Your genes don't change after a viral infection. But the instructions for reading them do. SARS-CoV-2 doesn't just infect and leave — it edits the chemical marks that tell your immune cells which genes to activate and which to silence. These edits persist for months, possibly years. They turn down your antiviral defenses while turning up chronic inflammation. And they may explain why Long COVID outlasts the virus that caused it.
This is epigenetics — the layer of molecular marks (methyl groups, acetyl groups, RNA modifications) that sit on top of DNA and control gene expression without altering the genetic sequence itself. Think of it as the difference between a musical score and a performance: the notes are the same, but the dynamics, tempo, and emphasis have been rewritten. In Long COVID, the immune system is playing the same genes — but in the wrong key.
Three distinct mechanisms of epigenetic reprogramming have been identified. They don't merely coexist. They cooperate.
Strike One: Silencing the Alarm
When SARS-CoV-2 enters a cell, one of its first proteins — NSP1 — begins a quiet sabotage. NSP1 is already known for blocking host translation (preventing cells from making their own proteins). But a 2024 study in PLOS ONE revealed something subtler: NSP1 induces the histone methyltransferase G9a to deposit repressive marks (H3K9me2) specifically at antiviral gene loci.
H3K9me2 is an epigenetic “keep out” sign. When placed on the histone proteins around which DNA is wound, it compacts the chromatin and prevents transcription factors from accessing the genes beneath. NSP1 doesn't need to enter the nucleus to do this — it operates through an intermediary protein called PRRC2B, recruiting G9a to silence the very genes the cell needs to fight the infection.
The effect is targeted. Integrating RNA-seq, ribosome footprinting, and ChIP-seq data, researchers found that NSP1 predominantly represses transcription of immune-related genes. Interferon-stimulated genes, pattern recognition receptors, antiviral effectors — the first-line alarm system — all dimmed.
The proof of mechanism came from the inhibitor. When researchers treated infected cells with UNC0638, a G9a inhibitor, antiviral gene expression was fully restored and viral replication dropped roughly tenfold. The alarm wasn't broken. It had been deliberately silenced — and could be turned back on.
A follow-up study in iScience (2025) confirmed that G9a is upregulated in COVID-19 patients with high viral loads. The virus doesn't just use G9a — it amplifies it. Patients who can't clear the virus efficiently may be the ones whose G9a-mediated silencing is most entrenched.
The Deeper Layer: Rewriting the RNA Playbook
The 2025 iScience study revealed that G9a's role extends far beyond histone modification. Through its interaction with METTL3 — the primary enzyme responsible for N6-methyladenosine (m6A) modification of RNA — G9a rewires the host's entire epitranscriptome.
m6A is the most common internal modification on mammalian mRNA. It affects RNA stability, translation efficiency, and localization. It's how cells fine-tune which proteins get made, how much, and for how long. SARS-CoV-2 hijacks this system through the G9a-METTL3 axis to simultaneously:
- Promote viral replication: m6A marks on the viral genome help it evade RIG-I innate immune sensing, suppressing the type I interferon response
- Drive T cell and NK cell exhaustion: m6A modifications upregulate PD-L1 and other exhaustion markers on host cells, disabling the adaptive immune clearance
- Amplify hyperinflammation: Cytokine and chemokine transcripts are stabilized and translated more efficiently
- Promote coagulation: Coagulation and angiogenesis-associated protein transcripts are upregulated
- Drive fibrosis: Fibrosis markers are enhanced through m6A-mediated translational control
G9a inhibition with UNC0642 reversed these changes in both infected cell lines and COVID-19 patient PBMCs. In vivo, UNC0642 reduced SARS-CoV-2 replication and infection-induced lung damage. The drug targets a host mechanism, not a viral one — meaning it works regardless of variant. The researchers noted “higher resistance barrier, broader activity against coronavirus strains/species, and potential synergy with other direct-acting antiviral drugs.”
This is what makes G9a extraordinary as a target. It sits at the intersection of two layers of epigenetic control — histone modification and RNA modification — and SARS-CoV-2 exploits both through the same enzyme. Blocking G9a doesn't just restore one set of genes. It rewrites both the chromatin landscape and the translational program back toward their pre-infection state.
Strike Two: Reprogramming the Factory
While NSP1 silences individual immune cells' defenses, a second epigenetic attack targets something more foundational: the stem cells that produce immune cells in the first place.
Cheong et al. published the key finding in Cell in 2023. During acute COVID-19, the IL-6 surge — the cytokine storm — reaches the bone marrow and epigenetically reprograms hematopoietic stem and progenitor cells (HSPCs). The reprogramming deposits activating marks (H3K4me1, H3K4me3, H3K27ac) at the promoters and enhancers of inflammatory gene loci. HSPCs don't just respond to inflammation during the acute infection — they remember it.
This is “central trained immunity.” The bone marrow becomes an inflammation factory. Every monocyte produced by these reprogrammed stem cells inherits the epigenetic marks — arriving in the bloodstream pre-loaded with open chromatin at inflammatory genes, ready to produce excessive cytokines at the slightest provocation. Since monocytes live only about one day, this inheritance mechanism is the only explanation for why hyper-inflammatory monocyte phenotypes persist for 12 months or longer after infection.
The reprogramming also skews hematopoiesis itself. Post-COVID HSPCs produce an excess of granulocyte-monocyte precursors (GMPs) relative to other lineages. The factory isn't just making inflammatory products — it's overproducing the inflammatory assembly line.
A Nature Immunology study (January 2026) added detail: Long COVID monocytes adopt a distinct transcriptional state, termed LC-Mo, characterized by profibrotic gene expression. These aren't normal monocytes behaving badly. They're fundamentally different cells, shaped by epigenetic marks set months earlier in the bone marrow.
The Synthesis: Brakes and Accelerator
The two strikes are complementary, not redundant.
Strike One (NSP1 → G9a → H3K9me2 + m6A rewiring) disables the brakes: antiviral defense genes are silenced, T/NK cells are exhausted, and the virus gains translational advantages. This helps the virus establish and maintain persistence.
Strike Two (IL-6 → HSPC reprogramming → trained immunity) floors the accelerator: inflammatory genes are permanently opened, monocytes are pre-armed for chronic inflammation, and the bone marrow produces them in excess. This drives the symptoms and tissue damage of Long COVID.
Together, they create a trap. The immune system can't clear the virus (because its antiviral programs are epigenetically silenced) but also can't stop responding to it (because its inflammatory programs are epigenetically amplified). Persistence and inflammation feed each other through two separate epigenetic mechanisms — and both outlast the acute infection.
This resolves a paradox that has puzzled Long COVID researchers: how can the immune system be simultaneously underperforming (failing to clear virus) and overperforming (producing chronic inflammation)? The answer is that these aren't contradictory states. They're two faces of the same epigenetic reprogramming — different marks on different genes, installed by different viral mechanisms, serving complementary functions.
The Scars That Remain
Do these epigenetic changes persist in chronic Long COVID? The direct evidence is still building, but converging lines suggest yes.
Balnis et al. identified 71 persistent differentially methylated regions (DMRs) in COVID-19 convalescents, with 90% located in gene promoter regions — including immune and inflammatory regulators. Nikesjo et al. (2025) found persistent DNA methylation alterations in lung immune cells of Long COVID patients, with pathway analyses highlighting immune dysregulation, Wnt signaling, and circadian rhythm involvement.
Zameer et al. (Clinical Epigenetics, 2026) extended this to neurological consequences: genome-wide DNA methylation profiling in COVID-positive patients revealed alterations in pathways linked to neurological and psychiatric dysfunction. The virus leaves epigenetic marks that reach beyond the immune system.
Epigenetic clock analyses show accelerated biological aging in severe COVID-19 patients — the molecular equivalent of the body aging faster than it should. Some researchers have described these as “irreversible epigenetic scars,” though others note partial restoration of normal profiles during recovery.
The open question is reversibility. Are these marks permanent, or do they decay? If IL-6 stops (because inflammation resolves, or because it's pharmacologically blocked), do the HSPC marks fade? If the virus is cleared (removing the trigger for NSP1 and G9a), does antiviral gene expression recover spontaneously?
Can We Rewrite the Rewrite?
Two therapeutic strategies emerge from the two-strike model, each targeting one layer of epigenetic damage.
For Strike One: G9a inhibitors. The preclinical evidence is compelling. UNC0638 restores antiviral gene expression. UNC0642 reverses multi-omic changes in patient PBMCs and reduces lung damage in vivo. The approach is host-directed and variant-proof. A US patent (#63/113,211) protects the indication.
But there's a critical gap: no G9a inhibitor has reached human clinical trials. As of early 2026, several candidates (MS152, DS79932728, MS8511) are in preclinical optimization, but challenges with oral bioavailability and membrane permeability have stalled clinical translation. By contrast, EZH2 inhibitors — targeting a related histone methyltransferase — have three compounds in clinical trials, showing the concept is viable. G9a is simply behind in the pipeline.
For Strike Two: IL-6 receptor antagonists. If IL-6 is the signal that reprograms bone marrow HSPCs, blocking it should prevent new reprogramming — and possibly allow existing marks to decay as HSPCs turn over. Tocilizumab and sarilumab are FDA-approved for other indications and readily available.
The evidence is suggestive. A case report in Frontiers in Immunology (2022) described a Long COVID patient on tocilizumab for rheumatoid arthritis whose symptoms resolved on IL-6R blockade, worsened (with new brain fog) when the drug was stopped, and resolved again when it was restarted. A 2026 emulated target trial in RA patients found that IL-6R antagonists were associated with 58% lower diagnosed Long COVID and 60% lower 12-month mortality compared to other biologics.
But no prospective trial has tested tocilizumab specifically for epigenetic reversal in Long COVID. The theoretical framework is strong — block the signal, prevent the reprogramming — but whether IL-6 blockade can reverse marks that are already established, rather than merely preventing new ones, remains unknown.
A combined approach — G9a inhibitors to restore antiviral defense (enabling viral clearance) plus IL-6R blockade to halt inflammatory reprogramming — would address both strikes simultaneously. But G9a inhibitors aren't available for human use yet, making this a future possibility rather than a current option.
What This Means
The epigenetic perspective reframes Long COVID. It isn't just lingering inflammation, or viral persistence, or immune exhaustion. It's a disease of altered instructions — an immune system running on rewritten code. The virus may be gone from the bloodstream, but its edits remain in the chromatin.
This also explains why Long COVID is so resistant to simple interventions. You can't fix reprogrammed stem cells with an antiviral. You can't resolve silenced antiviral genes with an anti-inflammatory. The two problems require two solutions, each targeting a different epigenetic layer. And neither solution is currently available in approved form for this indication.
The science is ahead of the medicine. The mechanisms are increasingly clear. The therapeutic targets are identified. What's missing is the clinical infrastructure to test them — G9a inhibitors ready for Phase 1, IL-6R blockade trials designed to measure epigenetic endpoints, and biomarkers to identify which patients carry which epigenetic scars.
Until then, the code remains rewritten. But we're learning to read it.
Key Sources
- Anastasakis et al. — PLOS ONE (2024): NSP1 induces G9a-mediated H3K9me2 silencing of antiviral genes; UNC0638 restores expression
- G9a-METTL3-m6A axis — iScience (2025): G9a rewires host m6A epitranscriptome via METTL3; UNC0642 reverses multi-omic effects in patient PBMCs and reduces lung damage in vivo
- Cheong et al. — Cell (2023): IL-6 epigenetically reprograms HSPCs; trained immunity persists 12+ months; skewed myelopoiesis
- Li et al. — Nature Immunology (Jan 2026): LC-Mo transcriptional state in Long COVID monocytes
- Balnis et al.: 71 persistent differentially methylated regions in COVID-19 convalescents
- Zameer et al. — Clinical Epigenetics (2026): DNA methylation alterations linked to neurological dysfunction
- Butzin-Dozier et al. (2026 preprint): IL-6R antagonists associated with 58% lower LC, 60% lower mortality in emulated target trial
- Frontiers in Immunology (2022): Tocilizumab case report — LC symptoms resolved on IL-6R blockade, recurred off drug
- G9a inhibitor pipeline review — J. Med. Chem. (2025): No G9a inhibitors in clinical trials; MS152, DS79932728, MS8511 in preclinical development