Peptides have found utility as vectors in theranostics, in large part because of their small size, fast pharmacokinetics, and synthetic flexibility which enables a single peptide to be labeled with different diagnostic or therapeutic radionuclides without changing its receptor affinity. With such two-pronged use, the same targeting principle with imaging agent can be used to first image target expression, verify pharmacokinetics in the specific patient, and then use the same logic now with a therapeutic payload to treat only those lesions that are visualized. In this way, the peptide backbone is able to provide an opportunity to straddle both companion diagnostics and personalized therapy, providing the quantitative, patient-specific information required for contemporary precision oncology while also being flexible to molecular adjustments that allow it to be targeted to both SPECT/PET imaging and radionuclide or drug delivery all in the same regulatory package.
Fig. 1 Scheme of peptide-based theranostic nanomaterials. 1,5
A single targeting strategy is employed in theranostics since any disparity between the diagnostic probe and the therapeutic molecule creates uncertainty regarding whether the therapeutic will accumulate in the same cells that were imaged. Diagnostic reagents that recognize similar but distinct epitopes, or possess mismatched affinities, for example, create the risk that the therapeutic payload will be excluded from lesions which were false positive on the scan, giving rise to misplaced confidence and undertreatment. Alternately, normal tissue that accumulates the diagnostic tracer but not the therapeutic version may be spared during imaging but injured during therapy, creating a therapeutic "hole in the donut". Identical peptide sequences for both targeting functions ensures that biodistribution, receptor binding, and clearance profiles are identical so that the image is a reliable predictor of the therapeutic footprint, and quantitative dosimetry can be made with confidence.
In molecular theranostics, a single receptor epitope is used as the targeting address for both a diagnostic radionuclide (for example, 68Ga for PET) and for a therapeutic radionuclide (for example, 177Lu for beta therapy), without altering the amino acid sequence of the peptide. Thus, the targeting ligand is a chemical chameleon that can be used for either imaging or therapy: its recognition function is unchanged but the radionuclide payload changes from one that will emit a photon that can be quantified in the body to one that will deposit cytotoxic radiation within target tissue. As the tracer that is imaged will be therapeutically identical to the agent that will be given to treat the cancer, patient selection and even response monitoring can be achieved with the same chemical entity. The entire theranostic concept is based on the diagnostic and therapeutic processes being connected by the same molecular recognition process. As the pharmacophore of the peptide is the same in both cases (same shape, charge and H-bonding pattern), the same tumor types (expressing the target with a high density) will bind both the tracer (imaging) and the therapeutic analog. This common logic between the tracer and the therapeutic provides an elegant solution to not having to extrapolate between two different molecules and allows the quantitative images to accurately predict absorbed dose.
If the agent that is used for diagnostics and therapeutics binds to different epitopes or if the affinities for the two entities are mismatched, then the distribution of both within the tumor will be different. A PET-hot lesion may not express the second epitope to which the therapy binds; or, the molecular size of the therapeutic construct may not allow it to penetrate the same micro-domains that the smaller imaging probe can. In either case, the therapeutic payload will not be delivered to the imaged cells; this results in sub-therapeutic dosing and clinical failure despite a positive scan. Discordance also invalidates dosimetry, because the image can no longer be used to estimate deposition of radiation; one would be forced to use the population-averaged model that fails to account for inter-lesion heterogeneity. If the targeting strategies for separate diagnostic and therapeutic entities are different, then the clearance kinetics will likely also be different. For example, the diagnostic may be cleared renally while the therapeutic analog is reticulo-endothelial. Normal-organ dose, in this case, will also be different and one will violate the ALARA principle. Further, if the linker chemistry is different, then one may have a premature release of therapeutic agent while the imaging probe remains intact, and therefore, a false sense of safety is perceived. With a single peptide, one will not have these variables. The SUV, measured on PET, will go into the absorbed dose directly and any off-target uptakes that are identified on a scouting scan will need to be shielded against during therapy.
Table 1 Unified vs Separate Targeting in Theranostics.
| Parameter | Unified Peptide Strategy | Separate Diagnostic/Therapeutic | Clinical Risk of Separation |
| Target engagement | Identical epitope | Possibly different epitopes | Undertreatment of imaged lesions |
| Clearance route | One pathway | Multiple pathways | Unpredictable organ dosimetry |
| Image-to-dose translation | Direct SUV-to-Gy | Extrapolated models | Under- or over-dosing |
| PK variability | Minimal | Compound-dependent | False-negative safety signals |
Targeting peptides represent a pharmacologically intermediate class that share properties with both large antibodies and small molecule agents and can be engineered with a set of properties that can be well suited for theranostic use. The relative small size of peptides allows them to extravasate and diffuse from the vasculature into the tissue quickly, while still being sufficiently small enough to be cleared by the kidneys, giving them a pharmacokinetic profile that is compatible with short half-life PET tracers and limit their exposure in normal tissue. Peptides can also be synthetically modified to include a diagnostic or therapeutic agent at a specific residue without interfering with their binding, as the peptide's binding region is different from the region modified for the payload. As the same targeting concept can be used for both therapy and imaging using peptides, there is less biological variability than in a single-agent strategy. The ability to directly transition from a quantitative imaging to personalized dosing and treatment monitoring can provide a streamlined work flow.
Fig. 2 Overview of the projected review. 2,5
Low molecular weight (<5000 Da): The relatively low molecular weight of peptides (～600 Da per residue, and usually <50 residues) leads to a small hydrodynamic volume, allowing them to diffuse into tissues more easily, in comparison to antibodies. This can be especially important in hypoxic regions of tumors. This property can also be used for target modulation purposes. This small size also allows peptides to be cleared rapidly from circulation. The rapid clearance of unbound tracer results in less radiation exposure of non-target organs and allows imaging and therapy to be performed on the same day, thus, reducing cost and time. The clearance is often highly predictable, which simplifies dosimetry and also allows short circulation half-life that is easily synchronized to the half-life of the diagnostic radionuclides provided by a generator, resulting in maximum tumor targeting at the time of imaging while the radioisotope is still at a high-enough activity. Target occupancy is high and rapid, which also leaves time for internalization and release of the radioactive payload within the tumor cell before clearance, particularly for peptides which internalize via receptor-mediated endocytosis. Overall, this creates a tight therapeutic window where the therapeutic activity can be efficiently killed while non-target tissues are minimally exposed. Peptides are predominantly cleared through renal or hepatobiliary routes, and both can be modulated to some extent with chemical modification. Hydrophilic peptides are usually excreted unchanged by the kidneys, and more lipophilic peptides have transient albumin binding and are ultimately cleared by the hepatobiliary system. These two common clearance profiles allow easy extrapolation of micro-dosing preclinical data to first-in-human dosimetry, allowing for easy prediction of organ radiation doses.
Selective modification of a lysine or cysteine side chain leaves a handle for bifunctional chelators (e.g. DOTA, NOTA or HYNIC) tailored to complex different radiometals under biologic conditions. Spacer units of PEG repeats, β-amino acids or rigid cyclic moieties separate the metal from the binding epitope to reduce steric occlusion and influence renal clearance. In this plug-and-play strategy, the development of an optimized targeting moiety is not limited by the choice of payload; affinity maturation can be decoupled from radiochemistry. Moreover, the same precursor can be labeled with different isotopes at the final stage, reducing GMP manufacturing and allowing a single active pharmaceutical ingredient (API) for both diagnostic and therapeutic applications. As the conditions required for radiometal complexation are mild (low pH and room temperature), the secondary structure of peptides is maintained, allowing diagnostic and therapeutic isotopes to be introduced sequentially or in parallel without affecting the pharmacophore. Orthogonal functional handles, such as azide-alkyne click chemistry, can be further exploited to install a fluorophore for intra-operative guidance in parallel with a therapeutic radionuclide, resulting in triple-function vectors spanning PET, fluorescence and radiotherapy. This chemical promiscuity allows targeting logic to be maintained while the payload suite is expanded, and preserves the "see what you treat" principle across multiple modalities, streamlining clinical translation.
Dynamic binding permits sampling of multiple copies of the receptor per cell during a single circulation time. Fast on-rates with moderate off-rates allow the ligand to unbind and diffuse further into tumor parenchyma, then rebind at downstream receptors, spreading the diagnostic or therapeutic payload throughout the lesion rather than concentrating it at the periphery. This behavior is particularly desirable for therapeutic isotopes whose cytotoxic range exceeds the originating cell; homogenous distribution both maximizes the bystander effect and minimizes toxicity to adjacent normal tissue. Repeatable targeting behavior also means that once the first imaging or therapy cycle has been completed and the peptide has cleared, the tumor is not permanently altered and further administrations can be given with the same pharmacokinetics. In contrast to irreversible covalent inhibitors that change the antigen and prevent further imaging, reversible peptide binding also permits longitudinal response monitoring and repeat dosing without accumulating off-target toxicity, a prerequisite for fractionated radioligand therapy.
Dual affinity and kinetic properties are required of a theranostic peptide. Peptides need fast clearance to allow high-contrast imaging and, conversely, retention for therapeutic effects. In addition, a theranostic peptide requires that the targeting geometry be maintained between the diagnostic and therapeutic versions of the peptide. These factors place limitations on design. The peptide must bind to its target with nanomolar affinity, independent of whether a diagnostic or therapeutic payload is attached. The peptide must be resistant to proteolysis for a time long enough to have therapeutic effect and be cleared quickly enough to be useful for imaging. In addition, it should be resistant to chemical modifications necessary for radiolabeling or drug attachment. These factors can be considered by careful placement of attachment sites distant from the binding epitope, incorporation of an appropriate linker length and flexibility and amino-acid sequence to confer metabolic stability, while remaining amenable to renal clearance.
Imaging-labeled peptides and drug-conjugated peptides risk functional divergence if the same attachment site leads to chemical competition; a chelator attached to the side-chain of a lysine or cysteine residue can mask the binding epitope through steric hindrance or charge reversal, leading to therapeutic conjugates with no affinity and tracer formats that are intact. To eliminate functional divergence, the chelator is often placed at the N- or C-terminus distal from the receptor-binding site, and separated by a rigid hydrophilic spacer. All conjugates, whether modified with a fluorophore, DOTA, or a cytotoxic warhead, should be verified with identical receptor-binding assays to ensure that diagnostic and therapeutic formats diverge by no more than twofold in Kd and internalization rates.
Kinetic conflicts arise because, for a long-lived therapeutic radioisotope like 177Lu (6.7 days), retention in tumors for days to weeks is required to deliver a killing dose, whereas the retention required for imaging over hours would create a high background that masks small lesions. Modulation of residence time makes use of a hybrid approach: the diagnostic formulation loads each peptide with a short-lived radioisotope for rapid clearance; the therapeutic formulation loads a long-lived radioisotope and adds an albumin-binding tag to lengthen circulation time by a factor of a few. In both cases the peptide itself and its tumor affinity are the same; this is required for quantitative correlation of the diagnostic images with the therapeutic dose.
The conjugation of a diagnostic radionuclide as well as of a therapeutic payload onto the same peptide backbone can introduce strain - chelators themselves can distort the backbone into a β-turn conformation, and larger prosthetic groups can sterically clash with the binding epitope. To accommodate this, peptides are often cyclized or stapled to fix the pharmacophore in a protease-resistant conformation that can also tolerate N-terminal chelation without unfolding. In addition, rigid spacers (such as proline-alanine hinges) can be inserted between the binding site and the chelator to eliminate payload-induced allosteric effects. Stability under both payload conditions is established by circular dichroism and ion-mobility MS, which verifies that the peptide's secondary structure does not change after each modification and that diagnostic and therapeutic formats are identical in targeting.
Although peptide theranostics offer the ideal direct translation from diagnostic to therapeutic development, it is important to recognize that development is not always straightforward and hidden complexities of a single molecule achieving both functions often complicate this approach and lead to clinical failure. The fastest pharmacokinetics is required of an imaging tracer to ensure a high ratio of target to non-target binding. However, longer circulation for high payload delivery is required for therapy. Hidden fragilities can be unveiled in the kinetic tension between achieving a theranostic that has optimized pharmacokinetics and pharmacodynamics for both imaging and therapy. These design frictions include small changes in payload size, charge or linker chemistry that can result in off-target binding or altered biodistribution or may inadvertently lead to increased metabolic stability or rapid in vivo deconjugation. This often transforms a valuable compound series into therapeutic failure. Additionally, the additional burden for dual optimization increases development costs, animal use and prolongs time to translation as each optimization step needs to be proven in both the imaging and therapy model. The following sections will highlight the most common and problematic pitfalls, including performance mismatch in the different conjugates, signal and therapeutic effect decoupling and the increased complexity of designing a single molecule that performs two different functions.
The other, less desirable example is if a diagnostic labelled with a positron-emitting radionuclide with a short half-life yields high-contrast tumor uptake in preclinical models, but the same targeting moiety attached to a therapeutic beta emitter demonstrates low lesion retention. In these cases the alterations in the chelator may induce a change in the conformation of the bioconjugate. This is especially relevant if the macrocyclic chelator cage to which the radionuclide binds necessitates a more rigid structure in the chelator-linker region in order to coordinate the radiometal. This may have an effect on binding affinity as well as the kinetics of internalization. Furthermore, each individual radionuclide has a decay process that can produce secondary fragments. For example, while positron emitters may be more benign, the radiolysis caused by the alpha particle decay cleaves sensitive peptide bonds, thus fragmenting the chelator-peptide conjugate prior to tumor localization. For these reasons it is important to not only confirm binding in preclinical studies but to also assess stability of the complex in serum and tumor homogenates with the final radiometal.
One of the most pernicious consequences is the absence of imaging signal without therapeutic effect, for example where those residualizing metabolites get sequestered in kidneys or liver, which produce hot spots that look like tumor uptake. This can happen when a peptide containing a labile bond is cleaved, releasing a radiometal-chelate fragment that is no longer bound to the targeting motif, and which is small enough to be subject to glomerular filtration but sufficiently residualizing to be reabsorbed by proximal tubules. This concentrates radiation in a radiosensitive organ but without delivering any cytotoxic payload to the lesion, meaning the PET image grossly overestimates the distribution of the therapeutic dose and so leads to subsequent fractions being under-prescribed by the clinician. The other common disconnect is therapeutic delivery without reliable imaging readout, which can happen when the conjugate is too tightly bound and an affinity sink is created at the tumor rim that blocks deeper penetration. This image is thus a bright rim but with a cold core, but the therapeutic isotope will in theory be subject to bystander effect and will reach cells in the core region, disconnecting the visualized dose coverage from the actual absorbed dose. In this case, one is forced to use a mathematical model rather than direct imaging as a correlate of absorbed dose, again defying the theranostic ideal of real-time treatment monitoring. Cleavable linkers and tuning the affinity down are two design strategies to prevent these scenarios from occurring.
The workload for theranostic iteration is greater because design decisions are double checked for their effect on both imaging and therapy: any decision to decrease charge to decrease renal uptake, must be counter-checked for effects on therapeutic residence time. A peptide with optimized biodistribution for 68Ga-PET, will often fail when the same peptide is radiolabelled with 177Lu and in vivo performance is determined. The peptide requires cycles of synthesis-radiolabelling-in vivo biodistribution-validation both for the imaging and therapeutic products, essentially doubling preclinical requirements. When multiple isotopes, spacers, and stability motifs are added into the parameter mix, combinatorial possibilities explode and high-throughput systems are necessary to explore them, at significant additional cost. Predictability of performance is hampered by the fact that tiny differences in cyclisation efficiency or N-terminal acetylation from batch to batch, can tip the balance from imaging efficacy to therapeutic inefficacy. Purity specifications, while usually simple and clear for small molecules, become less so for peptides. Mass spectrometry will address sequence fidelity, HPLC might be necessary to assess diastereomeric purity, and radiochemical stability will provide information on metallation robustness. Expectations for high batch-to-batch consistency would also require process controls that are too stringent for many academic synthetic groups, and the gap between a lead and a clinical candidate becomes unmanageable.
The overarching design principle that should guide peptide theranostic candidates towards bench-to-bedside success is that the diagnostic and therapeutic format should have identical targeting properties. In the absence of this property, the diagnostic signal (tracer uptake), used as a surrogate for receptor occupancy, poorly predicts the in vivo distribution of the therapeutic payload. This imperfect translation between diagnosis and therapy invariably leads to suboptimal dosing decisions and, ultimately, to suboptimal patient outcomes. This goal can be achieved using a variety of levers. The first approach uses imaging as a quantitative tool to predict therapeutic drug exposure, the second approach is to engineer modular platforms that can be dual functionalized without drifting from the molecular logic that ensured efficient targeting, and the last approach is to identify targets whose expression is not affected by the disease condition. Notably, these design criteria should be applied at the lead optimization phase rather than applied in a re-engineering step during the scale-up process.
This relies on the assumption that SUV obtained by PET accurately forecasts the absorbed radiation dose (expressed in gray units) that will be delivered to tumor cells by the therapeutic analog. To ensure that this is the case, a developer must perform serial imaging at multiple time points after administration of a micro-dose of the diagnostic tracer to generate time-activity curves that accurately model the clearance of the radionuclide, turnover of the target receptor, and perfusion within the lesion. Kinetic values are then input into dosimetry software that will model the absorbed dose per unit of activity administered (MBq). Clinicians can then determine the activity they need to administer in order to deliver a tumoricidal dose to a tumor while staying below the tolerance dose of normal-organ compartments. This quantitative assessment of lesion burden allows the imaging readout to go beyond a "hot-or-not" paradigm and instead become a treatment-planning tool that can be used to prescribe dosing to each patient, thus eliminating the practice of population-based dosing and the under-treatment or unacceptable toxicity that often results. First-line validation requires showing that the diagnostic and therapeutic versions have identical pharmacokinetics. This is demonstrated by labeling the same batch of a peptide with both a short-lived PET radionuclide and a microdose of the therapeutic radionuclide. The two tracers are then co-injected into an animal model, and the biodistributions are demonstrated to be concordant in tumor, kidney, liver, and marrow as well as other organ systems that express the target receptor. The uptake is found to be superimposable (within error of the assay) in each organ, thus showing that the diagnostic signal accurately represents the therapeutic delivery. Should any differences arise between the two curves, work to modify spacer length or chelator position must be repeated until both probes yield identical biodistributions.
Modular platforms are based on a conserved recognition loop (most commonly a cyclized or stapled motif which determines the selectivity for a given receptor) and bookending standardized attachment points for chelators, linkers, and warheads. This approach ensures that the pharmacophore is the same across payloads, making it certain that they will have the same receptor engagement and kinetics across diagnostic and therapeutic formats. In a common architecture a bicyclic peptide containing a linear receptor-binding domain between cysteine handles can be labeled with a 68Ga-DOTA on one arm and a 177Lu-DOTA on the other in one construct, creating a single bifunctional construct capable of both imaging and therapy in a 1:1 ratio. This sharing of a scaffold between diagnostic and therapeutic platforms can make regulatory approval easier, since the same peptide master file is used for both, and also simplifies production since one precursor material can be used for both the imaging and therapy batches, simplifying analytics and reducing the batch-to-batch variation that comes from introducing more complexity to one of the two production processes. Flexible architectures, on the other hand, are based around linkers (PEGylated or β-amino acid spacers) that separate the dynamics of the payload from the binding motif, such that each radionuclide or drug can be conjugated without perturbing the recognition fold. This design can be useful to account for the relative bulkiness of a therapeutic warhead, without sacrificing the rapid pharmacokinetics necessary for imaging. In such linkers, conditional release mechanisms can also be incorporated, such as pH-sensitive linkers that are stable for imaging but cleave upon reaching the more acidic tumor microenvironment, or enzyme-cleavable motifs that are only activated following cathepsin-mediated processing.
The target should be expressed uniformly from the time of diagnosis through disease progression and during all cycles of treatment. If expression is lost due to the treatment or progression, then the imaging agent is no longer valid and the therapy agent no longer has an address. Ideally the target is a lineage-marker antigen that is under the control of the oncogenic program rather than a transient stress response, such that it is still expressed under conditions of hypoxia and/or under therapy-induced selective pressure. Examples include prostate-specific membrane antigen (PSMA) or somatostatin receptor subtype 2, which can be expressed by several different tumor types that metastasize to specific locations and are retained through multiple lines of treatment. Expression should be determined across many patients and should be confirmed by longitudinal biopsies during therapy to show that the expression does not fade over time. It is a failure mode to have a lesion that is bright at the time of diagnosis and therefore amenable to therapy, only to have it fade by the time the radioactive component is available. The target epitope should be accessible on the cell surface of both early, well-perfused lesions as well as larger and/or fibrotic or necrotic tumors with reduced vascular permeability. For example, peptides that must be actively transported across the cell membrane or are dependent on proteolytic processing will not work for the hypoxic core of the lesion where these functions are downregulated.
Table 2 Theranostic Peptide Design Trade-Offs
| Design Parameter | Imaging Requirement | Therapy Requirement | Unified Strategy |
| Clearance rate | Rapid (renal) | Moderate (tumor retention) | pH-sensitive off-rate |
| Payload bulk | Small (hydrophilic) | Bulky (lipophilic) | Distal chelator placement |
| Backbone stability | Flexible (binding) | Rigid (stress tolerance) | Cyclisation + protease-resistant motifs |
| Linker chemistry | Stable in plasma | Cleavable in lysosome | Self-immolative spacer |
| Dosimetry | Qualitative uptake | Quantitative absorbed dose | Identical biodistribution |
Theranostic development requires more than pairing an imaging agent with a therapeutic payload. Success depends on whether both functions share a truly unified targeting behavior across different molecular formats and exposure conditions. Our targeting peptide services are designed to support theranostic programs by aligning peptide design, functional performance, and biological context from the earliest stages. Rather than optimizing imaging and therapeutic conjugates independently, we focus on preserving consistent targeting logic as peptides transition from diagnostic to therapeutic applications.
Unified Targeting Logic: A common failure point in theranostic programs is assuming that a peptide that performs well as an imaging probe will behave similarly once conjugated to a therapeutic payload. Our theranostic-oriented design approach begins with defining a unified targeting logic that governs both diagnostic and therapeutic formats. Peptide sequences are engineered to tolerate different conjugation chemistries and payload sizes while maintaining consistent target engagement. Binding affinity, kinetics, and structural stability are optimized within a functional window that supports both rapid imaging clearance and sufficient therapeutic exposure, avoiding designs that favor one application at the expense of the other.
Imaging-Therapy Alignment Workflows: To ensure alignment between imaging and therapy, we apply imaging-therapy alignment workflows that evaluate how peptide behavior changes across conjugate formats. This includes assessing differences in biodistribution, target retention, and non-specific uptake between imaging-labeled and drug-conjugated peptides. By identifying and correcting divergence early, we help ensure that imaging readouts remain predictive of therapeutic delivery, strengthening the foundational premise of theranostics.
Early Divergence Risk Identification: Not all targets that are suitable for imaging are equally suitable for therapeutic delivery, and vice versa. Our feasibility-first assessment focuses on identifying early divergence risks, such as targets with favorable imaging contrast but insufficient capacity for therapeutic accumulation. This early evaluation helps determine whether a shared targeting peptide can realistically support both diagnostic and therapeutic objectives, reducing the risk of advancing misaligned theranostic strategies.
Design Consistency Evaluation: Theranostic success depends on maintaining design consistency across multiple peptide constructs. We evaluate whether a peptide scaffold can support repeated functionalization without compromising stability, binding behavior, or selectivity. This design consistency evaluation informs whether a single peptide platform can serve as a reliable backbone for theranostic development or whether alternative strategies should be considered.
If your theranostic program shows strong imaging performance but inconsistent therapeutic outcomes—or if you are planning to translate a diagnostic peptide into a therapeutic format—an early technical discussion can help clarify the underlying constraints. Discuss your theranostic targeting strategy with our scientists to assess target feasibility, identify divergence risks, and define a peptide design approach that aligns imaging and therapy from the start.
References
