Liraglutide exerts analgesic, anti-inflammatory and anti-degradative actions in osteoarthritis (Meurot et al. 2022)

TL;DR

Combined in vivo (MIA mouse OA model) + in vitro (IL-1β-stimulated primary chondrocytes; LPS-stimulated RAW 264.7 macrophages) study showing that intra-articular (IA) liraglutide is analgesic, anti-inflammatory, and anti-catabolic in OA, with all effects reversed by the GLP-1R competitive antagonist exendin-9-39 — establishing GLP-1R as the on-target receptor in joint cells. The first paper to demonstrate GLP-1R protein expression in human OA articular cartilage and synovial membrane. The intra-articular delivery framing is critical — the authors cite the negative Gudbergsen 2021 systemic-liraglutide OA RCT ¹ as evidence that systemic delivery cannot achieve sufficient joint concentrations, motivating the IA pivot. Significant commercial COI: senior authors are CEOs of 4P-Pharma + 4Moving Biotech and hold patents on GLP-1 analogs for OA.

Design

In vivo (MIA mouse OA)

Sodium monoiodoacetate (MIA) model — the standard chemical-induction OA mouse model. MIA inhibits chondrocyte glycolysis, causing chondrocyte death and subsequent OA-like joint degradation with reproducible pain behaviour.

Animals: 12-week-old male wild-type C57Bl/6 mice (Janvier Labs)
Induction: single IA injection of 0.75 mg / 5 µL MIA in 0.9% saline (pH 7.4) into right knee on day 1
Short-term cohort (n=120 total):
- Saline/vehicle (n=19), MIA/vehicle (n=18), MIA/liraglutide groups at 1, 5, 10, 20 µg (n=17 each), MIA/dexamethasone 20 µg (n=15)
- Single IA treatment on day 3
- Pain assessment (von Frey filaments) on days 2 (randomization), 7, and 10
- Sacrifice day 11; histology
Long-term cohort (n=49 total):
- Saline/vehicle (n=9), MIA/vehicle (n=10), MIA/liraglutide 20 µg (n=10), MIA/dexamethasone 20 µg (n=10)
- Three IA injections: days 8, 15, and 22 (once-weekly × 3)
- Pain assessment days 7, 14, 21, 28; sacrifice day 29
Pain readout: paw withdrawal threshold via Semmes-Weinstein von Frey filaments (forces 0.008–1.0 g)
Histology: Krenn synovitis score on H&E sections (blinded assessment by two examiners)

In vitro

Murine primary chondrocytes isolated from articular cartilage of 6-day-old C57Bl/6 pups; chondrocyte phenotype confirmed by Sox9/Col2a1/Acan expression and Alcian blue/Safranin O staining. Used at passage 0.
RAW 264.7 murine macrophage cell line (ATCC TIB-71); used before passage 20.
Stimuli: IL-1β 2 ng/mL (chondrocytes) or LPS 100 ng/mL (RAW 264.7); 24 h
Liraglutide doses: 10-point dose response, 6.6 nM to 3.4 µM (24 h co-treatment); long-term anti-catabolic experiments at 50 nM × 72 h
Antagonist (mechanism): exendin-9-39 at 25, 50, 100 nM; co-treatment with 50 nM liraglutide
Readouts:
- Cytokines/mediators by ELISA: NO (Griess), PGE₂, IL-6, MMP-3, MMP-13
- mRNA by RT-qPCR: iNos, Cox2, Tnf-α (chondrocytes); Il-6, Cox2, Tnf-α, Mcp-1, Cd38, Erg-2 (macrophages); Mmp-3, Mmp-13, Adamts4, Adamts5 (chondrocyte catabolism)
- GAG release into supernatant (cartilage matrix degradation surrogate)

Human samples (limited scope — IHC only)

6 human OA patients undergoing arthroplasty at Saint-Antoine Hospital, Paris (3 women, 3 men; ages 68–83 women, 71–83 men; Mankin score 1.5–14/14). 10 mm² cartilage + subchondral bone plugs and synovial membrane samples; used solely for GLP-1R immunohistochemistry, not for any functional/intervention assays.

Key results

1. GLP-1R is expressed in articular cartilage and synovial membrane (human + mouse)

IHC of human OA knee sections from 6 patients: chondrocytes express GLP-1R protein in superficial and intermediate cartilage layers (Fig 1a). Synovial membrane shows positive GLP-1R in intima and blood vessel walls (Fig 1b). Non-OA mouse knee sections similarly show GLP-1R+ chondrocytes (tibial + femoral articular), meniscus, and bone marrow (Fig 1c–d). RT-qPCR on RNA from primary human chondrocytes (cartilage explants from 18 OA patients) confirms GLP-1R mRNA expression (data not shown — author claim).

Significance: first protein-level demonstration of GLP-1R expression in OA-affected joint tissues in humans. Establishes the molecular target for joint-directed GLP-1 RA therapy.

2. Single IA liraglutide injection is analgesic in short-term MIA mouse OA

Day 3 IA injection of liraglutide produced dose-dependent rescue of MIA-induced mechanical allodynia by day 7 and day 10 (Fig 2a):

Dose	p value vs MIA/vehicle (day 7)
1 µg	0.0196 (day 10)
5 µg	<0.0001
10 µg	0.0011
20 µg	<0.0001 (day 7); <0.0001 (day 10); superior to dexamethasone 20 µg, p<0.0001

Calculated EC50 = 11 µg (Fig 2b). 20 µg liraglutide outperformed the 20 µg dexamethasone positive control on day 10.

3. Repeated IA liraglutide is analgesic over 28 days, no body weight effect

Long-term cohort (weekly 20 µg IA × 3 injections, days 8/15/22):

Significant improvement vs MIA/vehicle: liraglutide day 14 p=0.0098, day 28 p=0.0038; dexamethasone day 14 p=0.0568 (NS), day 28 p=0.0002
After second and third injections (days 21, 28), liraglutide and dexamethasone effects were comparable
No significant body weight differences between treatment groups in either short-term or long-term studies (Suppl. Fig 1) — pain and synovitis effects dissociate from systemic metabolic actions of GLP-1R agonism

4. Liraglutide reduces synovitis (anti-inflammatory in vivo)

H&E + Krenn synovitis scoring at day 11 (Fig 3a–b):

Saline/vehicle vs MIA/vehicle: p<0.0001 (model induced synovitis)
MIA/liraglutide 20 µg vs MIA/vehicle: p=0.0099 (significant reduction)
MIA/dexamethasone 20 µg vs MIA/vehicle: p=0.1288 (NOT significant — liraglutide outperformed dex on this endpoint)

Synovitis-pain coupling (Fig 3c): linear regression of synovitis score vs day-10 von Frey paw withdrawal threshold yielded R² = 0.91, p<0.0001 — a remarkably tight observational coupling between joint inflammation and pain across animals.

5. Anti-inflammatory in vitro — chondrocytes (IL-1β-stimulated)

10-dose liraglutide × IL-1β-stimulated primary chondrocytes (Fig 4a):

Mediator	IC50 (nM)	95% CI
NO (Griess)	45	41–50
PGE₂ (ELISA)	48	43–53
IL-6 (ELISA)	38	34–44

mRNA (Fig 4b; doses 13.3, 53.1, 1700 nM): iNos, Cox2, and Tnf-α dose-dependently reduced (p=0.0286 vs IL-1β alone). LDH non-cytotoxic confirmed.

6. Anti-inflammatory in vitro — macrophages (LPS-stimulated RAW 264.7)

10-dose liraglutide × LPS-stimulated RAW 264.7 (Fig 5a):

Mediator	IC50 (nM)	95% CI
NO	38	34–44
PGE₂	54	49–59
IL-6	41	37–45

mRNA: Il-6, Cox2, Tnf-α dose-dependently reduced (p=0.0286).

7. Liraglutide drives M1→M2 macrophage repolarization

Fig 5c: in LPS-stimulated RAW 264.7 cells co-treated with liraglutide:

M1-associated genes ↓: Mcp-1, Cd38
M2-associated gene ↑: Erg-2

Establishes liraglutide’s effect on macrophage phenotype is not just cytokine suppression but a polarization shift toward the anti-inflammatory M2 state — the same state correlated with reduced cartilage damage in mouse synovial macrophage-depletion experiments cited in the Discussion.

8. GLP-1R is the on-target receptor (exendin-9-39 reversal)

Fig 6: 100 nM exendin-9-39 (GLP-1R competitive antagonist) completely reversed the anti-inflammatory effect of 50 nM liraglutide in both chondrocytes and macrophages (NO, PGE₂, IL-6 all restored to LPS/IL-1β-only levels). Dose-dependent reversal at 25, 50, 100 nM exendin-9-39. p=0.0022 (chondrocytes), p=0.026 / p=0.0022 (macrophages).

This is the cleanest mechanism statement in the paper — it confirms the effect is mediated by GLP-1R, not an off-target action of liraglutide.

9. Anti-catabolic on chondrocytes

Fig 7a: 10-dose liraglutide × IL-1β-stimulated primary chondrocytes → cartilage degradation enzymes:

Enzyme/marker	IC50 (nM)	95% CI
MMP-3 (secreted)	56	52–62
MMP-13 (secreted)	58	53–64
GAG release (rGAG)	45	31–67

mRNA reductions in Mmp-3, Mmp-13, Adamts4, Adamts5 (p=0.0286).

Long-term anti-degradative experiment (Fig 7c): 50 nM liraglutide × 72 h × 2 ng/mL IL-1β rescued IL-1β-induced GAG release from chondrocyte ECM (p=0.0227).

Negative human counterpoint cited in this paper

The Discussion explicitly addresses Gudbergsen et al. 2021 (Am J Clin Nutr 113:314–323; doi:10.1093/ajcn/nqaa328) ¹, which reported that daily systemic liraglutide injections in OA-with-obesity patients (after diet-induced weight loss) did NOT ameliorate OA-related pain. The Meurot authors interpret this as a failure of delivery — “probably because of poor access and hence poor local concentrations of liraglutide in the knee joint” — and pivot the therapeutic framing to intra-articular delivery. This is the central translational caveat of the paper: a positive IA mouse + in-vitro story exists in tension with a negative systemic-delivery human RCT. Resolution requires either an IA human trial or a systemic agent that achieves higher joint concentrations (the position taken by Qin et al. 2026 for semaglutide ², which reports positive systemic-delivery effects in mice).

Mechanism summary as stated by authors

Targeting inflammation and cartilage breakdown in two main cellular actors of the diseased joint, and alleviating pain in vivo, confers to liraglutide the properties of a potential disease modifier, which could constitute a new treatment for OA.

The proposed pathway: liraglutide → joint chondrocyte/synovial macrophage GLP-1R → reduced NF-κB / PI3K-Akt-driven inflammatory transcription → ↓ NO, PGE₂, IL-6, MMPs, ADAMTS → reduced synovitis (in vivo) → reduced pain (DRG nociceptor desensitization). The authors note that prior work (Chen et al. 2018 ref 35, Que et al. 2019 ref 37) implicates ER-stress reduction and PKA/CREB-mediated effects but emphasize that the analgesic mechanism of liraglutide had not been previously reported — this paper is the first to establish IA liraglutide analgesia in MIA OA.

Strengths

Integrated in-vivo + in-vitro design with cell-type dissection (chondrocyte vs macrophage)
Genetic-pharmacology mechanism confirmation via exendin-9-39 reversal
Dose-response IC50/EC50 quantification
Cross-species GLP-1R expression confirmation (human OA + mouse)
Synovitis-pain correlation (R²=0.91) is an elegant single-figure summary of the inflammation-pain coupling
Open access (CC-BY 4.0), full local PDF available

Limitations

Significant commercial COI — senior authors are CEOs of 4P-Pharma and 4Moving Biotech; hold patents on GLP-1 analogs for OA treatment (PCT/FR2013/051998, PCT/IB2018/059100). Independent replication is essential.
MIA model limitations — chemical chondrocyte-toxicity model, not slowly progressive degenerative OA; the authors note the need for follow-up in DMM (destabilization of medial meniscus) for structural endpoints. No DMM data in this paper.
In vitro and short-duration in vivo only — mechanism work is 24 h in cells; in vivo is up to 28 days. No long-term cartilage structural data.
Human data is IHC only — no functional or interventional human work in this paper. n=6 patients for cartilage IHC, n=18 for chondrocyte explant RT-qPCR; both small.
Anti-degradative claim relies on small-n in vitro work (“small number of experiments performed”; authors flag this as a limitation; require DMM in vivo to validate).
No body weight effect in mice is technically a strength for the local-action argument but means the paper doesn’t address whether systemic GLP-1 RA + weight loss adds incremental benefit.
Macrophage M1/M2 framing is acknowledged by authors as a simplification of in-vivo polarization plasticity.

Extrapolation to humans

Dimension	Status	Notes
Pathway conserved in humans?	yes	GLP-1R protein expression confirmed in human OA cartilage + synovium (this paper)
Phenotype conserved in humans?	partial	Pain/synovitis pathophysiology conserved; cartilage degradation enzymes (MMP-3/13, ADAMTS4/5) conserved
Replicated in humans?	no — counter-evidence exists	Systemic liraglutide failed in Gudbergsen 2021 ¹; no IA-liraglutide RCT exists. The closest contemporary human signal is the Qin 2026 ChiCTR2200066291 pilot for semaglutide (different agent, systemic dose) ²

Cross-references

liraglutide — compound page (stub as of 2026-05-08; should be expanded with this paper’s findings as the “joint / OA” section anchor)
glp1-agonists — class page; should add a “Joint and cartilage” section consolidating this study + Qin 2026
osteoarthritis — phenotype page (newly seeded 2026-05-08)
chronic-inflammation — local synovitis + macrophage M1/M2 axis is the hallmark intersection
glp1r — receptor (implicit stub); first protein-level expression confirmation in OA joint tissues here
Companion / counterpoint study: qin-2026-semaglutide-oa (semaglutide; weight-loss-independent; systemic delivery)
Negative human RCT for systemic liraglutide: ¹

Footnotes

doi:10.1093/ajcn/nqaa328 · rct · n=156 · Gudbergsen H et al. · Am J Clin Nutr 2021;113:314–323 · “Liraglutide after diet-induced weight loss for pain and weight control in knee osteoarthritis” · cited as reference 44 in ³ · daily systemic liraglutide did NOT ameliorate OA pain after diet-induced weight loss · primary source not yet retrieved; n and effect-size figures are derived from Meurot 2022 narrative — verify against original PDF before relying on quantitative claims ↩ ↩² ↩³ ↩⁴
qin-2026-semaglutide-oa · companion semaglutide-OA paper; systemic-delivery, weight-loss-independent · doi:10.1016/j.cmet.2026.01.008 ↩ ↩²
this study; doi:10.1038/s41598-022-05323-7 · Meurot et al. Sci Rep 2022;12:1567 ↩

🧬 Aging Wiki

Explorer

Liraglutide, a glucagon-like peptide 1 receptor agonist, exerts analgesic, anti-inflammatory and anti-degradative actions in osteoarthritis