Hypothesis

Berberine's glucose-lowering action depends on a microbiome-driven shift in bile acid composition that activates TGR5-mediated GLP-1 release, a pathway that can operate independently of AMPK.

Rationale

• Berberine enriches Akkermansia, Eubacterium and Ruminococcus while suppressing dysbiotic taxa, a change linked to altered microbial metabolism of primary to secondary bile acids [3].
• Secondary bile acids such as deoxycholic acid are potent agonists of the G-protein-coupled receptor TGR5, which stimulates GLP-1 secretion from intestinal L-cells [4].
• GLP-1 enhances glucose-dependent insulin secretion and suppresses glucagon, improving glycemia independent of direct hepatic AMPK effects.
• Although berberine activates AMPK in many tissues, several reports show glucose lowering persists when AMPK is genetically or pharmacologically blocked, suggesting an AMPK-independent route [4].
• Individual response variability reported in clinical trials may stem from baseline differences in bile acid pool composition or TGR5 expression, both shaped by the resident microbiota.

Predictions

If the hypothesis is correct, then in humans taking berberine:

1. Fasting plasma levels of secondary bile acids (e.g., deoxycholic acid, lithocholic acid) will rise proportionally to the increase in Akkermansia abundance.
2. Plasma active GLP-1 will increase concurrently with the bile acid shift, preceding measurable changes in HbA1c.
3. Pharmacological blockade of TGR5 (using a specific antagonist such as SBL-111) will attenuate berberine-induced GLP-1 rise and diminish its glucose-lowering effect, even when AMPK activity remains intact.
4. Conversely, inhibiting AMPK (with compound C) will not abolish the GLP-1 increase or glucose improvement if TGR5 signaling is preserved.
5. Participants with a baseline high ratio of secondary to primary bile acids will show a smaller incremental response to berberine, reflecting a saturated TGR5 pathway.

Experimental Design (n-of-1 crossover)

• Participants: 12 healthy volunteers with baseline fasting glucose 90-110 mg/dL.
• Periods: Four 4-week phases separated by 2-week washouts: (A) placebo, (B) berberine 1500 mg/day, (C) berberine + TGR5 antagonist, (D) berberine + AMPK inhibitor.
• Measurements (twice weekly): fasting glucose, HbA1c (monthly), plasma bile acids (LC-MS), active GLP-1 (ELISA), fecal 16S rRNA sequencing for Akkermansia abundance, peripheral blood mononuclear cell p-AMPK (Western blot).
• Analysis: Mixed-effects models testing phase effects; mediation analysis to assess whether bile acid changes mediate GLP-1 and glucose outcomes.

Falsifiability

The hypothesis is falsified if any of the following occur:

• Berberine fails to raise secondary bile acids despite microbiota shifts.
• GLP-1 does not increase, or its increase is not correlated with bile acid changes.
• TGR5 blockade does not reduce berberine's glucose-lowering effect while AMPK inhibition does.
• AMPK inhibition abolishes glucose improvement even when TGR5 signaling is intact.
• A negative result in any of these tests would reject the proposed microbiome-bile acid-TGR5-GLP-1 axis as the primary mechanism, directing focus back to AMPK-dependent or alternative pathways.