Therapeutic peptides fall into three classes: native hormones, engineered analogs, and heterologous leads discovered by screening, design, or phage display. They treat metabolic, malignant, infectious, and pain conditions through receptor activation, protein-protein interaction blockade, and…
Therapeutic peptides are amino acid chains that produce clinical effects by binding receptors, enzymes, or protein surfaces. They are not a single drug format. Products in this category range from small cyclic peptides to large branched constructs, antibody conjugates, and radionuclide-containing molecules. The category matters because peptides are naturally abundant in the body, where they act as hormones, neurotransmitters, growth factors, ion channel ligands, and anti-infectives. Most peptides of this kind work by binding cell surface receptors and triggering intracellular responses.
The size of the field can be measured. An industry overview of peptide therapeutics published November 4, 2020 reported that between 2015 and 2019 the U.S. FDA authorized 15 peptides or peptide-containing molecules out of 208 new drugs, a figure drawn from an editorial by de la Torre and Albericio in Molecules. Of those 208 authorizations, 150 were new chemical entities and 58 were biologics. Peptides therefore accounted for about 7 percent of new drug authorizations in that five-year window. The count deliberately includes molecules that are not simple linear chains, which is why they are described as peptides or peptide-containing molecules. The source does not name the 15 products, so the list cannot be independently inspected here. Clinically, therapeutic peptides are used for metabolic diseases, cancer, cardiovascular diseases, infectious diseases, pain, and hematological diseases. That scope statement comes from the reviewed material and does not quantify how often peptides succeed in each area.
Therapeutic peptides are usually divided into three classes.
Native peptides are natural sequences used for replacement therapy, supplying a peptide the body fails to make. Insulin is the canonical example. Roughly 100 years separate the first extraction of insulin from its clinical use in replacement therapy, and insulin remains essential for type 1 diabetes. Research on the disease continues; the FinnDiane LifeOne study is enrolling 400 people with type 1 diabetes by invitation, alongside healthy ageing controls, to examine how ageing affects the disease NCT07289204 . This is a cohort study, not a peptide trial, and it shows only that the research infrastructure around insulin-dependent disease remains active.
Analog peptides are engineered variants of native sequences. Their sequences are adapted to improve potency, half-life, selectivity, or resistance to degradation. Analogs reportedly constitute the majority of peptides in clinical trials, though the published material provides no trial count for that claim. The best documented analogs are the incretin-based drugs for type 2 diabetes. A Lancet review found that clinical trials of GLP-1 receptor agonists reduce HbA1c by 1 to 2 percent and produce 2 to 5 kg of weight loss, while DPP-4 inhibitors reduce HbA1c by 0.5 to 1 percent with no weight gain PMID 17098089 . These are trial-derived figures for what peptide analogs achieve in metabolic disease.
Heterologous peptides are sequences unrelated to any native peptide. They are discovered by screening natural compound libraries, by rational or computational design, or by phage display. Because they are not bound to the biology of a native ligand, heterologous peptides can deliver longer half-lives and novel functions, but they also carry an increased risk of side effects. That trade-off is asserted in the reviewed material without quantitative safety data attached.
The class a candidate belongs to shapes its development profile. Native peptides carry the lowest novelty risk but inherit the instability of natural sequences. Analogs preserve native signaling biology while improving pharmacokinetics. Heterologous peptides open new functions but enter the clinic with less predictable safety. The reviewed material offers no trial statistics comparing the classes, so those differences are directional rather than measured.
Peptides treat disease through several distinct mechanisms. Knowing which mechanism applies to a given peptide matters for evaluating claims about it.
Replacement therapy supplies a deficient peptide. Insulin in diabetes is the model: provide the missing hormone and restore its downstream signaling. Metabolic hormone modulation goes further, using analogs to tune receptor signaling rather than simply replace a missing hormone. The incretin data above are the clinical anchor. The same drug class is now being tested beyond metabolism. A 2025 review in Pharmacological Research reports that clinical trials of exenatide, liraglutide, and lixisenatide improved motor symptoms in Parkinson's disease, and that liraglutide improved cognitive function in Alzheimer's disease, while noting that long-term neuroprotective effects remain under investigation PMID 40344943 . The metabolic-neurology connection has independent biochemical support: a 2024 review concluded that dysregulation of specific kinases and phosphatases contributes to neuronal insulin resistance, linking impaired brain insulin signaling to Alzheimer's disease, sometimes called type 3 diabetes PMID 37930675 . These findings establish signals worth following, not proof that any GLP-1 receptor agonist slows neurodegeneration.
Targeted delivery uses a peptide's receptor specificity to carry a payload into selected cells. The peptide can deliver a cytotoxic substance or an imaging agent, which is the logic behind peptide-drug conjugates and radionuclide-containing peptides. No specific product is named in the material reviewed, and no efficacy data for this delivery mode are provided there.
Antiviral peptides inhibit viral replication. Peptide antivirals have been investigated against HIV, influenza, and hepatitis. Attention has since focused on SARS-CoV-2, which enters host cells by binding human ACE2 through its viral spike protein. Peptides can be designed to disrupt that protein-protein interaction PPI . The argument, set out by VanPatten and colleagues in Future Medicinal Chemistry, is that PPI interfaces involved in SARS-CoV-2 infection are dynamic and planar, and that peptides are better suited than small molecules to disrupt such broad, flat surface contacts. This is a structural hypothesis, not a clinical result. No peptide targeting the spike-ACE2 interaction has been authorized as a drug.
Ion channel modulation for pain targets the voltage-gated sodium channel NaV1.7, which plays a major role in human pain perception. Selectivity is the central problem. Of the 9 voltage-gated sodium channel subtypes, only 4 are involved in pain signaling; the other 5 have critical roles in heart and muscle function, so a nonselective blocker is dangerous. Spider-venom gating modifier toxins GMTs bind voltage-gated ion channels and are being investigated as selective NaV1.7 inhibitors Agwa et al., Journal of Biological Chemistry 2018 . A University of Queensland group has prepared GMT variants by solid-phase methods and studied how they interact with channels and lipid membranes to achieve selective inhibition. Structural work on these toxins shows that a hydrophobic patch and surrounding cationic ring promote NaV channel promiscuity, meaning the natural molecules hit multiple channel subtypes. Engineering a selective analgesic therefore means editing precisely the features that make the natural toxins broad-spectrum. The evidence stops at mechanism and target engagement; no GMT-based NaV1.7 inhibitor is approved.
Epigenetic modulation is a newer peptide application. Epigenetic changes are involved in many or all cancers, making methylation modulators promising drug candidates. Romidepsin, a peptide-derived epigenetic drug, was approved for different forms of T-cell lymphoma in 2009 and 2011. Beyond approved drugs, the chromatin modulator DPY30 facilitates histone H3K4 methylation by binding ASH2L and plays an important role in hematologic malignancies. Work by Hao Jiang's group at the University of Alabama shows that a peptide derived from ASH2L can disrupt DPY30-ASH2L binding and suppress blood cancer cell growth. This is preclinical mechanism work, and targeting DPY30-ASH2L binding is a potential strategy for treating hematologic malignancy, not an established therapy.
Cyclotides , cyclic peptides that resist proteolysis because of their highly constrained structure, are investigated as scaffolds rather than as standalone drugs. They can present epitopes in a constrained, proteolysis-resistant form. A University of Queensland group linked non-immunogenic cyclotides to VHH7, a nanobody targeting murine class II MHC molecules, and showed that the conjugates can target antigen-presenting cells Kwon et al., ACS Chemical Biology 2018 . The promising feature is delivery of an epitope to the antigen-presenting compartment in a stable format. Whether this works safely and effectively in human vaccines or immunotherapy is unestablished.
Most therapeutic peptides must be synthesized, and synthesis is where many candidates stall. Natural linear peptides have poor proteolytic stability, which limits their direct therapeutic use. Making them clinically useful usually means modifying them: cyclization, non-natural residues, or other structural changes that resist degradation and preserve active conformation. Those modifications put the molecule into the category called complex peptides and impose the synthesis burden that defines the field. Advances in solid-phase peptide synthesis have made it possible to prepare customized peptides with improved function and greater resistance to degradation, but those advances shift the difficulty rather than remove it.
Solid-phase peptide synthesis SPPS assembles a peptide stepwise on a solid support, adding one residue at a time through repeated coupling and deprotection cycles. The chemistry works well for short sequences and progressively worse for long ones. Long sequences accumulate impurities, side products, and incomplete reactions over many cycles, and each partial reaction is carried forward. The result is that even small amounts of incomplete reaction at each step can drastically reduce final purity and yield. Aggregation, secondary structure, steric hindrance, and conformational effects add further failure modes, especially in sequences prone to folding and in cyclic or branched constructs.
The practical guidance that follows from the chemistry is specific. Deprotection must be driven to completion, and real-time monitoring can optimize reaction times to ensure complete deprotection instead of relying on fixed schedules. Real-time monitoring is one of the few controls that directly addresses the failure cascade: an incomplete deprotection poisons the next coupling, and the error propagates through every later cycle. Monitoring converts deprotection from a fixed-time assumption into a verified step. Purity loss scales with chain length, so purification strategy should be planned before synthesis begins, not after the crude product fails. Sequences that aggregate or form secondary structure will not be rescued by longer reaction times alone; the synthesis design itself must be adjusted. Cyclic and conjugated products add steps and purification burden, and that complexity should be budgeted in time and cost. For a team planning a complex peptide, the questions to answer before choosing a route are: how long is the sequence, where is aggregation likely, which residues will be modified or cyclized, and what purity the downstream assay requires. The reviewed material does not say which specific synthetic strategies best overcome aggregation in long peptides; that remains an open technical question.
The evidence base across this field is uneven, and it is worth separating what is proven from what is asserted.
Well supported. The incretin analog data are strong. Trials summarized in The Lancet report specific HbA1c and weight outcomes PMID 17098089 . The GLP-1 receptor agonist results in Parkinson's and Alzheimer's disease come from clinical trials but are symptom-level improvements with unresolved long-term neuroprotection PMID 40344943 . Romidepsin is an approved drug. These are the anchors of the field.
Asserted without data. The claims that therapeutic peptides offer high activity, chemical and biological diversity, and low toxicity, and that they are relatively easy to produce at low cost compared with protein biologics, are not supported by comparative numbers in the material reviewed. The low-cost claim in particular comes from a vendor source with a commercial interest in selling synthesis equipment, and it should be treated as promotional rather than independent data. The same caution applies to advice that a particular type of synthesizer is necessary; instrument specifications are not evidence of therapeutic efficacy.
Thin or absent. The 15 FDA-authorized peptides from 2015 to 2019 are not named, so the count cannot be checked product by product. No statistics are provided for the share of native versus analog versus heterologous peptides in trials. The side-effect risk of heterologous peptides is asserted without incidence data. The spider-venom GMT and cyclotide work is mechanistic and preclinical. The DPY30-ASH2L peptide work is preclinical cell and mechanistic data.
The registered trial record shows how much of the surrounding research is not peptide pharmacology. A completed 28-participant study of concurrent training in prediabetes NCT03502304 and a completed 20-participant study of extra virgin olive oil and red wine polyphenols in obesity NCT03101436 test lifestyle and dietary interventions. A resistance training study in coronary artery disease and heart failure with reduced ejection fraction NCT04638764 has unknown status. A phase 4 trial of tight glycemic control in acute myocardial infarction was terminated after enrolling 40 participants NCT00237471 . None of these registrations evaluates a peptide therapeutic. They matter here only to calibrate expectations: the clinical infrastructure around metabolic and cardiovascular disease is broad, and peptide drugs must compete with, or complement, older interventions in the same space.
Unresolved questions remain. Which specific peptide or peptidomimetic designs can disrupt the SARS-CoV-2 spike-ACE2 interaction with efficacy and safety in people? How can spider-venom GMTs be engineered to inhibit NaV1.7 without affecting the other 5 channel subtypes that cardiac and muscle function depend on? Can cyclotide scaffolds deliver constrained epitopes in human vaccine or immunotherapy settings? Can DPY30-ASH2L peptide disruptors be optimized into peptidomimetics or small molecules with acceptable pharmacology? Which SPPS strategies best control aggregation and secondary structure in long therapeutic peptides? None of these has a settled answer. The difference between a peptide that works in a biochemical assay and one that works in a patient remains the widest gap in the field.
Peptides referenced: Liraglutide, Exenatide, Lixisenatide, Glucagon, GLP-1.
Related reading: Balancing Chemistry and Timelines in Complex Peptide Synthesis, PEC Purification for GLP-1 Manufacturing: Liraglutide Case Study, Hybrid Fragment Synthesis Expands Peptide Manufacturing Options, Ab Biotechnology Cuts Costs with In-House Synthesis.