Reviewed 17 May 2026

Example research study

Fact-Checking a KAT2A Preprint on Epigenetic Licensing in Rheumatoid Arthritis Synovial Macrophages

A preprint proposes KAT2A-mediated H3K9 acetylation as the "commitment checkpoint" between RA flare resolution and chronic joint destruction. BioSkepsis verified every mechanistic claim against PubMed — confirming the core epigenetic and metabolic links while exposing selective citation, unsupported CD40 claims, and critical omissions about MerTK reversal and PRG4-dependent barrier regulation.

TL;DR The preprint's core chain — KAT2A licenses Il1b/Nlrp3 transcription via H3K9ac, drives aerobic glycolysis, and suppresses NRF2 — is well-established (PMID: 37313329). The CX3CR1+ TREM2+ synovial barrier maintained by JAM1/ZO-1/Claudin-5 is confirmed (PMID: 31391580). However, the preprint omits evidence that MerTK function reverses in late-stage RA (PMID: 38493215), that PRG4 regulates the metabolic switch upstream of KAT2A (PMID: 39372933), and that barrier reversibility has already been demonstrated via mitochondrial transfer and xanthine oxidase inhibition. The CD40-high phenotype attribution is not reported in any cited paper. BioSkepsis flagged three unverified citations where the paper cited did not support the specific claim made.

What the Preprint Claims About KAT2A and Synovial Macrophage Licensing

The preprint (Proposal 1.1) describes a mechanistic chain in which CX3CR1+ TREM2+ lining macrophages maintain a physical joint barrier using tight junction proteins (JAM1, ZO-1, Claudin-5). Upon inflammatory challenge, KAT2A (GCN5) catalyses H3K9 acetylation at the Il1b and Nlrp3 promoters, simultaneously licensing a "CD40-high" pathogenic phenotype and driving a metabolic switch to aerobic glycolysis. This switch suppresses the NRF2 antioxidant pathway, which would otherwise repress pro-inflammatory gene transcription. The preprint frames this as a "commitment checkpoint" whose reversibility "remains untested."

BioSkepsis ran a multi-step fact-check across eight analytical queries — verifying each mechanistic claim, checking cited effect sizes, identifying contradictory evidence, assessing the selectivity of references, evaluating experimental methods, and testing the novelty claim. The results paint a more nuanced picture than the preprint presents.

Confirmed: KAT2A-Mediated H3K9ac and NLRP3 Inflammasome Priming in RA Macrophages

The preprint's central epigenetic claim is directly supported. Zhang et al. (2023) demonstrated that KAT2A expression is elevated in PBMCs of active RA patients (p < 0.0001 versus healthy controls) and correlates with clinical arthritis scores in CIA mice (PMID: 37313329). Pharmacological inhibition with MB-3 reduces H3K9ac enrichment at Il1b and Nlrp3 promoters (p < 0.001) and suppresses both priming and activation stages of the NLRP3 inflammasome.

The metabolic coupling is equally robust. KAT2A inhibition decreases the extracellular acidification rate (ECAR), intracellular ATP, and lactic acid production in LPS-stimulated macrophages, confirming that KAT2A is required for the glycolytic reprogramming that sustains inflammasome activation (PMID: 37313329). The NRF2 axis is validated in the same study: KAT2A suppresses NRF2 protein expression, and its inhibition restores NRF2 enrichment at pro-inflammatory promoters. This regulation depends on the metabolite itaconate — in IRG1-deficient macrophages lacking itaconate, KAT2A inhibitors lose their ability to upregulate NRF2 unless exogenous itaconate derivatives are supplied.

Confirmed: The CX3CR1+ TREM2+ Synovial Barrier and Its Tight Junction Architecture

The physical synovial barrier is a landmark finding from Culemann et al. (2019), published in Nature. Using 3D light-sheet fluorescence microscopy and single-cell RNA sequencing, the study identified CX3CR1+ TREM2+ resident macrophages forming a dense intimal lining maintained by JAM1 (F11r), ZO-1 (Tjp1), and Claudin-5 (Cldn5) (PMID: 31391580). These macrophages constitute approximately 40% of total synovial macrophages under steady-state conditions and locally renew from CX3CR1+ mononuclear cell pools embedded in the synovial tissue.

Local depletion of CX3CR1+ macrophages in Cx3cr1-cre R26-iDTR mice accelerates arthritis onset and neutrophil influx, confirming their protective function. The barrier's disintegration — loss of tight junctions and reorganisation of the macrophage layer — is a consistent feature of inflammatory arthritis onset across multiple models (PMID: 31391580, PMID: 39372933).

What the Preprint Gets Wrong or Overstates About Macrophage Phenotyping

The preprint attributes a "CD40-high pathogenic phenotype" to the KAT2A-licensed state. BioSkepsis found no mention of CD40 in any of the cited papers — not in PMID: 37313329 (KAT2A/H3K9ac), not in PMID: 31391580 (synovial barrier), and not in any of the other references the preprint lists. The CD40 attribution is unsupported by the provided evidence (Not Reported).

The preprint also lists three pharmacological agents for testing reversibility — givinostat (HDAC inhibitor), I-BET151 and JQ1 (BET inhibitors) — but none of these appear in the cited literature in the context of RA or KAT2A-mediated licensing. Only MB-3 is validated as a KAT2A inhibitor in the relevant studies (PMID: 37313329, PMID: 41476772).

Selective Citation: What the Preprint Should Have Cited in Synovial Macrophage Biology

BioSkepsis identified four categories of critical omissions in the preprint's 14-reference bibliography.

Omitted evidence that qualifies or contradicts the preprint's claims

Omission	What the literature shows	PMID
MerTK stage-dependent reversal	In late-stage D2T RA (R4RA cohort), MERTK expression positively correlates with inflammatory infiltration — opposite to its "pro-resolving" role in early disease	38493215
PRG4/Lubricin metabolic regulation	PRG4 loss precedes and licenses the macrophage glycolytic switch via xanthine oxidase and HIF-1α; febuxostat restores the CX3CR1+ TREM2+ barrier	39372933, 39696446
Barrier reversibility via mitochondrial transfer	ADSCs repair the barrier by transferring mitochondria to Atf3-high macrophages, restoring oxidative phosphorylation	40038808
KAT2A ferritinophagy pathway	KAT2A promotes NCOA4-dependent ferritinophagy and cGAS-STING activation — alternative pro-inflammatory routes independent of NRF2	41476772

The omission of the R4RA cohort data (PMID: 38493215) is particularly significant. The preprint frames MerTK+ macrophages as definitively "pro-resolving," but clinical data from 133 difficult-to-treat RA patients shows that MERTK expression is upregulated in lympho-myeloid pathotypes and positively correlates with pro-inflammatory cytokines in established disease. This reversal is consistent with ADAM17-mediated proteolytic shedding of MerTK, which produces a soluble decoy receptor (sMer) that actively prevents anti-inflammatory signalling (PMID: 28067670, PMID: 38673989).

The Novelty Claim: Has Barrier Reversibility Already Been Demonstrated in RA Models?

The preprint's framing — that the reversibility of the KAT2A commitment checkpoint "remains untested" — is contradicted by two independent studies in the existing literature.

Wang et al. (2025) demonstrated that intra-articular injection of ADSCs restores the CX3CR1+ macrophage barrier in serum-transfer arthritis mice. The mechanism involves mitochondrial transfer via tunnelling nanotubes to an Atf3+ Ccl3+ subset of metabolically impaired macrophages, depleting the pro-inflammatory population and restoring oxidative phosphorylation (PMID: 40038808). Separately, Elsaid et al. (2024) showed that the xanthine oxidase inhibitor febuxostat preserves CX3CR1+ TREM2+ macrophages, reduces glycolytic flux and HIF-1α levels, and maintains Claudin-5 expression in Prg4-inactivated mice (PMID: 39372933, PMID: 39696446).

The individual components of the KAT2A axis (H3K9ac at Il1b/Nlrp3, glycolytic switch, NRF2 suppression) were established in 2023 (PMID: 37313329). The barrier identity (JAM1, ZO-1, Claudin-5) was established in 2019 (PMID: 31391580). The preprint's integration of these systems as a "commitment checkpoint" is a reasonable synthesis, but it is an extension of established paradigms — not a first-of-its-kind discovery.

BioSkepsis Evidence Grading: How Each Claim Scored on the Verification Pipeline

Evidence status of each mechanistic claim in the preprint

Claim	Evidence Grade	Key PMIDs
KAT2A licenses Il1b/Nlrp3 via H3K9ac	Direct, High	37313329
KAT2A drives aerobic glycolysis	Direct, High	37313329
KAT2A suppresses NRF2 via itaconate	Direct, High	37313329
CX3CR1+ TREM2+ barrier with JAM1/ZO-1/Claudin-5	Direct, High	31391580, 39372933
KAT2A glycolysis causes barrier disintegration	Derived, Medium	37313329, 40038808
CD40-high pathogenic phenotype	Not Reported	—
Givinostat/BET inhibitor reversibility	Not Reported	—
MerTK is purely pro-resolving	Contradicted in late RA	38493215

Experimental Method Limitations: Seahorse, ChIP-seq, and Lineage-Specific Deletion in RA Models

The preprint proposes Seahorse XF metabolic flux analysis, ChIP-seq/ATAC-seq, and Cx3cr1-Cre lineage-specific deletion as its core methods. BioSkepsis identified specific pitfalls in the literature for each.

Seahorse assays measure averaged population responses and cannot distinguish metabolic shifts in resident macrophages from those in infiltrating monocyte-derived cells without prior sorting (PMID: 39696446). For ChIP-seq, ATAC-qPCR amplicons can extend into regions of closed chromatin, leading to impaired amplification and underestimation of accessibility at specific promoters (PMID: 37615937). The Cx3cr1-creER system requires a 4-week rest period after tamoxifen pulsing to avoid labelling circulating monocytes; without this, the preprint's claim of STRM-specific effects may be confounded (PMID: 31391580). Conditional knockout models achieve approximately 94% — not complete — reduction in target gene expression (PMID: 39372933).

Who Should Use Citation-Grounded Preprint Fact-Checking

Frequently asked questions

What is KAT2A and why does it matter in rheumatoid arthritis?

KAT2A (also known as GCN5) is a lysine acetyltransferase that catalyses H3K9 acetylation at the promoters of pro-inflammatory genes Il1b and Nlrp3 in macrophages. Its expression is elevated in RA patients and CIA mice, and pharmacological inhibition with MB-3 reduces joint inflammation and bone destruction (PMID: 37313329).

What is the CX3CR1+ TREM2+ synovial barrier?

A population of tissue-resident macrophages in the synovial intimal lining that forms a physical and immunological barrier, maintained by tight junction proteins JAM1, ZO-1, and Claudin-5. First described by Culemann et al. in Nature (PMID: 31391580), its disintegration is a hallmark of inflammatory arthritis onset.

How does KAT2A connect to aerobic glycolysis in macrophages?

KAT2A is required for the metabolic reprogramming from oxidative phosphorylation to aerobic glycolysis. Inhibition reduces ECAR, intracellular ATP, and lactic acid production, demonstrating a direct link between epigenetic licensing and metabolic switching (PMID: 37313329).

Does KAT2A inhibition restore the NRF2 antioxidant pathway?

Yes. KAT2A suppresses NRF2 protein expression and its enrichment at Il1b and Nlrp3 promoters. Inhibition with MB-3 increases NRF2 levels and recruits NRF2 as a transcriptional repressor. This effect depends on the metabolite itaconate via the IRG1 pathway (PMID: 37313329).

What does BioSkepsis do differently from a general-purpose LLM when fact-checking a preprint?

BioSkepsis verifies every citation against PubMed with a three-stage check: entity matching, conclusion alignment, and evidence grading. It flags unverified citations and distinguishes between directly supported claims, derived inferences, and unsupported assertions — something a general-purpose LLM cannot do because it lacks grounded citation verification.

Has the reversibility of the synovial barrier already been demonstrated?

Yes. ADSCs repair CX3CR1+ macrophages by transferring functional mitochondria (PMID: 40038808), and the XO inhibitor febuxostat preserves CX3CR1+ TREM2+ macrophages and Claudin-5 expression in Prg4-knockout mice (PMID: 39372933, PMID: 39696446). The preprint's claim that reversibility "remains untested" is contradicted.

What critical evidence did the preprint selectively omit?

The preprint omits R4RA cohort data showing MerTK positively correlates with inflammation in late-stage RA (PMID: 38493215), the role of PRG4/Lubricin in regulating the metabolic switch upstream of KAT2A (PMID: 39372933), and alternative KAT2A pathogenic mechanisms via ferritinophagy and cGAS-STING (PMID: 41476772).

Fact-check your next biomedical preprint with PubMed-grounded citations

BioSkepsis verifies every claim against the primary literature — flagging unsupported attributions, selective citations, and evidence gaps that general-purpose LLMs miss.

Start free

Sources & further reading

1. Zhang Y et al. Targeting KAT2A inhibits inflammatory macrophage activation and rheumatoid arthritis through epigenetic and metabolic reprogramming. MedComm. 2023;4(3):e306. PMID: 37313329
2. Culemann S et al. Locally renewing resident synovial macrophages provide a protective barrier for the joint. Nature. 2019;572(7771):670–675. PMID: 31391580
3. Elsaid KA et al. Proteoglycan 4 (Lubricin) and regulation of xanthine oxidase in synovial macrophage as a mechanism of controlling synovitis. Res Sq. 2024 (preprint). PMID: 39372933
4. Elsaid KA et al. Proteoglycan 4 (Lubricin) and regulation of xanthine oxidase in synovial macrophage as a mechanism of controlling synovitis. Arthritis Res Ther. 2024;26(1):214. PMID: 39696446
5. Wang L et al. Adipose-derived stem cells attenuate rheumatoid arthritis by restoring CX3CR1+ synovial lining macrophage barrier. Stem Cell Res Ther. 2025;16(1):111. PMID: 40038808
6. Nerviani A et al. Axl and MerTK regulate synovial inflammation and are modulated by IL-6 inhibition in rheumatoid arthritis. Nat Commun. 2024;15(1):2398. PMID: 38493215
7. Zuo Z et al. Inhibitors of KAT2A alleviate the progression of AKI by alleviating macrophage ferritinophagy. ImmunoTargets Ther. 2025;14:1515–1529. PMID: 41476772
8. Cai B et al. MerTK receptor cleavage promotes plaque necrosis and defective resolution in atherosclerosis. J Clin Invest. 2017;127(2):564–568. PMID: 28067670
9. Lahey KC et al. Regulation of Mertk surface expression via ADAM17 and γ-secretase proteolytic processing. Int J Mol Sci. 2024;25(8):4404. PMID: 38673989
10. Zhou L et al. Hepatic danger signaling triggers TREM2+ macrophage induction and drives steatohepatitis via MS4A7-dependent inflammasome activation. Sci Transl Med. 2024;16(738):eadk1866. PMID: 38478630
11. Shi S et al. TREM2 in MASH: integrating lipid metabolism and immune response. Front Immunol. 2025;16:1604837. PMID: 40636122
12. Chakraborty S et al. Trained immunity of alveolar macrophages enhances injury resolution via KLF4-MERTK-mediated efferocytosis. J Exp Med. 2023;220(11). PMID: 37615937