Long COVID, Candida, and Klebsiella: When the Gut Loses Its Ecological Control

One of the patterns I have been paying much closer attention to in post-COVID cases is the repeated loss of normal gut ecology. In some of my clients after COVID, I have seen depletion of beneficial anaerobic organisms, reduced butyrate-producing bacteria, expansion of facultative organisms such as E. coli, Klebsiella, Citrobacter and Enterobacter, evidence of barrier dysfunction and altered inflammatory markers, and in some cases elevated levels of Candida albicans. It is easy to look at those findings and call them dysbiosis, but I think that description is incomplete. The question I keep asking is much more fundamental: what changed in the host that allowed these organisms to expand in the first place?

That question is central to what I call the Host Capacity Model. I do not think the microbiome determines its composition independently of the host. The host continuously constructs the environment in which microorganisms have to compete. It regulates oxygen availability, nitrate production, bile acids, mucus, antimicrobial peptides, secretory IgA, epithelial metabolism, intestinal transit, nutrient availability, immune pressure, barrier integrity and the physical distance between microbes and intestinal tissue. If those conditions change, microbial ecology changes with them. So when I see elevated E. coli, Klebsiella or Candida, I am not only asking what can suppress the organism. I am asking what host function was lost that normally prevented that organism from becoming dominant.

But I think the model has to go one step further. Once the microbial ecosystem changes, bacteria and fungi do not remain passive. They begin generating metabolites, inflammatory molecules and toxins that act back on the host. Those products can alter epithelial signaling, mitochondrial respiration, redox balance, immune activation, iron-sulfur proteins and neuroactive metabolic pathways. The sequence may therefore begin with host dysfunction selecting a new microbial ecosystem, but after that ecological shift occurs, the new microbial ecosystem can begin damaging the same host systems that originally allowed it to expand. The direction of causality becomes circular. The host changes the microbiome, and then the changed microbiome begins changing the host.

One possible upstream entry point after COVID is intestinal ACE2. Most people know ACE2 because SARS-CoV-2 uses it as an entry receptor, but ACE2 also performs important physiological functions in the intestine. ACE2 is highly expressed on absorptive enterocytes, particularly in the small intestine, where it associates with the neutral amino-acid transporter B0AT1, encoded by SLC6A19. ACE2 helps B0AT1 localize correctly at the intestinal brush border, and B0AT1 transports several neutral amino acids, including tryptophan.

This is important because tryptophan is not simply a serotonin precursor. Tryptophan participates in nutrient sensing, mucosal immunity, microbial metabolism, antimicrobial defense and epithelial signaling. If intestinal ACE2 physiology becomes disturbed, one possible downstream consequence is altered B0AT1-dependent amino-acid transport, which may then alter intracellular tryptophan availability and downstream nutrient-sensing pathways such as mTOR. That gives us a possible sequence in which intestinal injury after SARS-CoV-2 alters ACE2-related physiology, changes B0AT1-dependent amino-acid handling, changes tryptophan availability, alters antimicrobial signaling and weakens ecological control of the intestinal microbiome. I do not think this explains every case of Long COVID, and I do not assume persistent ACE2 dysfunction in every person, but it provides a biologically coherent mechanism through which intestinal injury could begin changing the host environment before the microbial abnormalities become obvious.

Tryptophan becomes even more interesting when we look at what happens to it after it enters the intestine. The host can metabolize tryptophan through the serotonin pathway or through the kynurenine pathway, while intestinal microbes can convert tryptophan into indole-derived metabolites. Some of these microbial metabolites activate the aryl hydrocarbon receptor, AhR. This creates an important communication pathway between the microbiome, mucosal immune system and epithelial surface.

Certain microbial indole metabolites can activate AhR in mucosal immune cells, including populations involved in IL-22 production. This is one of the reasons I keep coming back to IL-22 in my Host Capacity Model. IL-22 sits almost exactly at the interface between the immune system and the intestinal epithelium. It is produced by immune populations including ILC3 cells and certain T-cell subsets, but one of its most important intestinal targets is the epithelium itself. I think of IL-22 as one of the signals through which the mucosal immune system tells the epithelial layer that something has changed in the microbial environment and that antimicrobial defense, repair and barrier protection need to increase.

When IL-22 reaches epithelial cells, it activates signaling pathways including STAT3 and can promote antimicrobial programs, epithelial regeneration, barrier repair and changes in epithelial glycosylation. This matters because the intestinal barrier is not simply a passive wall. The epithelium is constantly sensing the microbial environment, replacing damaged cells, secreting antimicrobial molecules, maintaining mucus, remodeling junctions and controlling the physical distance between the microbiome and the epithelial surface. A healthy microbiome does not mean bacteria are allowed to sit directly against intestinal cells. The host actively controls microbial geography.

That is why I am interested in the entire tryptophan-AhR-IL-22 pathway rather than simply asking whether tryptophan is low. If beneficial organisms responsible for normal tryptophan metabolism disappear, the host does not merely lose bacterial diversity. It may lose an important set of microbial signals that normally support epithelial defense. Less appropriate microbial indole production may mean altered AhR signaling, altered IL-22 signaling, weaker antimicrobial defense and weaker epithelial repair. Once microbial organisms can approach the epithelium more closely, microbial products gain greater access to pattern-recognition receptors, inflammatory signaling increases and the host begins generating the same respiratory compounds that can favor facultative organisms.

This pathway also changes how I think about Candida albicans. I have seen elevated Candida albicans in some clients after COVID, particularly in people who have also had repeated antibiotic exposure or major depletion of beneficial anaerobic organisms. I do not automatically interpret that as a fungus that simply appeared and now needs to be killed. I ask what happened to colonization resistance. Were bacterial competitors lost? Did microbial tryptophan metabolism change? Did AhR ligand production change? Did IL-22-dependent mucosal defense become inadequate? Did secretory IgA fall? Did mucus integrity change? Did oxygen availability increase? Those are much more important ecological questions.

Once Candida albicans becomes more active, it can also begin contributing back to epithelial injury. During invasive or hyphal growth, Candida albicans can produce candidalysin, a peptide toxin capable of damaging epithelial cells and activating inflammatory signaling. That means a damaged ecological niche may allow Candida to expand, and the expanding fungus can then create more epithelial stress, which increases inflammation and permeability, which further changes the environment for both bacteria and fungi. Again, the relationship becomes circular.

This brings me to one of the most important parts of the Host Capacity Model: the host does not control the microbiome only through immunity. It also controls it through bioenergetics. The healthy colon is profoundly anaerobic, but it is not anaerobic because oxygen never reaches it. Oxygen is continually delivered through the intestinal circulation. The important question is what happens to that oxygen before it reaches the lumen.

Healthy differentiated colonocytes consume large amounts of oxygen through mitochondrial oxidative metabolism, and one of their major fuels is butyrate. Butyrate is produced by obligate anaerobic bacteria through fermentation. Colonocytes absorb butyrate and oxidize it through mitochondrial metabolism. Butyrate is converted into acetyl-CoA, enters the TCA cycle, generates NADH and FADH2 and drives the mitochondrial electron transport chain. That process consumes oxygen, which makes the colonocyte an oxygen sink.

Oxygen arrives from the circulation, but healthy colonocyte mitochondria consume much of it before it can reach the microbial compartment. This creates a steep spatial oxygen gradient in which blood is oxygenated, oxygen progressively falls toward the epithelial surface and the intestinal lumen remains strongly anaerobic. This is what I mean by physiological epithelial hypoxia. I am not talking about starving intestinal tissue of oxygen. I am talking about maintaining the normal spatial distribution of oxygen that keeps the luminal microbial environment anaerobic.

The relationship is self-supporting. Obligate anaerobes produce butyrate, butyrate fuels colonocyte mitochondrial respiration, colonocytes consume oxygen, low oxygen preserves obligate anaerobes and those organisms continue producing butyrate. The host and microbiome maintain one another. Once one part of that loop collapses, the ecology can begin shifting in the opposite direction.

Imagine what happens after infection, inflammation, repeated antibiotics, severe dietary restriction or another major ecological disruption. Butyrate-producing organisms decline, butyrate production falls and colonocytes receive less of their preferred oxidative fuel. Their metabolism can shift away from high mitochondrial oxidation and toward greater glycolytic metabolism. The important consequence is that they consume less oxygen. The epithelial oxygen sink weakens and more oxygen reaches the mucosal surface.

That change can completely alter microbial competition. Facultative organisms such as E. coli can use oxygen for respiration, while many strict anaerobes cannot. The host has therefore changed the energetic rules of the ecosystem. This is where PPAR-gamma becomes one of the most important pieces of the model.

PPAR-gamma is a transcriptional regulator involved in cellular differentiation, lipid metabolism and inflammatory signaling. In colonocytes, microbial butyrate supports PPAR-gamma signaling. PPAR-gamma helps maintain the differentiated oxidative phenotype of the epithelial cell, supporting fatty-acid and butyrate oxidation and keeping mitochondrial oxygen consumption high. But PPAR-gamma also restrains expression of NOS2, the gene encoding inducible nitric oxide synthase, or iNOS. This means PPAR-gamma directly connects the two respiratory resources I am most interested in: oxygen and nitrate.

When butyrate and PPAR-gamma signaling are maintained, colonocytes continue consuming oxygen while excessive NOS2 activity is restrained. This simultaneously limits oxygen availability and nitrate generation at the microbial interface. I think of this as a form of respiratory restriction. The host is actively depriving facultative organisms of the terminal electron acceptors they would need to gain a major energetic advantage over the normal anaerobic ecosystem.

Now reverse that system. Butyrate declines, PPAR-gamma signaling declines, colonocyte oxidative metabolism falls, oxygen consumption falls and more oxygen reaches the microbial surface. At the same time, reduced PPAR-gamma restraint allows NOS2 and iNOS activity to rise. Nitric oxide and reactive nitrogen chemistry increase, and nitrate becomes more available. One host metabolic disturbance has now changed the microbial environment in two directions simultaneously. There is more oxygen and more nitrate.

That is a major ecological transition because many healthy colonic organisms rely heavily on fermentation, while facultative organisms can use respiration. During respiration, electrons derived from nutrients enter an electron transport chain and eventually have to be transferred to a terminal electron acceptor. Humans primarily use oxygen. Facultative bacteria are more flexible. An organism such as E. coli can use oxygen when it is available and can switch toward nitrate respiration when oxygen becomes limited but nitrate is present.

Nitrate therefore is not simply another nutrient. It allows the organism to conserve energy differently from carbon substrates that were already present. That can produce a major growth advantage over neighboring organisms restricted largely to fermentation. This is why I increasingly think microbial ecology needs to be interpreted in terms of electron donors, electron acceptors, respiratory capacity and host metabolism, rather than only in terms of which foods are feeding a particular organism.

So when I see elevated E. coli, the deeper question becomes: why is the host environment suddenly giving E. coli access to respiratory electron acceptors that were previously restricted?

Once facultative Gram-negative organisms expand, exposure to microbial inflammatory structures such as LPS increases. LPS activates innate immune signaling through pathways including TLR4. Inflammatory signaling increases, NOS2 and iNOS increase, nitric oxide increases, reactive oxygen and nitrogen chemistry increase and nitrate availability increases. Now the host inflammatory response is generating more of the same terminal electron acceptor that nitrate-respiring organisms can exploit. The organism expands further, more inflammatory microbial products reach the host, more inflammation develops and more nitrate is generated.

Meanwhile, continued disruption of obligate anaerobes reduces butyrate production even further. Lower butyrate means weaker PPAR-gamma signaling, lower epithelial oxidative metabolism and less oxygen consumption. More oxygen reaches the microbial surface. The system now reinforces itself from both directions. One loop is driven by inflammation and nitrate. The other is driven by loss of butyrate and oxygen leakage.

Inflammation can also create other respiratory opportunities. The intestinal lumen contains sulfur compounds such as thiosulfate. During inflammation, reactive oxygen species can oxidize thiosulfate into tetrathionate. Certain organisms, particularly Salmonella enterica serovar Typhimurium, can use tetrathionate as a terminal electron acceptor and gain a respiratory advantage over competing fermenters. I would not generalize this mechanism to every E. coli, Klebsiella or Citrobacter strain because respiratory capacity is species- and strain-dependent, but the larger principle is important: inflammation changes the respiratory chemistry of the gut and creates metabolic opportunities for organisms equipped to exploit it.

Electron acceptors are only one side of the equation. Organisms also need electron donors. Formate is one example. It is generated during microbial fermentation and can enter bacterial respiratory metabolism through formate dehydrogenases. Lactate is another important example. When colonocytes shift away from mitochondrial oxidative metabolism and toward glycolysis, host lactate production can increase. Certain facultative organisms can exploit that lactate. So epithelial metabolic dysfunction may simultaneously provide more substrate and more respiratory opportunity. The host can begin supplying more lactate as an electron donor while allowing more oxygen or nitrate to function as electron acceptors.

This is why I think simply calling the endpoint “dysbiosis” can hide the deeper biology. The energetic architecture of the ecosystem has changed.

Once dysbiosis becomes established, bacteria and fungi are not simply occupying space. They are chemically active organisms releasing metabolites, inflammatory molecules and toxins into an epithelium that may already have reduced capacity to handle them. The consequences depend on the organism, concentration, anatomical compartment, host detoxification capacity and integrity of the epithelial barrier.

Hydrogen sulfide is a good example of why concentration and capacity matter. The colon normally encounters microbially generated H2S, and colonocytes possess mitochondrial sulfide-oxidation machinery that can detoxify and metabolically process sulfide. At physiological concentrations, H2S participates in normal signaling and metabolism. But that capacity is finite. At sufficiently high exposure, H2S can inhibit mitochondrial cytochrome-c oxidase, Complex IV, and suppress respiration.

That creates another possible feedback loop. If microbial sulfide production exceeds epithelial handling capacity, mitochondrial respiration can fall. If mitochondrial oxygen consumption falls, more oxygen remains available near the epithelial surface. That can favor facultative organisms and further destabilize the anaerobic ecosystem. The point is not that hydrogen sulfide is universally harmful. The important variables are dose, compartment and host metabolic capacity.

This principle likely applies to many microbial metabolites. A molecule may be beneficial at one concentration and harmful at another. The key question is not whether the metabolite is inherently good or bad. The key question is whether the host has the biochemical capacity to process the amount being produced.

This brings me to another mechanism that I think may be important in chronic inflammatory gut states: iron-sulfur clusters. Iron-sulfur clusters are cofactors built from iron and sulfur atoms and are required for several critical metabolic proteins. Mitochondrial aconitase contains an iron-sulfur cluster. Complex I contains multiple iron-sulfur centers required for electron transfer. Complex II, succinate dehydrogenase, also contains iron-sulfur centers. These structures are essential for normal mitochondrial metabolism, but they are also vulnerable to oxidative and nitrosative stress.

The mechanism I am most interested in is not that every bacterial metabolite directly destroys iron-sulfur clusters. It is that microbial inflammatory products can increase iNOS, nitric oxide, superoxide and peroxynitrite, and that redox environment can damage Fe-S-dependent proteins and respiratory machinery. If mitochondrial aconitase becomes impaired, normal TCA-cycle flux changes. If Complex I becomes impaired, NADH oxidation and electron flow toward the CoQ pool are disturbed. If Complex II becomes impaired, succinate oxidation and electron transfer toward ubiquinone are affected.

This means Fe-S damage can change much more than ATP output. It can alter NADH/NAD ratios, succinate handling, ROS production, TCA-cycle flux and mitochondrial signaling. That becomes particularly relevant because succinate is not simply a TCA-cycle intermediate. It can also function as a signaling metabolite through SUCNR1 and participate in inflammatory and oxygen-sensing pathways.

So the loop can become even deeper. Dysbiosis increases inflammatory microbial exposure. Inflammation increases ROS and reactive nitrogen species. Oxidative and nitrosative stress damage Fe-S-dependent enzymes and respiratory proteins. Mitochondrial respiration becomes less efficient. Colonocytes consume less oxygen. More oxygen reaches the microbial surface. Facultative organisms gain a larger respiratory advantage. Their expansion increases inflammatory signaling again.

In other words, microbial inflammation may begin damaging the mitochondrial machinery that normally maintains the anaerobic microbial niche.

The intestinal barrier then becomes part of the same problem. The barrier is metabolically expensive. Epithelial cells need ATP for cytoskeletal organization, junction remodeling, membrane trafficking, ion gradients, mucus secretion, cell migration, wound healing and constant cellular turnover. Butyrate oxidation is one major part of that energy system, while the creatine-phosphocreatine system may provide another layer of rapid energetic buffering.

If mitochondrial oxidative metabolism becomes impaired while cellular energy-buffering systems are also compromised, epithelial resilience may fall. I am not suggesting that creatine deficiency explains post-COVID gut dysfunction, but I do think epithelial energy capacity deserves much more attention. A barrier can only remain functional if the cells maintaining it have enough energy to continuously repair and regulate it.

Once barrier integrity deteriorates, microbial products gain greater access to the mucosal immune system. More LPS reaches innate immune receptors, inflammatory signaling rises, NOS2 and iNOS increase, nitrate production rises and respiratory opportunity increases. Barrier dysfunction is therefore not necessarily a separate downstream consequence. It can become another amplifier of the same ecological loop.

The normal oxygen gradient is also sensed by HIF, hypoxia-inducible factor. HIF helps epithelial cells adapt to their physiological low-oxygen environment and regulates pathways involved in barrier protection, metabolism and tissue resilience. This is why I think it is important to distinguish physiological epithelial hypoxia from pathological tissue hypoxia. I am not trying to make the intestine oxygen-deprived. I am interested in restoring the normal condition in which blood remains adequately oxygenated while healthy epithelial metabolism consumes enough oxygen that very little reaches the luminal microbiome.

An inflamed intestine can even contain both states simultaneously. Deeper tissue may become pathologically hypoxic because inflammation increases oxygen demand or alters microvascular function, while reduced epithelial oxidative metabolism allows abnormal oxygen availability at the microbial interface. So the important question is not simply whether oxygen is high or low. The important question is where the oxygen is and which cells are consuming it.

There is also another branch of this model that may help explain some of the neurological symptoms I see in complex post-COVID cases, and this brings us back to tryptophan. During inflammation, cytokine signaling can increase activity of the kynurenine pathway. One important entry enzyme is IDO1, which directs tryptophan toward kynurenine metabolism. TDO and, in some settings, IDO2 can also contribute.

This matters because sustained inflammatory activation changes the metabolic destination of tryptophan. Instead of asking simply whether tryptophan is high or low, I think about flux. How much tryptophan is entering microbial indole metabolism? How much remains available for serotonin synthesis? How much is being directed into the kynurenine pathway? And what metabolites are being produced downstream of kynurenine?

This is where I am particularly interested in quinolinic acid. One route of the kynurenine pathway moves from tryptophan to kynurenine and then through downstream intermediates including 3-hydroxykynurenine and 3-hydroxyanthranilic acid before eventually reaching quinolinic acid. Quinolinic acid is also a normal intermediate in de novo NAD synthesis, so it is not simply a toxin. Again, concentration and compartment matter.

But excessive quinolinic acid in the nervous system is important because quinolinic acid is an endogenous NMDA-receptor agonist. This creates a direct connection between inflammatory tryptophan metabolism and glutamatergic signaling.

An important compartment distinction is that peripheral quinolinic acid does not simply move freely from the intestine into the brain. Kynurenine crosses the blood-brain barrier much more readily. Once kynurenine enters the CNS, activated microglia and macrophage-lineage cells can metabolize it further and generate quinolinic acid locally.

This is one of the pathways I suspect may be relevant in a subset of clients with post-COVID neurocognitive symptoms, sensory sensitivity, sleep disruption, internal agitation and other manifestations that could be compatible with excessive excitatory signaling. I would not diagnose elevated quinolinic acid from symptoms alone, but the mechanism is important enough that I think it deserves investigation.

Quinolinic acid does not only activate NMDA receptors directly. It can also increase glutamatergic signaling by stimulating glutamate release and impairing astrocytic glutamate uptake. Normally, astrocytes continuously remove extracellular glutamate and recycle it through the glutamate-glutamine system. If glutamate release increases while glutamate clearance becomes less efficient, extracellular glutamate can rise. At the same time, quinolinic acid itself is stimulating the NMDA receptor.

So in the same system you can potentially have greater glutamate release, less efficient glutamate clearance and an additional endogenous NMDA agonist present at the same time.

This is why I think quinolinic acid may be more important than simply describing someone as having “high glutamate.” It gives us a mechanistic pathway through which inflammation could increase glutamatergic tone.

IDO activation can also change the competition for tryptophan. If inflammation drives more tryptophan toward kynurenine metabolism, less substrate may remain available for other metabolic destinations, including serotonin synthesis. I do not think it is accurate to say that increased IDO automatically causes low serotonin because serotonin metabolism is compartment-specific and depends on transport, tryptophan hydroxylase activity, enterochromaffin biology, neuronal metabolism and many other factors. The more precise way to think about it is metabolic repartitioning. Chronic inflammation may shift more tryptophan toward kynurenine metabolism and potentially away from other pathways.

The IL-22 relationship also needs to be treated carefully. I do not think the pathway is simply IDO rises, therefore IL-22 falls. Kynurenine itself can interact with AhR, so IDO, AhR and IL-22 do not form a simple linear suppression pathway. What I am more interested in is whether the quality, source and location of AhR signaling changes.

In a healthy intestinal ecosystem, microbial metabolism converts tryptophan into indole-derived AhR ligands directly at the mucosal surface. Those metabolites interact with ILC3 cells and other immune populations and help support an organized IL-22-mediated epithelial defense program. If beneficial indole-producing organisms are lost while host inflammatory tryptophan catabolism increases, the system may shift away from locally organized microbial indole-AhR-IL-22 signaling and toward a very different kynurenine-dominant metabolic environment.

That does not guarantee low IL-22, but it may mean that the normal mucosal signaling architecture is disrupted. If IL-22-mediated epithelial antimicrobial defense and repair become insufficient or dysregulated, barrier repair may become less effective, microbial proximity increases and inflammatory signaling rises. That feeds directly back into the nitrate and oxygen loops.

This creates a possible gut-brain feedback system. Intestinal dysbiosis increases inflammatory microbial signals. Barrier dysfunction increases immune exposure. Systemic inflammation increases kynurenine-pathway flux. Circulating kynurenine can enter the CNS. Activated CNS immune cells can metabolize kynurenine toward quinolinic acid. Quinolinic acid can stimulate NMDA receptors, increase glutamate release and impair astrocytic glutamate clearance. Increased glutamatergic signaling and NMDA activation increase intracellular calcium demand and can increase oxidative stress and mitochondrial burden.

Now the loop is no longer confined to the intestine.

CNS dysfunction can feed back through autonomic pathways that regulate intestinal motility, secretion, vascular tone, immune signaling and gut transit. So the gut can influence the CNS, and the CNS can influence the gut.

This is why I do not think the gut, immune system, mitochondria and brain should be interpreted as separate systems in chronic post-infectious disease.

When I put all of these pathways together, the model becomes much larger than the statement that COVID causes dysbiosis. A viral or inflammatory intestinal insult can alter ACE2-related physiology and amino-acid handling. Tryptophan signaling changes. Antimicrobial control may weaken. Microbial tryptophan metabolism and AhR-IL-22 signaling may change. Obligate anaerobes decline. Butyrate falls. PPAR-gamma signaling weakens. Colonocyte mitochondrial oxidation falls. The epithelial oxygen sink weakens. More oxygen reaches the microbial surface.

At the same time, inflammation increases NOS2 and nitrate production. Facultative organisms gain access to oxygen and nitrate and obtain a respiratory advantage. They expand. Their LPS and other microbial products stimulate more inflammation. Fungal organisms such as Candida albicans may gain an ecological opportunity and can add further epithelial stress. Microbial and inflammatory chemistry increases ROS and reactive nitrogen species. Those molecules can impair Fe-S-dependent enzymes and mitochondrial respiratory machinery. Mitochondrial respiration becomes less efficient. Colonocytes consume less oxygen. More oxygen reaches the microbial niche.

Barrier energy capacity falls. Permeability increases. More microbial products reach the immune system. Inflammation increases again. At the same time, inflammatory signaling redirects more tryptophan toward kynurenine metabolism. Kynurenine can enter the CNS. Activated CNS immune cells can generate quinolinic acid. Quinolinic acid can amplify NMDA and glutamate signaling. Neuroinflammation and autonomic dysfunction can then feed back into intestinal physiology.

That is the broader loop I am trying to understand.

Antibiotics can also push this system in the same direction. Antibiotics can be necessary and lifesaving, but if they remove butyrate-producing anaerobes, butyrate falls, PPAR-gamma signaling can weaken, colonocyte oxidative metabolism falls and more oxygen reaches the microbial surface. Loss of bacterial competition can also weaken fungal colonization resistance. This is particularly relevant to what I have seen clinically because some of my post-COVID clients with elevated Candida albicans also have histories of repeated antibiotic exposure and major depletion of beneficial anaerobes.

Again, I am not saying that oxygen explains every case of Candida or that antibiotics explain every post-COVID microbial abnormality. I am saying the ecological pattern makes sense. The treatment may remove organisms that were helping maintain the very host environment required for colonization resistance.

This is also why recurrence after antimicrobial treatment does not always mean the antimicrobial was too weak. Imagine reducing E. coli temporarily while butyrate remains low, PPAR-gamma signaling remains impaired, colonocyte mitochondrial respiration remains weak, Fe-S-dependent metabolic pathways remain under oxidative or nitrosative stress, oxygen continues reaching the microbial interface, NOS2 remains elevated, nitrate remains available, the barrier remains compromised, IL-22-mediated repair remains inadequate and motility remains abnormal.

The organism may return because the environment that selected for it never changed.

You changed the population.

You did not necessarily change the selection pressure.

The same principle may apply to Candida albicans. If the ecological systems that normally constrain fungal growth remain impaired, killing the fungus does not automatically recreate the environment that kept it suppressed in the first place.

The opposite mistake would be looking at this model and assuming that simply adding butyrate solves the problem. I do not think it is that simple either. A metabolite being present does not mean the cell can use it correctly. Butyrate has to reach the colonocyte, enter the cell, be activated metabolically, move through beta-oxidation, generate acetyl-CoA, enter the TCA cycle, pass through aconitase and other enzymes, generate NADH and FADH2, feed the respiratory chain and ultimately increase oxygen consumption.

If Fe-S-dependent enzymes are damaged, if Complex I or Complex II is impaired, if the CoQ pool is dysfunctional, if Complex IV is inhibited or if the cell is otherwise unable to sustain oxidative metabolism, simply increasing luminal butyrate may not reproduce normal physiological flux.

That is why I care much more about metabolic flux than simply metabolite concentration.

The same principle applies to PPAR-gamma, IL-22, tryptophan, creatine and almost every other pathway in this model. Mechanism should tell us what needs to be investigated. It should not automatically become another supplement.

One of the major areas I am actively working on in the Host Capacity Model is therefore how to restore physiological epithelial hypoxia. Again, I do not mean making intestinal tissue hypoxic. I mean restoring the host’s ability to consume oxygen in the correct compartment so that very little oxygen reaches the luminal microbial environment.

The questions I am increasingly asking are whether the anaerobic ecosystem can recover, whether butyrate-producing capacity can recover, whether the colonocyte can actually oxidize butyrate, whether PPAR-gamma signaling can normalize, whether mitochondrial respiration can recover, whether excessive NOS2 and nitrate production can decline, whether sulfide exposure is within epithelial handling capacity, whether Fe-S-dependent enzymes can recover from inflammatory redox stress, whether the epithelial barrier can regenerate, whether microbial tryptophan metabolism can recover, whether AhR-IL-22 signaling becomes appropriately organized again, and whether secretory IgA, mucus, bile acids and motility can return to a state that makes facultative organisms less competitive.

That is what I mean by restoring Host Capacity.

The microbiome is not simply a collection of organisms. It is an ecosystem constrained by metabolism. The host regulates oxygen. Inflammation generates nitrate. The epithelium produces metabolic substrates such as lactate. The microbiome produces butyrate, indoles, sulfide, formate and thousands of other molecules. Bacteria transform tryptophan into AhR ligands. Immune cells produce IL-22. IL-22 communicates back to the epithelium. Fungi interact with the barrier and can produce virulence factors capable of injuring epithelial cells. Microbial inflammatory products increase ROS and reactive nitrogen species. Those molecules can affect mitochondrial respiratory proteins and iron-sulfur-dependent enzymes. Mitochondrial dysfunction changes oxygen consumption. Changes in oxygen consumption alter microbial ecology. Inflammation changes tryptophan flux. Kynurenine can connect peripheral inflammation with the CNS. Quinolinic acid can connect inflammatory tryptophan metabolism with NMDA-receptor and glutamate signaling.

Every one of these pathways can feed into another.

That is why I increasingly think we cannot properly understand the microbiome without simultaneously thinking about mitochondrial bioenergetics, immunology, microbial respiration, epithelial biology, redox chemistry and neuroimmune metabolism.

When I see elevated E. coli, Klebsiella, Citrobacter, Enterobacter or Candida albicans after COVID, I do not automatically see those organisms as the beginning of the disease. Sometimes they may initially be the ecological readout of a host environment that has already changed. Once they expand, however, their metabolites, inflammatory signals and toxins can make them part of the mechanism maintaining that altered environment.

The sequence I am investigating is therefore not simply that COVID causes dysbiosis. It is that COVID or another major insult may change host physiology, altered host physiology changes the microbial niche, the altered niche selects a different microbial ecosystem, that microbial ecosystem generates metabolites and inflammatory signals that further impair epithelial and mitochondrial function, mitochondrial dysfunction changes oxygen consumption, inflammation generates nitrate, oxygen and nitrate select for facultative respiration, barrier failure increases immune exposure, inflammatory signaling changes tryptophan metabolism, kynurenine can contribute to CNS quinolinic-acid production and glutamatergic signaling, and the altered host continues selecting for the same microbial ecosystem.

That is the loop I am trying to understand.

And it is why the central question in my Host Capacity Model is becoming less:

What do I need to kill?

and increasingly:

What did the host lose the capacity to do?

Because if we can understand and restore the host functions that maintained the original ecosystem, the microbiome may no longer have the same metabolic reason to remain dysbiotic.

That, to me, is a much deeper way of thinking about the gut after COVID.