Koo & Morrow 2025 — HERV amplification in PASC monocytes

One-paragraph summary

Single-cell-RNA-seq study of monocytes from patients with post-acute sequelae of COVID-19 (PASC, n=12). Using a Window-based HERV Alignment (WHA) method that scores per-locus HERV transcript depth, the authors identify three autonomous HERV loci expressed across all 12 PASC patients, with the most amplified locus residing within an intron of JAKMIP2 — a host gene whose own transcription, along with that of neighboring genes, was also elevated. Comparable HERV expression patterns are seen in monocytes from patients with influenza, dengue, and sepsis but not in 31 healthy controls. The interpretation: viral infections (and sepsis) induce innate immune memory in monocytes via epigenetic remodeling of hematopoietic stem and myeloid precursor cells, and that remodeling persists into the post-acute phase, producing co-amplified HERV-host transcript patterns. The mechanism is consistent with a durable upstream driver of post-acute symptoms — the same kind of durability the project's HERV-mitochondrial-inflammation loop is designed to explain. For Q40 (the bioinformatic pipeline), this paper provides empirical evidence that HERV reactivation in PASC extends beyond HERV-W ENV — the three persistently-expressed loci in PASC monocytes are candidates for inclusion in the Q40 protein universe and may broaden the cgas/STING-binding screen beyond syncytin-1.

Claims as triples

• viral_infection — causes → HERV_reactivation [evidence: distinct HERV expression patterns in COVID, influenza, dengue, sepsis monocytes; confidence: established (reinforces existing edge)]
• HERV_reactivation — present_in → post_covid_syndrome [evidence: 3 HERV loci expressed in all 12 PASC patients; confidence: emerging]
• HERV_reactivation — modulates → microglia [evidence: HERV-host co-amplification in monocyte lineage suggests innate-immune memory remodeling, with myeloid-microglial overlap; confidence: bridging]
• HERV_reactivation — bridges → fm_autoimmune [evidence: cross-condition extrapolation from PASC to FM-HERV+ subset; confidence: bridging]

Methods note

scRNA-seq data analyzed via custom Window-based HERV Alignment (WHA) method (≥9 windows = positive call; ≤8 = negative). Control set: 31 normal individuals. Comparison sets: hospitalized COVID-19, influenza, dengue, sepsis (each as published scRNA-seq cohorts). PASC sample: 12 patients sampled at early (<7 days post recovery) and late (>14 days) timepoints. Azimuth application used for monocyte-lineage identification within PBMCs. No interventional arm.

Limitations

• Small PASC cohort (n=12). Replication in larger PASC cohorts and FM cohorts is needed before the three persistently-expressed loci can be promoted to confirmed FM-relevant targets.
• Monocyte-only readout. Whether the same HERV loci are amplified in T cells, B cells, or tissue-resident macrophages in PASC patients (or FM patients) is open.
• Window-based alignment method is custom. WHA performance benchmarking against orthogonal methods (e.g., Telescope, hervquant) would strengthen the locus-level claims.
• No protein-level data. The paper shows HERV transcripts; whether the three persistent loci produce functional protein products with cgas-STING-binding capability is the Q40 question.
• No FM patients in the cohort. Direct demonstration in FM-HERV+ subset is open.

Open questions raised

• Are the same three persistent HERV loci amplified in FM-HERV-W-positive patients, or are FM and PASC HERV signatures distinct? Tractable with stored FM PBMC scRNA-seq + WHA method.
• Does the JAKMIP2-intronic locus produce a translated protein, and if so, does it have predicted mitochondrial-localization signal? Direct Q40 candidate — should be added to the seed protein universe in q40_pipeline/herv_seed_proteins.tsv.
• Does the persistent HERV-host co-amplification predict response to temelimab or to broad anti-retroviral therapy? Sankar et al rituximab work in HERV-altered ME suggests broader B/plasma-cell-directed therapy may be relevant.

Triangulation notes

• Q40 pipeline input expansion. This paper provides three concrete HERV loci to add to the Q40 seed protein universe beyond the curated 10. The JAKMIP2-intronic locus in particular is a high-priority addition because its host-gene-overlap pattern suggests transcriptional read-through and possible chimeric protein products.
• Compatible with the HERV-mito-loop framework. The loop's defining feature (durability via type-I IFN-driven HERV de-silencing) predicts persistent HERV expression in chronic post-viral states, which is what Koo & Morrow observe in PASC monocytes 14+ days post recovery.
• Strengthens B7 bridge (HERV ↔ FM autoimmune via reactivated HERV-W ENV) — even if the specific persistent loci differ between PASC and FM, the phenomenon of durable HERV reactivation in post-viral chronic disease is now empirically anchored across multiple acute infections.
• Reinforces Martín-Martínez 2025's HERV-W-negative-but-HERV-altered finding. Both papers argue the same point: HERV stratification in chronic post-viral patients must extend beyond HERV-W ENV alone.

Bridges

• Q40 candidate input. Three persistent HERV loci (including the JAKMIP2-intronic locus) added to the seed-protein universe as additional Step 1 inputs.
• B7 reinforcement — HERV ↔ FM autoimmune via reactivated HERV-encoded proteins. Multi-locus rather than single-locus.