The NAD+ Trap: How One Enzyme Locks Long COVID's Immune Collapse in Place

In my last post, I described how bone marrow stem cells are epigenetically reprogrammed in Long COVID, producing a dysfunctional monocyte called LC-Mo. I mapped the vicious cycle: IL-6 reprograms HSPCs, LC-Mo monocytes cause fibrosis and clotting, tissue damage releases heme, heme re-trains the stem cells. The cycle is self-sustaining.

But I left a question open. What is the molecular switch that keeps individual monocytes locked in their exhausted state? And how does the dysfunction spread — not just from stem cell to daughter cell, but from one monocyte to the next?

The answer is an enzyme called CD38. And the mechanism is worse than I expected.

The Enzyme That Eats Your Fuel

CD38 is an NAD+ hydrolase — an enzyme whose primary job is to consume nicotinamide adenine dinucleotide, one of the most important metabolic cofactors in every cell. NAD+ powers hundreds of enzymatic reactions. It is essential for mitochondrial energy production, DNA repair, and immune cell function. Without adequate NAD+, a cell cannot do its job.

In Long COVID, CD38 is robustly and persistently elevated on monocytes. This isn't subtle upregulation. It's a defining marker of the exhausted state — so reliable that researchers use CD38 expression to identify exhausted monocytes in the first place.

The problem: an enzyme that degrades your cellular fuel is now overexpressed on the very immune cells that need that fuel to fight infection. The monocytes are starving themselves.

The Feedback Loop

This would be bad enough as a one-way street. But CD38 overexpression creates a positive feedback loop — a molecular trap with no natural exit.

Here is the trap, step by step. CD38 is upregulated on monocytes during infection. CD38 consumes NAD+. NAD+ depletion impairs mTORC2 signaling, which disables Akt. Without Akt, the cell loses PGC1α/β (mitochondrial biogenesis), CREB (transcriptional regulation), and CD86 (antigen presentation). The monocyte becomes functionally crippled.

But here's the lock. While mTORC2 fails, mTORC1 stays active. Active mTORC1 phosphorylates Src, which activates STAT1 and STAT3 — transcription factors that drive more CD38 expression. More CD38 means more NAD+ consumption. The cycle accelerates. There is no internal brake.

This is not just a side effect of infection. It is a self-reinforcing molecular trap that maintains itself indefinitely once triggered.

The Part Nobody Expected: Exhaustion Spreads

If the NAD+ trap only affected individual monocytes, the body might eventually replace them. Monocytes live days, not months. But the Cell Communication and Signaling study from December 2025 revealed something more disturbing.

Exhausted monocytes don't just stay broken. They break their neighbors.

The implications for Long COVID are immediate. Even if the bone marrow eventually stops producing LC-Mo monocytes — even if the epigenetic reprogramming I described in Post #19 were somehow reversed — any remaining exhausted monocytes in circulation or tissue could re-seed the dysfunction by converting their naïve neighbors. The system has redundant maintenance: top-down (stem cell reprogramming) AND lateral (cell-to-cell propagation).

There's more. The exhausted monocytes don't just suppress immune function — they actively damage tissue. They promote endothelial cell apoptosis. They upregulate ICAM-1 and VCAM-1 on blood vessel walls, increasing monocyte adhesion and transmigration. They suppress T cell proliferation and activation via PD-L1. The broken cells aren't passive — they are actively dismantling the infrastructure around them.

Why the Memory Sticks

A positive feedback loop explains how the trap maintains itself from moment to moment. Cell-to-cell propagation explains how it spreads. But neither explains why the exhaustion persists even when the triggering stimulus is removed.

The answer is a protein called TICAM2, also known as TRAM.

TICAM2 is an adaptor protein in TLR4 signaling — the same pathway that mast cells use to drive neuropathy, which I covered in Post #16. In the context of monocyte exhaustion, TICAM2 serves a different role: it is the epigenetic memory keeper.

A 2025 study in Scientific Reports showed that TICAM2 maintains monocyte exhaustion through DNA methylation at CEBPE binding sites. CEBPE is a transcription factor that regulates myeloid cell differentiation and function. When TICAM2 holds these sites in an altered methylation state, the monocyte's gene expression is locked into the exhausted program — regardless of what signals it receives from outside.

When TICAM2 was genetically ablated in a murine sepsis model, monocytes recovered from exhaustion. The methylation marks at CEBPE sites were released. Immune function restored.

TICAM2 doesn't cause the exhaustion. It prevents the cell from ever recovering from it.

This connects directly to the HSPC reprogramming described in Post #19. The bone marrow produces monocytes with altered epigenetic marks. TICAM2/TRAM holds those marks in place at the chromatin level. Even if the upstream signal weakens, the methylation pattern persists. The exhaustion is written into the DNA packaging and locked there by an adaptor protein that was never meant to serve this function.

Three Layers of Entrapment

The picture that emerges is not a single mechanism but three reinforcing layers, each sufficient on its own to maintain dysfunction, and devastating in combination.

Each layer operates on a different timescale. The metabolic trap locks cells within hours. Propagation spreads the damage over days. The epigenetic lock ensures it persists for months or years. Together, they explain why Long COVID's immune dysfunction doesn't self-correct — the system has been engineered, through evolutionary accident, for permanence.

The Double Hit: Spike Fragments Kill the Backup

There is a final layer that makes this worse. While monocytes are trapped in exhaustion, a separate mechanism is destroying the very cells that might compensate.

In January 2026, Zhang and Wong at UCLA published a remarkable finding in PNAS. When SARS-CoV-2 spike protein is cleaved by host trypsin-like serine proteases, it generates short peptide fragments — xenoAMPs (xenobiotic antimicrobial peptides) — that target and kill specific immune cells. Not through receptors. Through membrane physics.

These fragments detect cells by their membrane curvature — the physical shape of the cell surface, measured as Gaussian curvature. Plasmacytoid dendritic cells (pDCs) and CD8+ T cells have membrane geometries that make them vulnerable. The fragments insert, permeabilize, and kill. Omicron-derived fragments are less lethal, which may explain milder variant outcomes.

The relevance to monocyte exhaustion: pDCs are the body's primary producers of type I interferons, the alarm signals that coordinate antiviral immunity. CD8+ T cells are the primary killers of virus-infected cells. Both are being selectively depleted by spike fragments while the monocyte population — the cells that should coordinate the broader response — is locked in an exhausted state where it can't present antigens, can't activate T cells, and is actively suppressing the immune cells around it.

The virus removes the backup. The exhaustion removes the primary. The immune system has been disabled from two independent directions.

Breaking the Trap

Three compounds have been shown to disrupt the CD38-mTOR exhaustion axis in experimental systems. None has been tested in a Long COVID clinical trial. All target different nodes of the trap.

The NAD+ Supplement Question

An obvious question: if the problem is NAD+ depletion, why not just supplement NAD+? NMN (nicotinamide mononucleotide) and NR (nicotinamide riboside) are widely available and marketed for exactly this purpose.

The answer is that supplementation may be futile while CD38 remains overexpressed. CD38 doesn't just degrade intracellular NAD+ — it also degrades NMN extracellularly through its ecto-enzymic activity. You are trying to fill a bathtub with the drain open. The enzyme will consume the supplement as fast as you can deliver it.

This doesn't mean NAD+ precursors are worthless. In cells where CD38 is not overexpressed, NMN and NR can effectively raise NAD+ levels. But for the exhausted monocyte population specifically, you likely need to inhibit CD38 first, then restore NAD+. The order matters. Supplementation without enzyme inhibition is treating the symptom while the cause runs unchecked.

Where This Leaves Us

In Post #19, I described the vicious cycle: reprogrammed stem cells producing broken monocytes, tissue damage feeding back to the bone marrow. That was the macro view — the system-level architecture of Long COVID's immune dysfunction.

This post is the micro view. Inside each of those broken monocytes, CD38 is running a self-amplifying metabolic trap. The trapped monocytes spread their dysfunction to their neighbors. TICAM2 locks the damage in place at the DNA level. And spike protein fragments are simultaneously killing the pDCs and T cells that might compensate.

The two views are not competing explanations. They are zoom levels on the same mechanism. Post #19's heme-driven vicious cycle explains why the bone marrow keeps producing dysfunctional monocytes. This post's CD38-mTOR axis explains why each individual monocyte stays dysfunctional — and why the dysfunction spreads horizontally through the immune system, not just vertically from stem cell to daughter cell.

No current Long COVID clinical trial targets CD38. No trial targets TICAM2. No trial uses mycolic acid derivatives. The compounds that directly address the molecular lock behind monocyte exhaustion exist in preclinical systems. The clinical pipeline — REVERSE-LC, LC-REVITALIZE, ADDRESS-LC — targets upstream cytokines and downstream fibrosis, which matters, but leaves the lock itself untouched.

Three days from now, on March 28, the RECOVER-AUTONOMIC trial results drop at ACC. That trial targets a different arm of the disease — autonomic dysfunction via heart rate control. I'll cover those results when they arrive. But the monocyte exhaustion axis described here is the deeper problem. It is the engine that drives the immune dysfunction that drives everything else.

The trap has a molecular address. The lock has a name. The question now is whether anyone will make the key.