AI-Optimized Whey Protein Peptides Achieve 89% ACE Inhibition and Lower Blood Pressure in Rats

Large language models identified an optimal enzyme combination that produced whey protein peptides with 89% ACE inhibition, which survived digestion and significantly lowered blood pressure in hypertensive rats while reducing inflammation and boosting antioxidants.

Jiang, Shuai et al.·Journal of agricultural and food chemistry·2025·
RPEP-116212025RETHINKTHC RESEARCH DATABASErethinkthc.com/research

Quick Facts

Study Type
Not classified
Evidence
Not graded
Sample
Not reported

What This Study Found

The AI-optimized multienzyme combination MC5 achieved 89.08% ACE inhibition at 1 mg/mL, significantly outperforming single-enzyme hydrolysis. After simulated digestion, ACE inhibition decreased by only 6.87%. In hypertensive rats, MC5 reduced systolic blood pressure to 125 mmHg and diastolic to 89 mmHg. The treatment significantly lowered TNF-α and IL-6, increased SOD, GSH-Px, GR, and CAT activity, reduced serum renin (1.25-fold) and ET-1 (1.04-fold), and increased NO content 3.15-fold. Four potent peptides were identified: LPEW, LKPTPEGDL, LNYW, and LLL.

Key Numbers

How They Did This

Large language models were used to predict optimal multienzyme combinations for whey protein hydrolysis. The best combination (MC5) was validated through in vitro ACE inhibition assays and simulated gastrointestinal digestion stability testing. In vivo efficacy was assessed in spontaneously hypertensive rats, measuring blood pressure, inflammatory markers, antioxidant enzymes, renin, ET-1, and NO. Molecular docking identified the most potent individual peptide sequences and their ACE binding modes.

Why This Research Matters

This study bridges AI, food science, and cardiovascular health. By using deep learning to optimize enzyme combinations, the researchers achieved ACE inhibition levels (89%) approaching pharmaceutical-grade potency from a common food protein. The critical addition of in vivo data — showing actual blood pressure reduction in rats — elevates this beyond typical in vitro peptide studies. The digestion stability finding (only 6.87% loss) addresses a major concern with food-derived bioactive peptides.

The Bigger Picture

This study exemplifies the convergence of AI-driven drug discovery with functional food development. Using large language models to optimize bioactive peptide production from commodity food proteins could revolutionize how we develop dietary interventions for chronic diseases. The combination of high ACE inhibition, digestion stability, and confirmed in vivo efficacy places these whey peptides among the most promising food-derived antihypertensive candidates reported to date.

What This Study Doesn't Tell Us

The in vivo study was conducted in spontaneously hypertensive rats, which may not fully represent human hypertension. Specific rat numbers per group and treatment duration were not detailed. Human clinical trials are needed. The four identified peptides need individual testing to confirm their contribution to the overall effect. Manufacturing scale-up and cost-effectiveness were not addressed. The LLM methodology details were not fully described.

Questions This Raises

  • ?Would these whey peptides produce clinically meaningful blood pressure reduction in human hypertension?
  • ?Can the AI-optimized enzyme combination be commercially scaled for industrial whey protein processing?
  • ?Do the four identified peptides (LPEW, LKPTPEGDL, LNYW, LLL) work synergistically or could individual peptides be developed as supplements?

Trust & Context

Key Stat:
89% ACE inhibition + in vivo blood pressure reduction LLM-optimized whey peptides achieved near-pharmaceutical ACE inhibition with only 6.87% loss after digestion, confirmed by actual blood pressure reduction to 125/89 mmHg in hypertensive rats
Evidence Grade:
This study combines in vitro optimization with in vivo validation — a stronger evidence package than most food-derived peptide studies. However, the animal data is from a rat model, and human clinical trials are needed. The AI methodology adds novelty but needs reproducibility verification.
Study Age:
Published in 2025, this study represents the cutting edge of AI-driven bioactive peptide development, combining large language models with traditional food science in a novel way.
Original Title:
Deep Learning-Driven Optimization of Antihypertensive Properties from Whey Protein Hydrolysates: A Multienzyme Approach.
Published In:
Journal of agricultural and food chemistry, 73(2), 1373-1388 (2025)
Database ID:
RPEP-11621

Evidence Hierarchy

Meta-Analysis / Systematic Review
Randomized Controlled Trial
Cohort / Case-Control
Cross-Sectional / ObservationalSnapshot without intervening
This study
Case Report / Animal Study
What do these levels mean? →

Frequently Asked Questions

How did AI help discover these blood pressure-lowering peptides?

The researchers used large language models — the same type of AI behind tools like ChatGPT — to predict which combination of enzymes would best break down whey protein into ACE-inhibiting peptides. This AI-driven approach found a five-enzyme combination that achieved 89% ACE inhibition, far better than any single enzyme alone. The AI also helped identify the four most potent individual peptides.

Could drinking whey protein lower blood pressure?

Regular whey protein products contain some bioactive peptides, but not in the concentrated, optimized form produced in this study. The key innovation is using a specific multi-enzyme process to maximize ACE-inhibiting peptides. A specialized supplement or functional food based on this process could potentially support blood pressure management, but human clinical trials are needed to confirm the effect.

Read More on RethinkPeptides

Related articles coming soon.

Cite This Study

RPEP-11621·https://rethinkpeptides.com/research/RPEP-11621

APA

Jiang, Shuai; Mo, Fan; Li, Wenhan; Yang, Sirui; Li, Chunbao; Jiang, Ling. (2025). Deep Learning-Driven Optimization of Antihypertensive Properties from Whey Protein Hydrolysates: A Multienzyme Approach.. Journal of agricultural and food chemistry, 73(2), 1373-1388. https://doi.org/10.1021/acs.jafc.4c10830

MLA

Jiang, Shuai, et al. "Deep Learning-Driven Optimization of Antihypertensive Properties from Whey Protein Hydrolysates: A Multienzyme Approach.." Journal of agricultural and food chemistry, 2025. https://doi.org/10.1021/acs.jafc.4c10830

RethinkPeptides

RethinkPeptides Research Database. "Deep Learning-Driven Optimization of Antihypertensive Proper..." RPEP-11621. Retrieved from https://rethinkpeptides.com/research/jiang-2025-deep-learningdriven-optimization-of

Access the Original Study

Study data sourced from PubMed, a service of the U.S. National Library of Medicine, National Institutes of Health.

This study breakdown was produced by the RethinkPeptides research team. We analyze and report published research findings without making health recommendations. All interpretations are based solely on the published abstract and study data.