⚠️ Auto-extracted by Claude on 2026-05-18 — full PDF not available in a local paper archive (not_oa; returns not_oa). Numerics drawn from secondary citations; Kong et al.’s reported values are flagged where exact numbers require primary-PDF confirmation. Verify all quantitative claims against an institutional-access copy before relying on them. PMC version does not exist (NCBI ID converter returns “Identifier not found in PMC”). no-fulltext-access

Berberine is a novel cholesterol-lowering drug working through a unique mechanism distinct from statins

Kong W-J, Wei J, Abidi P et al. (Jiang J-D, senior author) · Nat Med 10(12):1344–1351 · 2004 · DOI: 10.1038/nm1135 · PMID: 15531889

Citation impact (as of 2026-05-18): 1,355 citations · FWCI = 18.30 · citation_percentile = 100

TL;DR

This is the origin paper for the “berberine works distinctly from statins” framing. Kong et al. report that berberine lowers serum cholesterol and LDL-C in hyperlipidemic hamsters and a small human cohort, and that the mechanism operates through post-transcriptional stabilization of ldlr mRNA — a 3’UTR-dependent mechanism requiring ERK signaling — rather than through SREBP-2 activation or cholesterol-biosynthesis suppression, the canonical statin target at hmgcr. The paper is widely cited in subsequent berberine reviews as the mechanistic foundation for AMPK-independent, cholesterol-lowering effects.

---

Study design

The paper combines three experimental levels conducted in parallel rather than sequentially:

Level	System	Key readout
In vitro	HepG2 human hepatocyte cells	LDLR mRNA half-life, LDLR protein expression, SREBP-2 activation, ERK pathway involvement
In vivo	Hyperlipidemic hamster model (specific strain/induction protocol: verify in primary PDF)	Serum TC, LDL-C, TG at 100 mg/kg/d × 10 d
Human pilot	n=32 hypercholesterolemic adults; 500 mg BID × 3 months	Serum TC, LDL-C, TG, HDL-C

The primary mechanistic arm is in vitro. Human and hamster arms establish translational plausibility but are not powered for definitive clinical conclusions.

---

Key findings

1. LDLR mRNA stabilization — in vitro

Kong et al. reported that berberine treatment of HepG2 cells increased LDLR protein expression by a mechanism operating at the post-transcriptional level. Specifically:

• LDLR mRNA half-life was extended approximately 3-fold in berberine-treated HepG2 cells relative to vehicle controls, via a mechanism dependent on the LDLR 3’ untranslated region (3’UTR). no-fulltext-access — exact half-life values (baseline and treated) require primary PDF confirmation.
• The effect was demonstrated to be independent of SREBP-2 activation: berberine did not induce nuclear translocation of SREBP-2 or activate the sterol-response element (SRE) promoter, confirming the mechanism is upstream and independent of the cholesterol-sensing pathway that statins operate through (by inhibiting hmgcr → cholesterol depletion → SREBP-2 cleavage → LDLR transcription). This dissociation is the paper’s central mechanistic claim.
• ERK pathway involvement: pharmacological inhibition of ERK1/2 blocked berberine-induced LDLR mRNA stabilization, implicating MAPK/ERK signaling as a required intermediate. The precise ERK substrate acting on the LDLR 3’UTR was not identified in this paper. no-mechanism — the bridging molecular target between ERK activation and 3’UTR-mediated mRNA stabilization was not resolved here.

Dimension	Status
Pathway conserved in humans?	partial — LDLR 3’UTR AU-rich element (ARE) regulation exists in human cells; HepG2 is a human-derived line, supporting relevance; primary hepatocyte confirmation pending
Phenotype conserved in humans?	yes (human pilot arm; see below)
Replicated in humans?	partial — small open-label n=32 cohort; RCT replication followed in subsequent decade

2. In vivo efficacy — hyperlipidemic hamster

Kong et al. reported the following lipid changes in hyperlipidemic hamsters treated with berberine 100 mg/kg/day × 10 days (values drawn from secondary citations; verify exact figures against primary PDF): no-fulltext-access

Lipid parameter	Reported change
Serum total cholesterol (TC)	↓ ~26%
LDL-C	↓ ~28%
Triglycerides (TG)	↓ ~35%
HDL-C	not prominently reported as changing

These magnitudes are widely reproduced in reviews citing Kong 2004. Exact p-values and group sizes require primary PDF.

3. Human pilot cohort (n=32)

The human arm enrolled 32 hypercholesterolemic adults treated with berberine 500 mg twice daily for 3 months. Kong et al. reported (values from secondary citations; verify in primary): no-fulltext-access

Lipid parameter	Reported change
LDL-C	↓ ~21% (range 12–32% across individuals)
Serum total cholesterol	↓ ~29%
Triglycerides	↓ ~35%

This arm was uncontrolled (open-label, no placebo arm explicitly described in secondary citations). The n=32 size is insufficient to draw definitive conclusions. It established proof-of-concept for human lipid-lowering activity and seeded a substantial subsequent clinical literature. needs-replication — subsequent RCTs have replicated the lipid-lowering direction but this pilot itself is single-arm.

---

Significance for berberine biology

This paper established three canonical claims in the berberine literature that have been repeatedly cited since 2004:

1. LDLR mRNA stabilization is a primary mechanism of berberine’s cholesterol-lowering effect — distinct from any effect on cholesterol biosynthesis or SREBP-2-mediated transcription.
2. ERK1/2 signaling is required for this stabilization — linking berberine’s metabolic effects to a MAPK axis (this is mechanistically distinct from, and potentially complementary to, berberine’s AMPK-activating effects described in subsequent work).
3. Berberine’s mechanism is additive-in-principle with statins — because the mechanistic targets do not overlap (LDLR mRNA stability via ERK vs. HMGCR inhibition via statin), the two agents could theoretically be combined. This framing has driven subsequent combination-therapy investigation. needs-replication — formal combination RCTs at scale remain limited.

The paper predates the main AMPK-activation literature for berberine (Yin et al. 2008 and related). It therefore represents the cholesterol-mechanism arm of berberine’s biology, not the nutrient-sensing / longevity arm.

---

Limitations

• Human arm is open-label, single-arm, n=32 — no placebo control, no randomization described in available secondary accounts. Cannot rule out regression-to-mean or dietary confounding.
• In vitro model — HepG2 cells are a hepatocellular carcinoma line with known metabolic differences from primary hepatocytes; results must be verified in primary hepatocyte or in-vivo human liver context.
• ERK mechanism not fully resolved — the specific RNA-binding protein or ARE-regulatory complex downstream of ERK that executes mRNA stabilization was not identified in this paper.
• Hamster lipid physiology differs from human lipoprotein metabolism in important ways (e.g., hamsters lack CETP activity equivalent to humans, affecting HDL metabolism); magnitude extrapolation to humans is approximate.
• Single-lab provenance — the mechanistic claims (SREBP-2 independence, ERK dependence, 3’UTR specificity) originated from one group and required subsequent independent confirmation before becoming consensus. needs-replication

no-fulltext-access needs-replication

---

Cross-references

Page	Relationship
berberine	Primary subject compound; this is the canonical origin citation for LDLR-mechanism claims
ldlr	Direct target: LDLR mRNA is stabilized; protein expression increased
hmgcr	Mechanistic contrast: statin target; Kong 2004 establishes that berberine does NOT work through HMGCR/SREBP-2 axis
lipoprotein-metabolism	Pathway context: LDLR-mediated LDL uptake; berberine acts on a post-transcriptional node within this pathway
ampk-activators	Berberine’s AMPK-activation mechanism is covered here; distinct from the LDLR mRNA-stabilization mechanism in Kong 2004

---

Footnotes