
I want to write more on this with specific mechanistic detail, but here’s where I’m directing my reading:
Lysosomal function as the rate-limiting step. TFEB is the master regulator of lysosomal biogenesis and ALP gene expression. If the block is at fusion or lysosomal acidification, restoring TFEB activity matters more than triggering more autophagosomes. Nutrient signaling, calcium signaling, and Nrf2 all feed into TFEB.
Nrf2 activation as a flux restorer. This is where the work I’m doing with Dr. Christine Houghton becomes directly relevant. Sulforaphane is one of the most potent Nrf2 activators we have, and Nrf2 drives expression of multiple autophagy-lysosome pathway genes. If the problem in post-COVID dysfunction is a sustained blockade of ALP flux, sulforaphane is one of the few interventions with a credible mechanism for re-opening the pipe.
NAD⁺ restoration as a parallel intervention. Without NAD⁺, SIRT3 can’t function, and without SIRT3 function, mitophagy quality control is compromised even if autophagy flux is restored. NAD⁺ precursors paired with CD38 modulation may need to be running alongside anything that re-opens the autophagy pipe.
Spermidine and trehalose as mTOR-independent inducers. Both have evidence for promoting autophagy through pathways that bypass mTOR, and both have signals in viral and post-viral contexts. They’re not magic, but in the right patient they may be useful adjuncts.
The timing question. Acute COVID and Long COVID likely require opposite interventions. During acute infection, you may want to limit the autophagy machinery the virus is using to replicate. After clearance, you may want to maximally restore flux to clear residual damage. The wrong intervention at the wrong stage probably explains a good portion of the contradictory clinical reports.
What I think is novel here
The autophagy-COVID literature is large but mostly virological. The Long COVID mitochondrial literature is growing but mostly descriptive. The HCM framework I’ve been building treats colonocyte bioenergetic failure as the upstream node in chronic gut dysfunction. What I haven’t yet seen articulated is the integration:
SARS-CoV-2’s blockade of autophagy flux is a credible upstream trigger for the exact bioenergetic-epigenetic loop that the HCM places at the center of chronic dysfunction — and the colonocyte’s specific metabolic profile makes it the tissue where this trigger most easily locks in.
If this is right, it reframes Long COVID gut dysfunction. It’s not a microbial problem with metabolic consequences. It’s a metabolic problem with microbial consequences, set in motion by a viral interference with cellular cleanup that the affected tissues have particular difficulty resolving on their own.
I’m going to keep pulling on this. If you work in this space and you see where I’m wrong, I want to hear it. The individual pieces all have literature behind them. The synthesis is mine and synthesis is exactly where reasoning errors hide.
More to come.