Targeting peptides represent an attractive approach to overcome one of the biggest hurdles of nucleic acid therapeutics: tissue selectivity. Organ- or cell-specific peptide ligands can be conjugated to siRNA, mRNA or antisense oligonucleotides to redirect these therapeutics away from the liver and towards diseased tissue. Peptides are modular, chemically defined targeting modules that can be synthesized in tandem with the nucleic acid, allowing homogeneous conjugates rather than the heterogeneous lipid nanoparticles. The small size of peptides maintains the pharmacokinetic advantages of the nucleic acid backbone and provides an additional layer of recognition to interact with receptors on target cells, which should improve therapeutic index and minimize off-target toxicity.
Fig. 1 Formulation of peptide-based nanoparticles in the presence of different nucleic acids and their cellular internalization.1,5
Tissue selectivity is another key hurdle as nucleic acid therapeutics are programmable in sequence specificity but lack inherent mechanisms to distinguish between target cells and bystander tissues. The small size and polyanionic nature result in rapid glomerular filtration and widespread distribution, which subject liver, spleen, kidney, and bone marrow to supratherapeutic doses and consequent activation of innate immune sensors and organ toxicity. In addition, delivery platforms in use today enrich payload selectively in organs of first-pass metabolism, thus creating a therapeutic ceiling where the amount of drug needed to be effective in the disease site exceeds tolerable systemic dose. The inability to restrict activity to diseased microenvironments leads to small therapeutic indices, frequent adverse events, and clinical attrition for indications beyond liver or ocular diseases.
Unmodified nucleic acids are able to spread systemically: their polyphosphate backbone is hydrophilic and thus not membrane-permeable, which additionally induces their ultrafiltration through the kidneys. They tend to get trapped in filtering organs, mainly kidney tubules, hepatic sinusoids, and reticuloendothelial cells of the spleen where they activate toll-like receptors and release inflammatory cytokines. Due to the systemic spread, only a small fraction of an administered dose reaches the target tissue. Binding of nucleic acids to cells of these healthy organs causes inadvertent gene silencing or expression and/or activation of the immune system. As a result, the administered doses must be lowered to tolerable levels, thus lowering their efficacy. Tissue specific targeting is not encoded in the structure, so not even cell-penetrating peptides (CPPs), unless designed to be stealthy, are able to avoid binding to serum proteins and losing their ability to home to the target. Toxicity can be initiated by several factors, such as immune activation by toll-like receptors (TLRs) 3, 7 and 8 (to which double-stranded RNA can bind), complement activation (initiated by cationic nanoparticles), or simple metabolic stress (associated with the uptake of large amounts of nucleic acids by the liver hepatocytes). Accumulation of nucleic acids in kidneys leads to tubular epithelial cell apoptosis and proteinuria, while accumulation in the liver results in increased transaminases and bilirubin levels. As a result, the doses required for substantial gene silencing or replacement protein production in many solid tumors are far above the doses that could be safely administered. A choice between poor efficacy and severe toxicity is an unacceptable situation for any drug developer, and has been the cause of many failed nucleic acid-based therapeutics.
Lipid nanoparticles and polymer complexes are biased for the liver and spleen because of apolipoprotein E binding and reticuloendothelial clearance. As a result, the hepatocyte is the default target of most currently used LNPs. Although this is desirable for liver-directed therapy, this feature imposes a severe limitation on their application to other solid malignancies or neurological diseases. Attempts to re-target LNPs with alternate lipid formulations or targeting ligands have led to only incremental improvements, as the particle itself is still rapidly filtered by the liver. When other organs are reached, the payload concentration is often several orders of magnitude lower than that reached in the liver and thus would require orders of magnitude higher doses to reach therapeutic levels, reintroducing systemic toxicity. In short, it is currently impossible for today's platforms to achieve sufficient penetration of lung, pancreas, or brain parenchyma at therapeutic levels without producing an unacceptable hepatic dose. Most carriers cannot differentiate between a normal cell and a malignant cell in the same organ, they simply deliver payload to all cells in which the target pathway is expressed. In liver, this means that both hepatocytes and malignant hepatocytes will take up siRNA, risking hepatotoxicity. In a tumor mass, LNPs will accumulate preferentially in the vascularized periphery of the tumor but not in hypoxic regions at its core where the same receptors may be downregulated, missing inner tumor nests. This lack of microenvironmental selectivity leads to a lower therapeutic index and promotes resistance as sub-lethal dosing of partially treated cells will select for escape variants. Peptide-targeted systems represent a way to recognize neo-epitopes only found on malignant cells, but it is non-trivial to incorporate such peptides into nanoparticles without destabilizing the particle or losing encapsulation efficiency of nucleic acids.
Table 1 Peptide-Targeted vs Conventional Nucleic Acid Delivery
| Feature | Conventional LNPs | Peptide-Targeted Systems |
| Biodistribution | Liver/spleen dominant | Receptor-directed |
| Tissue discrimination | Poor | High (neo-epitope dependent) |
| Immune stimulation | TLR activation possible | Reduced via masking |
| Dose ceiling | Systemic toxicity | Tissue-specific dosing |
| Clinical applicability | Liver/eye limited | Broad (if validated) |
Targeting peptides act to transform nucleic acid delivery from passive to active and receptor-mediated recognition. Targeting peptides selectively recognize cell-surface markers unique to diseased tissue in vivo. This causes selective binding of the nucleic acid to areas of need while simultaneously avoiding normal organs and tissues. This "molecular sieving" reduces the circulating peptide and nucleic acid exposure in normal organs and decreases dose-limiting toxicities and increases the therapeutic window. Peptides can be appended to any type of nucleic acid cargo, including siRNA, mRNA, antisense oligonucleotides or gene editing ribonucleoproteins, without significant re-formulation making them a versatile modular platform for nucleic acid medicines. Optimization of peptide-mediated nucleic acid delivery requires a balance of receptor affinity, serum stability, and endosomal escape to ensure that targeting leads to functional intracellular delivery rather than mere membrane binding.
Cell-surface recognition is the first step: peptides bearing RGD or bombesin mimetic motifs bind to integrins or GRP receptors that are overexpressed on the surface of tumor cells, or GalNAc conjugates that bind to asialoglycoprotein receptors on hepatocytes. The complex will be endocytosed into the cell by clathrin- or caveolae-mediated endocytosis and internalized into an endosome. For effective delivery, the endosomal escape step is critical for release of the peptide-nucleic acid complex before the endosome is degraded in the lysosome. This can be achieved by fusing CPPs such as TAT to the cargo or by including CPPs that form branching networks within the endosome and disrupt endosomal membranes through the proton-sponge effect. The density and turnover of the receptor can therefore directly influence the selectivity and efficiency of loading; receptors with a low expression level might be easily saturated with the complex, whereas receptors with a high recycling rate might result in multiple recycling back to the cell surface and a limited release to the cytosol. For this reason, the choice of receptor also needs to consider trafficking kinetics as well as endosomal acidification kinetics to ensure that binding to a receptor results in functional delivery.
A key benefit of active targeting, as compared to passive targeting, is the separation of delivery from filtration by size. Passive accumulation of lipid nanoparticles is primarily liver-based and non-selective, so the siRNA or mRNA can accumulate in Kupffer cells, splenic macrophages, and other non-targeted compartments. Systems with peptide guidance, by contrast, leverage molecular recognition to avoid accumulation in "sink" organs and to instead deliver siRNA or mRNA specifically to any cell that expresses the cognate receptor. This also decreases the overall dose needed for therapy and reduces the probability of non-specific gene modulation of healthy tissues. In addition, active targeting can be used to target extracellular-matrix-embedded tumors, or target endothelial barriers such as the blood-brain barrier, which are otherwise not accessible to passively diffusing therapeutics. The downside is that manufacture is more complex, as peptide-nucleic acid conjugates must be purified from unconjugated reagents, and receptor binding must be validated across disease models to ensure that active uptake results in active silencing or expression.
Targeting peptides can be broadly compatible with the various nucleic acid modalities. For siRNA, branched cationic CPPs can be used to form reducible nucleic acid networks that electrostatically condense the duplex, protecting it from serum RNases, and are released in the cytosol upon reduction. For mRNA, peptide-lipid hybrids or CPPs fused to lipid anchors can mediate endosomal escape and improve translation efficiency in primary cells and in vivo. Antisense oligonucleotides benefit from receptor-mediated uptake to achieve efficient internalization, and this avoids the high phosphate backbones required for electrostatically mediated delivery that drive renal clearance. Gene-editing tools, including CRISPR ribonucleoproteins, can be directly guided by CPPs or be packaged in CPP-engineered lipid nanoparticles that deliver the RNP complex directly to the nucleus. Payload stability is a shared problem across all modalities, as the peptide must avoid proteolysis and the nucleic acid must remain intact until cytosolic delivery can occur. To this end, successful designs use D-amino acids, N-methylation, or cyclisation in the peptide backbone and modified nucleotides (2'-O-methyl, phosphorothioate) in the cargo.
Table 2 Comparative delivery modalities for nucleic acid therapeutics
| Delivery Strategy | Targeting Mechanism | Tissue Discrimination | Key Advantage | Primary Limitation |
| Lipid nanoparticles | ApoE-mediated liver uptake | Poor (hepatotropic) | High siRNA loading | Off-target hepatotoxicity |
| Peptide-siRNA conjugates | Receptor-mediated endocytosis | Excellent (marker-driven) | Small size, deep penetration | Limited payload capacity |
| GalNAc conjugates | Asialoglycoprotein receptor (liver) | Liver-specific | High potency in hepatocytes | Restricted to hepatic targets |
| Aptamer chimeras | Nucleic acid-protein binding | Moderate (affinity-dependent) | Dual function (target + drug) | Susceptible to nuclease degradation |
An optimal peptide nucleic acid (PNA) conjugate design must balance three conflicting parameters: receptor-binding affinity, serum stability, and endosomal escape. In contrast to small-molecule drugs, nucleic acids require intracellular delivery for functional activity. This constraint necessitates that the peptide carrier serves a dual role of targeting ligand and membrane-translocation device. Therefore, some of these potential tradeoffs are the dual functionality in PNA conjugates, that the Cationic residues that aid in PNA complexation may lead to non-specific serum protein binding, while hydrophobic motifs that aid in membrane penetration may also decrease aqueous solubility. Also, the peptide: nucleic acid stoichiometry, which eventually dictates the overall nanoparticle size and charge can also dramatically impact biodistribution. PNA conjugates are therefore often created with iterative cycles of rational design, synthesis, biophysical characterization, and functional testing to ensure an appropriate targeting affinity without sacrificing payload integrity or cellular uptake.
The first strategy is covalent attachment, which has the advantages of a defined stoichiometry and a stable linkage, typically a disulfide (that can be reduced in the cytosol) or an amide bond (at a terminal residue). Covalent binding of the peptide to the nucleic acid prevents dissociation during circulation, with consequent premature release of the payload. The covalent conjugation chemistry, however, must be carefully controlled, so as not to interfere with nucleic acid hybridization or peptide folding. An alternative is non-covalent assembly of the complex, most commonly through electrostatic interactions between cationic peptide domains (e.g., a poly-arginine motif) and the nucleic acid backbone. Such a complex is a dynamic nanoparticle whose size can be adjusted through the mixing ratio. Non-covalent complexes are also much easier to prepare and offer modular loading of the cargo, but present issues of batch-to-batch variability, and they may dissociate upon dilution in the blood, leading to less efficient targeting and degradation of the unbound nucleic acids by nucleases. The conjugation approach also has direct effects on pharmacokinetics and targeting specificity. Covalent conjugates have a predictable half-life and tumor uptake, but the binding epitope may be sterically occluded by the attached nucleic acid, leading to a loss in affinity. Non-covalent assemblies can reach a high peptide/nucleic acid ratio, which increases avidity but the resultant positive surface charge leads to non-specific adhesion to anionic basement membranes, which in turn leads to accumulation in lung and kidney as well as tumor. This non-specific sequestration can be partially addressed by the use of a flexible linker (e.g., polyethylene glycol or a β-alanine repeat) to decouple the nucleic acid cargo from the recognition domain.
Endosomal escape is the rate-limiting step for nucleic acid function. But the same properties that enhance membrane disruption tend to limit targeting specificity. Peptides that are highly cationic form nucleic acid complexes readily and stimulate macropinocytosis but have limited selectivity due to non-specific binding to healthy tissue. Peptides with high affinity for a given receptor are prone to internalization via clathrin-mediated endocytosis and entrapment in late endosomes, with subsequent cargo retention. Designers are forced to calibrate hydrophobic moment and net charge to a narrow range where endosomal escape is potent in the target cell but is lost in bystander cells. Cyclization can pre-organize the peptide in a conformation that is still compatible with receptor binding but restricts the conformational entropy required for non-specific membrane insertion. Achieving a balance between these competing objectives involves a series of structural compromises: including D-amino acids increases protease resistance, which comes at the potential cost of altered receptor recognition; adding fusogenic motifs can improve escape but leads to higher off-target cytotoxicity. One typical compromise is the use of conditional chemistries, such as pH-sensitive histidine pairs that protonate in the acidic environment of the endosome, to induce membrane destabilization in this target compartment alone. A second strategy is to use cleavable linkers that release the nucleic acid cargo after internalization, thus allowing the peptide to adopt an endosomolytic conformation in a way that does not expose this activity during systemic circulation.
The peptide-to-nucleic acid ratio is critical in controlling the size, charge, and morphology of the nanoparticles and their delivery functionality. Excess peptide results in positively charged particles that can aggregate in serum, bind proteins nonspecifically, and are rapidly cleared by the liver. However, not using enough peptide does not allow full encapsulation of the nucleic acid cargo, which can be degraded by nucleases or fail to be taken up by the cell. Ratios are usually within a small range in which the particles are small (<50 nm) and close to neutral in charge, allowing them to penetrate deeply into the tumor while reducing nonspecific interactions. The functionality of the delivery is determined not by the cell uptake, but by its gene silencing or expression efficiency. This in turn depends on their ability to escape the endosome and release the nucleic acid into the cytosol. Developers avoid using high excesses of peptides that may drive uptake by adding conditional release linkers that are cleaved upon receptor binding, thus releasing the nucleic acid from the peptide carrier, and preventing the peptide from translocating more non-target cells. High peptide content is more costly in the manufacturing process and increases immunogenicity and off-target toxicity. Peptides that self-assemble into larger structures like amyloid-like fibrils can be especially prone to toxicity and retention in liver and spleen due to their superstructures being difficult to degrade. Minimizing excess peptide may be done by iterative titration experiments to find the lowest concentration of peptide that fully saturates the nucleic acid without forming large assemblies. Alternatively, it may be possible to incorporate nuclear localization signals or endosomal escape motifs directly into the nucleic acid sequence. This decreases the peptide load required and allows for lower carrier to cargo ratios without losing efficacy. This self-limiting method allows each peptide to only deliver its payload once, reducing off-target toxicity.
Peptides are inherently limited by several factors that make their development to clinical agents a challenge. Natural ligands are often present at receptors at high concentrations, tumor or cell types often express varying levels of targets which reduce the selectivity of a targeting peptide and peptides are inherently unstable due to proteolysis, which reduces the half-life in vivo. All of these factors are interrelated, the high binding affinity required to overcome the natural ligand leads to reduced fraction of bound peptide at the target, variations in tissue expression leads to altered pharmacokinetics, and a short half-life requires more frequent dosing, which is often associated with increased costs and higher toxicity. To date, strategies to address these factors include cyclization of peptides, incorporation of non-natural amino acids and protective groups that can be conditionally removed in vivo, all of which significantly increases the synthetic burden and are only a partial solution to overcoming these issues. The following sections briefly review how these factors limit the use of peptides as targeting moieties.
Fig. 2 Overview of Targeting and Delivery Strategies for Peptide-based Delivery Systems.2,5
Most receptors have physiologic roles and are occupied in their natural state by endogenous growth factors and hormones that are found at micromolar concentrations in the circulation. Thus the ligand must compete for a finite number of available surface epitopes and the affinity advantage of the targeting peptide may be insufficient to displace endogenous ligands. This is especially true for integrins bound to ECM proteins or for receptors with high-affinity serum ligands (e.g., transferrin). Peptide binding may also be dynamically exchanged, which truncates residence time and intracellular payload build-up. This limitation can be partially overcome by choosing a receptor that has a low circulating ligand concentration, or by designing a peptide that binds to an allosteric site distinct from the natural ligand pocket, but such opportunities are limited and receptor family-specific. In addition, target receptors are only present at a limited number of copies per cell (thousands to millions). Thus when the sites become saturated, no further increase in uptake can be achieved, which sets a pharmacodynamic upper limit. Clearance receptors in the liver and kidney can be occupied by the peptide at high doses, leading to paradoxically longer systemic half-life and unwanted increase in off-target exposures. High-dose saturation of target receptors also interferes with expected dose-response relationships by violating the assumption of linearity at doses greater than the receptor occupancy threshold. This makes it difficult to predict therapeutic outcome as a function of administration rate. This can be partially overcome by using a split dosing regimen or a receptor-boosting approach to transiently increase expression of the target, but these increase clinical complexity and can trigger off-target biological feedback responses.
Receptor density can vary several-fold between nodules in the same tumor, metastatic sites, or patient populations. A peptide selected for high-expression lesions will thus have subtherapeutic payload delivery to low-expression micrometastases. This can lead to a heterogeneous response where some clones are eradicated and others survive to seed relapse. Expression in stromal cells and tumor-associated macrophages also dilutes the available concentration for targeting malignant cells. The density variability confounds the predictability of peptide-based delivery. Pharmacokinetic modeling with regard to receptor density averages therefore do not adequately predict the spatially mosaic expression of in vivo tumors. Validation in multiple cell lines and patient-derived xenografts covering the expression spectrum is therefore critical to avoid overestimation of clinical performance. In many cases, even if a receptor is enriched on diseased tissue, low-level expression on critical organs such as liver sinusoids, renal glomeruli, or pulmonary capillaries can lead to a "bystander effect" where peptide is also taken up in parenchymal compartments. This partial selectivity can therefore drastically shrink the therapeutic window: the dose needed for effective tumor uptake will saturate normal tissue receptors at the same affinity. Thus, dose-limiting toxicities may be reached before complete tumor coverage. The issue is compounded when receptors are down-modulated during therapy; radiation or chemotherapy may cause antigen downregulation in tumors while normal tissue expression remains at baseline levels. The risk-benefit ratio will then progressively shift toward treatment-limiting toxicities. Many peptides therefore fail not for lack of affinity but for lack of absolute biological exclusivity.
Linear peptides composed of L-amino acids are susceptible to serum peptidases, including aminopeptidases, carboxypeptidases, and endoproteases like chymotrypsin. Administration is quickly followed by proteolytic degradation that results in the truncation of recognition epitopes, and the appearance of inactive fragments that compete with the full length peptide for receptor binding. C-terminal amidation and N-terminal acetylation of peptides can modestly improve their metabolic stability, however, these modifications rarely result in plasma half-lives of >1 hour. Other modifications such as cyclization, D-amino acid insertion, or hydrocarbon stapling can also be utilized, though they often have a detrimental impact on receptor affinity when the modification forces the peptide away from the binding conformation. Metabolically stable peptides that do last long enough in the body can also elicit anti-drug antibodies, especially when coupled to immunogenic drug payloads like radiometals and cytotoxins, which further restricts repeat dosing. The short half-life of peptides in circulation, desirable for imaging, complicates their development as therapeutics because it requires prolonged periods of time during which the peptide must remain on target. The rapid clearance of peptides in the kidneys and liver also narrows the therapeutic window for tumor penetration and endosomal escape and often forces developers to use larger bolus doses or continuous infusions that lead to higher systemic exposures. In the case of therapeutic radionuclides, the fact that the peptide half-life is in minutes while the isotope half-life is measured in days means that the carrier protein or peptide is completely metabolized before the isotope is fully decayed, resulting in poor use of the injected radioactive material and necessitating higher injected activities.
Targeting peptides are useful when the target tissue expresses a surface marker that is both abundant, relative to other organs, and capable of active internalisation, while off-target tissues express the same marker at insignificant levels and lack the endocytic machinery. In this case the peptide is a molecular zip code, diverting the nucleic acid payload to where it can be internalised and be of function, and away from tissues where it would otherwise be dumped by passive mechanisms. When non-specific uptake prevails, by scavenger receptors, charge-mediated adsorption or phagocytic clearance, or when the target receptor is obscured by tight endothelium, the peptides confer little over molecular weight and synthetic complexity.
The first prerequisite is high target-to-background expression ratios: a receptor that is ten-fold more highly expressed on tumor hepatocytes than on normal hepatocytes creates a concentration gradient that peptides can leverage, transporting siRNA preferentially into malignant cells while sparing normal parenchyma. This fold-change advantage is necessary because even the most selective peptide will interact with some off-target receptors; a large differential ensures that therapeutic signal exceeds background noise. Equally important are active internalization mechanisms: tissues that mediate clathrin-dependent endocytosis, caveolin-dependent uptake or macropinocytosis can rapidly internalize peptide-nucleic acid conjugates and traffick them toward functional compartments. Tumor endothelium, activated macrophages and certain epithelial cancers offer these pathways, while quiescent fibroblasts or acellular stroma do not, rendering peptide targeting futile.
The dominance of non-specific uptake can make targeting peptide specificity irrelevant: if Kupffer cells in the liver or macrophages in the spleen take up > 80 % of the injected dose independent of the targeting moiety, the engineering of a selective peptide will be futile. In these organs, strategies that provide a protective shield (PEGylation, glycosylation or size modulation) can be more useful than receptor targeting, because the limiting factor is avoidance of the scavenger receptors rather than increased binding to the tumor. Similarly, poor accessibility of the target can render a peptide targeting strategy useless: targets hidden from circulation by basement membranes, blood-brain barrier tight junctions or tumor necrotic cores are inaccessible for peptides circulating in the blood vessels which can only bind to surface exposed antigens. Tissue penetrating peptides or ultrasound-responsive vectors which transiently open vascular pores are more suitable strategies in these settings. Finally, if the receptor is expressed ubiquitously (e.g., integrins on all endothelium), then peptide targeting can simply divert the dose from one normal tissue to another, without any net gain in tumor selectivity.
Achieving tissue selectivity with nucleic acid drugs requires more than efficient cellular uptake. In many programs, delivery failure arises from mismatches between peptide targeting behavior, nucleic acid payload properties, and tissue-specific biology. Our peptide design services are structured to support targeted nucleic acid delivery by integrating payload compatibility, selectivity optimization, and feasibility assessment from the outset. Rather than treating targeting peptides as interchangeable carriers, we design peptide systems that are explicitly tailored to the constraints of nucleic acid payloads and the intended tissue context.
Payload-Aware Peptide Engineering: Nucleic acids impose distinct physicochemical and biological constraints that directly influence peptide design. Our payload-aware peptide engineering approach evaluates how peptide sequence, charge distribution, and structure interact with nucleic acid cargo such as siRNA, mRNA, or antisense oligonucleotides. Peptides are engineered to support stable association with nucleic acids without inducing excessive non-specific uptake or compromising tissue selectivity. This includes controlling electrostatic interactions to avoid peptide-driven biodistribution and designing architectures that preserve targeting function after payload association.
Selectivity-Driven Optimization: Improving tissue selectivity requires prioritizing specificity over maximal uptake. Our selectivity-driven optimization strategy focuses on enhancing receptor-mediated delivery while actively minimizing non-target interactions. Affinity, peptide density, and structural features are tuned to match the expression profile and internalization behavior of the intended tissue target. This helps ensure that improved delivery arises from active targeting rather than passive accumulation, reducing off-target exposure and variability.
Target Validation Logic: Not all receptors that bind targeting peptides support effective nucleic acid delivery. Our feasibility assessment applies a target validation logic that evaluates receptor expression, internalization pathways, and compatibility with nucleic acid release requirements. This assessment helps determine whether a target is likely to support functional delivery to the cytosol or nucleus, rather than simply mediating surface binding or endosomal sequestration.
Early Risk Identification: Tissue-selective delivery programs often fail due to risks that are identifiable early, such as heterogeneous target expression or strong competition from endogenous ligands. We focus on early risk identification to flag these limitations before extensive optimization efforts are undertaken. By identifying feasibility constraints early, we help teams allocate resources toward targets and designs with a higher probability of delivering meaningful tissue selectivity.
If your nucleic acid program shows broad biodistribution, inconsistent tissue targeting, or limited functional activity despite efficient uptake, an early technical discussion can help clarify whether peptide design or target selection is the limiting factor. Discuss your targeted nucleic acid delivery challenge with our team to evaluate tissue-specific feasibility, identify design risks, and define a peptide strategy aligned with your delivery objectives.
References
