Why You Can't Swallow Most Peptides
Oral Peptide Delivery
<2% Bioavailability
Most peptides that reach your stomach are more than 98% destroyed before entering the bloodstream. Three barriers make oral peptide delivery one of pharmaceutical science's hardest problems.
Kommineni et al., Pharmaceutical Research, 2023
Kommineni et al., Pharmaceutical Research, 2023
View as imageEvery GLP-1 injection, every insulin pen, every subcutaneous peptide dose is a reminder that the pharmaceutical industry has not solved oral peptide delivery. The reason most peptides require injection is straightforward: your digestive system is a peptide-destroying machine. It is designed to break dietary proteins and peptides into individual amino acids for absorption. It does not distinguish between a therapeutic peptide and the proteins in your lunch.[1] The result: oral bioavailability for most peptides falls below 2%, and for many, it is effectively zero. This article breaks down the three barriers, explains why oral semaglutide is the exception, and covers the strategies being developed to overcome this problem. For a broader look at the delivery landscape, see our pillar article on The Future of Oral Peptide Drugs.
Key Takeaways
- Most therapeutic peptides have oral bioavailability below 2%, which is why nearly all peptide drugs require injection
- Three sequential barriers destroy oral peptides: gastric acid (pH 1-3) denatures structure, proteases (pepsin, trypsin, chymotrypsin) cleave bonds, and the intestinal epithelium blocks absorption of molecules above ~500 Da
- Oral semaglutide (Rybelsus) achieved 0.4-1% bioavailability using SNAC (sodium N-[8-(2-hydroxybenzoyl)amino] caprylate), a permeation enhancer that creates a transient local pH shift
- Cyclosporine A, a cyclic peptide of 1,202 Da, achieves ~30% oral bioavailability through N-methylation, cyclization, and lipophilicity (Wang & Craik, Biopolymers, 2016)
- Combining SNAC with C10 (capric acid) in oral tablets improved gastric peptide delivery in a 2025 preclinical and clinical study (Niu et al.)
- Self-immolative conjugate technology enabled oral peptide delivery by responding to intestinal inflammation signals (Cheng et al., Science Advances, 2026)
Barrier 1: Stomach Acid Destroys Peptide Structure
The human stomach maintains a pH between 1 and 3 during digestion. This acidic environment serves a biological purpose: it denatures dietary proteins, unfolding their three-dimensional structures to expose peptide bonds for enzymatic cleavage.
Therapeutic peptides face the same treatment. At pH 2, the protonation state of amino acid side chains changes, disrupting hydrogen bonds and electrostatic interactions that maintain peptide conformation. For peptides whose activity depends on specific folding (most peptides longer than 10-15 amino acids), acid denaturation alone can destroy biological function even before any bond is cleaved.
Pepsin, the dominant gastric protease, is maximally active at pH 3. It preferentially cleaves peptide bonds adjacent to hydrophobic amino acids (phenylalanine, tyrosine, leucine). A 28-amino-acid peptide like ghrelin contains multiple pepsin cleavage sites. Within minutes of gastric exposure, it is fragmented into pieces too small to bind its receptor.
The stomach provides about 15-90 minutes of acid exposure depending on gastric emptying time. During this window, peptides are simultaneously denatured by acid and cleaved by pepsin. Enteric coatings can bypass this barrier by encapsulating the peptide in pH-sensitive materials that dissolve only in the neutral environment of the small intestine. But surviving the stomach is only the first problem.
Barrier 2: Intestinal Proteases Shred What's Left
The small intestine presents an even more aggressive enzymatic environment. Pancreatic secretions deliver a battery of proteases into the duodenum:
- Trypsin cleaves after basic amino acids (arginine, lysine)
- Chymotrypsin cleaves after aromatic and large hydrophobic residues
- Elastase cleaves after small hydrophobic residues (alanine, glycine, serine)
- Carboxypeptidase A removes C-terminal hydrophobic amino acids
- Carboxypeptidase B removes C-terminal basic amino acids
Together, these enzymes provide overlapping specificity that covers nearly every possible peptide bond. A peptide that escapes pepsin in the stomach faces trypsin and chymotrypsin in the duodenum. Any fragments that survive those encounter elastase and the carboxypeptidases.
The half-life of most linear peptides in simulated intestinal fluid is under 30 minutes. For some, it is under 5 minutes. The degradation is so rapid and complete that even very high oral doses produce no detectable intact peptide in portal blood.
The brush border membrane of intestinal epithelial cells adds another layer of enzymatic attack. Membrane-bound peptidases (aminopeptidase N, dipeptidyl peptidase IV, angiotensin-converting enzyme) sit on the luminal surface of enterocytes and cleave peptide fragments as they approach the absorption surface. Even peptides that survive the luminal proteases face a final enzymatic gauntlet at the absorption site itself. This is why DPP-4 inhibitors (sitagliptin, saxagliptin) work for diabetes: they block the brush border enzyme that would otherwise destroy GLP-1 and GIP before absorption into the bloodstream.
The sheer redundancy of proteolytic defenses means that no single modification or formulation trick can overcome the enzymatic barrier for most peptides. Successful strategies typically combine multiple protective approaches: acid-resistant coatings, protease inhibitors, permeation enhancers, and absorption promoters working together.
Strategies to combat enzymatic degradation include D-amino acid substitution (proteases cannot cleave mirror-image bonds), cyclization (closing the peptide backbone into a ring eliminates vulnerable termini), and co-administration of enzyme inhibitors that temporarily suppress protease activity in the intestinal lumen.
Barrier 3: The Intestinal Wall Blocks Absorption
Even if a peptide survives acid and enzymes intact, it still must cross the intestinal epithelium to reach the bloodstream. This is arguably the most fundamental barrier.
The intestinal epithelium is a single layer of cells connected by tight junctions. These junctions allow small molecules (under roughly 500 Daltons) to pass through paracellular spaces between cells. Most peptide therapeutics are well above this cutoff. Semaglutide weighs 4,113 Da. Insulin weighs 5,808 Da. Even small bioactive peptides like BPC-157 (1,419 Da) exceed the paracellular limit.
Transcellular absorption (crossing through the cell) is theoretically possible but requires the molecule to pass through two lipid bilayer membranes (apical and basolateral). Peptides are generally too hydrophilic and too large for passive transcellular diffusion. The exceptions, like cyclosporine A (1,202 Da, oral bioavailability ~30%), succeed because they are extremely lipophilic, N-methylated, and cyclic, properties that make them behave more like small molecules than typical peptides.[2]
Permeation enhancers address this barrier by transiently opening tight junctions or increasing transcellular transport. SNAC, the enhancer used in oral semaglutide, works primarily by creating a local pH microenvironment that promotes transcellular absorption of the peptide through gastric epithelium.
How Oral Semaglutide Solved (Part of) the Problem
Oral semaglutide (Rybelsus) is the first and, as of 2026, only oral GLP-1 receptor agonist. Its success depended on a specific permeation enhancer: SNAC (sodium N-[8-(2-hydroxybenzoyl)amino] caprylate).[3]
SNAC works through a specific mechanism at the gastric epithelium. It creates a local pH increase near the stomach wall, which protects semaglutide from acid denaturation in the immediate vicinity of the absorption site. It also promotes monomeric dissolution of semaglutide and enhances transcellular transport across gastric epithelial cells.
The net result is an oral bioavailability of approximately 0.4-1%. That sounds terrible. It is, by small molecule standards. But it works because semaglutide is potent enough that even 1% absorption from a 14 mg oral tablet delivers a therapeutically effective dose. The injectable version uses 0.25-2.4 mg. The oral version compensates for 99% loss with a massively higher starting dose.
This approach has inherent limitations. It requires fasting (food in the stomach dilutes the SNAC concentration gradient and reduces the local pH effect). It requires taking the tablet with no more than 4 ounces of water. It is sensitive to gastric motility, stomach pH variability, and individual differences in gastric emptying. For more on how this works, see How SNAC Makes Oral Semaglutide Possible.
Solis-Herrera et al. (2024) reviewed the SNAC mechanism in detail and concluded that it represents a paradigm shift for oral peptide delivery, though its applicability to peptides less potent than semaglutide remains limited.[4]
What Makes a Peptide Orally Absorbable
Cyclosporine A taught pharmaceutical science the rules. Wang and Craik (2016) analyzed the chemical features that make certain cyclic peptides orally bioavailable:[2]
Cyclization. Closing the peptide backbone into a ring eliminates free N- and C-termini (the primary targets of exopeptidases). It also constrains conformational flexibility, reducing the entropic penalty of membrane crossing.
N-methylation. Replacing amide NH groups with N-methyl groups does three things. It blocks hydrogen bond donation (reducing aqueous solubility and increasing membrane permeability). It shields the peptide bond from protease recognition. And it reduces the number of rotatable bonds, improving oral absorption.
Lipophilicity. Oral peptides need to be hydrophobic enough to cross lipid membranes. Cyclosporine A has a calculated LogP of approximately 2.9, placing it firmly in the range of orally absorbed small molecules.
Size limit. The "rule of 5" for small molecules (molecular weight under 500 Da) does not strictly apply to cyclic peptides, but there is still an upper bound. The largest reliably orally absorbed cyclic peptides are around 1,000-1,200 Da. Above that, even with cyclization and N-methylation, passive permeability drops precipitously.
Intramolecular hydrogen bonding. In nonpolar environments (like lipid membranes), peptides that can form intramolecular hydrogen bonds shield their polar groups from the hydrophobic interior of the membrane. Cyclosporine A is a master of this: it adopts a "closed" conformation in nonpolar solvents where all amide groups are internally satisfied. This chameleonic behavior, alternating between an "open" conformation in water and a "closed" conformation in membranes, is a hallmark of orally bioavailable cyclic peptides.
Molecular flexibility. Paradoxically, some conformational flexibility helps oral absorption. Rigid molecules may not be able to adopt the membrane-permeable conformation when they reach the lipid bilayer. The optimal oral peptide is constrained enough to resist proteolysis but flexible enough to conformationally adapt during membrane crossing.
These rules explain why most therapeutic peptides cannot be made oral simply by chemical modification. Semaglutide (4,113 Da) is far above the size limit. Insulin (5,808 Da) even more so. Neither can be cyclized without losing activity. Formulation-based strategies (SNAC, enteric coatings, nanoparticles) are required for these larger molecules.
The Next Generation: Emerging Delivery Technologies
Several technologies are advancing beyond SNAC.
Niu et al. (2025) demonstrated that combining SNAC with C10 (capric acid, a medium-chain fatty acid permeation enhancer) in erodible oral tablets improved gastric peptide delivery in both preclinical and clinical studies. The dual-enhancer approach achieved higher and more consistent absorption than either enhancer alone.[5]
Cheng et al. (2026) published in Science Advances a self-immolative conjugate system that releases peptide cargo specifically in response to intestinal inflammation. The conjugate survives the upper GI tract intact and breaks apart only when it encounters reactive oxygen species (elevated in inflamed intestinal tissue), releasing the active peptide at the site where it is needed. This inflammation-triggered approach could enable oral delivery of anti-inflammatory peptides directly to diseased gut tissue.[6]
Tyagi et al. (2023) used systems biology and peptide engineering to optimize an oral peptide formulation through clinical translation, demonstrating that computational modeling of intestinal absorption barriers can guide rational dosage form design rather than relying on empirical screening.[7]
Microneedle patches represent a parallel approach that sidesteps the oral route entirely. Rather than fighting the GI tract, microneedle patches for peptide delivery use painless arrays of dissolving needles to deliver peptides directly through the skin. This approach achieves bioavailability comparable to subcutaneous injection without the compliance challenges of syringes. For GLP-1 agonists specifically, several microneedle patch formulations are in clinical development as alternatives to both injection and oral delivery.
The fundamental question for each peptide therapeutic is whether oral delivery is worth pursuing at all, or whether alternative non-injection routes (intranasal, transdermal, buccal, rectal) offer a better cost-benefit ratio for that specific molecule's properties.
These approaches share a common principle: they do not make the peptide itself oral. They engineer the delivery system to protect the peptide through the GI tract and promote absorption at a specific site. The peptide remains fragile. The packaging does the work. For context on related delivery innovations, see Lipidated Peptides: Why Adding Fat Makes Peptides Last Longer and Intranasal Peptide Delivery.
The Bottom Line
The oral bioavailability problem for peptides is not a single barrier but three sequential ones: gastric acid denatures structure, intestinal proteases cleave bonds, and the epithelial wall blocks absorption of molecules above ~500 Da. Oral semaglutide's success with SNAC showed that formulation engineering can overcome these barriers for sufficiently potent peptides, but the 0.4-1% bioavailability sets a high floor on required dosing. Next-generation approaches combining multiple permeation enhancers, inflammation-responsive conjugates, and computational delivery optimization are expanding the toolkit, but oral delivery of large peptides like insulin remains an unsolved challenge.