The peptide–drug conjugate (PDC) is less a molecule than a molecular stage play, the set of which is the linker. The functional groups decorating the peptide backbone, payload warhead, and linker spacer are the dramatis personae whose chemical reactivity, steric mood, and electronic tone determine whether the play's therapeutic climax will be on-target release or silent systemic quenching. The linker must therefore be viewed as a programmable, solvent-excluded micro-environment, one that converts extracellular stability into intracellular freedom via the precisely timed rupture of otherwise physiologically unremarkable bonds. Fluency in orthogonal ligation handles, pH-rescued cleavage kinetics, and secondary interactions that can either mask or flag the construct for enzymatic editing are therefore required to master this environment. In short, functional group choice is the language through which medicinal chemists whisper instructions to the pharmacological future of the patient.
Fig. 1 Basic structure of PDC.1,5
Linker chemistry is the keystone of a PDC's risk–benefit profile. Too stable, and it will deliver the payload unscathed through the circulatory system to the endosome, where it will be trapped and converted to costly, inactive metabolites. Too labile, and it will release its burden next to healthy vessels, reviving the dose-limiting toxicity that the peptide had been developed to overcome. The design space between these two extremes is narrow, and is defined by adjusting the hybridization of heteroatoms, ring strain of pre-activated electrophiles, and redox potential of protected thiols so that it is only cleaved in the presence of an intracellular handle (acidic pH, reducing environment, or tissue-specific protease). In this way, linker chemistry is not a simple tether, but an encoded biosensor, whose key is written in the metabolic language of the target cell.
Fig. 2 Different types of polypeptide-drug linkers commonly used in the rational design of poly peptide-drug conjugates.2,5
The idea of using a "cleavable linker" is analogous to the working of an hourglass, where the "sand" is the linked payload. It flows only when an external stimulus tips the balance between release and non-release in favor of release, as with hydrazone bonds that hydrolyze slowly at physiological pH, but once in the low-pH milieu of the late endosome (which they enter like other nanoparticles) undergo accelerated hydrolysis due to the protonation gradient across the organelle membrane. On the other hand, disulfides "listen" to the high level of glutathione in the cytosol to catalyze the breakage of the disulfide linkage (i.e. this small molecule, which exists in reduced form inside the cell, acts as the stimulus). The same disulfide bond remains stable in oxidizing extracellular environment, and thus the payload is not prematurely released while it is circulating in the blood. An alternative is "non-cleavable linkers", where the whole linker-peptide conjugate is digested inside the lysosome and it is hoped that the remaining small molecule-amino acid conjugate will still have sufficient activity to produce the desired therapeutic effect. In other words, it bets on complete metabolism of the protein moiety rather than on a cleavable linker; on the other hand, if the linker–drug conjugate does not work as well as the drug alone, one has to redesign the drug, rather than the linker, which is a much more complex and difficult proposition. The theoretical appeal of both strategies is a factor that influences the decision in the preclinical phase, where the difference between the two may determine the success or failure of an otherwise well-designed ADC, and where the exact mechanisms of proteolysis inside the target cell may be poorly understood.
Traceless release ideally releases the payload in its native form, without any chemical trace of the conjugation. Self-immolative spacers can provide such an effect by covalently entrapping the drug as a carbamate or carbonate that is activated by an enzymatic step prior to release, so that upon cleavage of the initiating bond, the spacer unzips through a series of cyclisation or 1,6-elimination events which eject a CO2 or a neutral heterocycle and regenerate the original, unmodified hydroxyl or amine of the free drug. The beauty of the strategy is that the entire spacer should be inert to premature cleavage (so long as the initiating bond remains), but that once that first bond is cleaved the whole structure should undergo rapid self-destruction before other hydrolytic events become significant. The kinetic compression is most appreciated when the payload is a steroid or kinase inhibitor, where a single structural appendage can be enough to utterly compromise the activity. The same electronic perturbations that make for rapid self-destruction, however, can also make the spacer vulnerable to nonspecific nucleophiles (plasma amines and thiols, for example), so the choice of substituents will also include electron-withdrawing groups that increase the activation energy just enough to reach the target without being destroyed by the plasma, but not so much that intracellular enzymes can't activate the linker. Optimization of these orthogonal rate constants is more art than science, and relies on an iterative process of medicinal insights and analytical oversight until the linker finally disappears, untraceably.
Functional groups are the chemical language through which peptides and payloads communicate to form a conjugate. All handles have their own steric signature, nucleophilic reactivity and pH-dependency which dictates the rate and site of ligation, as well as the ultimate in vivo stability and immunogenicity of the conjugate. The conjugation process thus starts with a word choice, i.e. which functional alphabet will be encoded on the peptide backbone so that it may covalently 'speak' to the drug without getting the message 'lost in translation' (hydrolyzed) along the way. We survey three of the most fluent functional group alphabets currently in use medicinally.
Amine and carboxyl couplings represent the longest chapter of bioconjugation syntax, while remaining among the most flexible. The primary aliphatic nucleophilicity of lysine ε-amines is poised for reaction with activated esters, isothiocyanates, or carbamoyl chlorides at physiological pH, to give strong urethane or amide bonds, whose hydrolytic half-lives can greatly surpass the expected circulation half-life of the conjugate. The reaction is quite tolerant to small pH excursions but can proceed quickly and under mild conditions to not disturb the folding of disulfide-rich peptides. In most cases, carboxylic acids are found on the C-terminus of the peptide, or on the side chains of aspartic acid or glutamic acid residues, and are activated prior to coupling. This is often done with a water-soluble carbodiimide which catalyzes an O-to-N acyl shift when an external amine is present. Because this reaction is reversible, the stoichiometry can be pushed to completion by temporarily reducing the dielectric constant of the solvent with a cosolvent that will not disrupt the peptide secondary structure. There can be a concern for cross-reactivity when both amine and carboxylate groups are present on the same macromolecule; in such cases, regioselectivity is imposed by orthogonal protection/deprotection chemistry, wherein the deprotection conditions for the protecting groups on one of the groups are chosen to not disturb the newly formed amide bond. One recent refinement has been the development of picolyl esters that self-destruct following intracellular reduction, thereby converting a stable linkage into a trigger for traceless release of the payload once the conjugate has been taken up into the cytosol. In this manner, the amine–carboxyl pair continues to be developed from a simple condensation reaction into a context-sensitive on–off switch.
Thiols bring a redox-responsive language that is orthogonal to the amines-carboxylic acids one. The sulfhydryl of cysteine reacts with maleimides, bromoacetamides, or pyridyl disulfides under nucleophilic conditions at pH when lysine is still protonated, leading to the formation of thioether or disulfide bonds, whose reversibility is dictated by the redox environment. The thioether bond is essentially irreversible in extracellular conditions, which is convenient when a permanent conjugation is needed, but can be a limitation when a payload release in the cell is required. Disulfides, on the other hand, are stable in the oxidizing environment of the plasma, but rapidly cleaved in the glutathione-enriched cytosol, providing an intrinsic release mechanism without further synthetic burden. The thiol handle also allows site-specific conjugation, as cysteine is less abundant than lysine, and mutating a single surface-exposed residue to cysteine will provide a unique conjugation site. The difficulty here is when the peptide backbone already features native disulfides, and strategies to ensure no scrambling by using pairing algorithms and transiently protecting disulfides are required. More recently, cyclic disulfide mimics have been used, which are recognized by the endogenous thiol/disulfide system, but features a strain on the ring, which favors a faster cleavage in the cytosol, while remaining kinetically stable in circulation. This way, sulfur chemistry affords a redox-responsive vocabulary that can be leveraged within the overall conjugation story.
Click chemistry provides a panel of pairwise orthogonal reactions that are fast, selective and biocompatible such that conjugation can be achieved in situ in living systems without interfering with the native biochemical vernacular. The copper-catalysed azide–alkyne cycloaddition remains the prototypical example, generating a 1,2,3-triazole linkage that is metabolically inert and thermodynamically stable; but the residual copper can drive oxidative stress, motivating an emerging transition to copper-free alternatives like the strain-promoted azide–alkyne cycloaddition (SPAAC). In this reaction, pre-distortion of the alkyne through a cyclooctyne ring tail accelerates the reaction to obviate the need for metal catalysis, albeit at the cost of a small hydrophobic burden. Tetrazine–alkene inverse-electron-demand Diels–Alder (IEDDA) reactions extend the rate constant frontier still further to allow conjugations that go to completion within seconds at micromolar concentrations, a kinetic edge that can be decisive when labelling internalizing peptides in a competitive environment. The real strength of click chemistry comes from its orthogonality: azides, alkynes, tetrazines and alkenes are virtually non-existent in mammalian biochemistry, and the conjugation can therefore occur against a silent backdrop. This property permits a sequential double-click strategy in which one handle is used to attach a payload and the second to append a fluorophore or solubilizing polymer, all without the need for protecting groups or chromatographic interruption. As the bioorthogonal palette broadens to include nitrile-oxides, sydnones, and quadricyclanes, the peptide chemist gains an ever richer vocabulary for scripting conjugations that are fast, clean, and stealthy.
In practice, the decision between cleavable and non-cleavable linkers is not a switch but a matter of risk management. Cleavable linker designs require faith in the pathophysiological micro-environment: that acidic pH, the reducing cytosol or localized proteolysis will come to play at a time when the payload can still be heard to sing. Non-cleavable linker designs, in contrast, place all their faith in the lysosome's ability to chop up the peptide carrier into innocuous amino-acid debris, and in the lysosome's ability to leave the warhead–linker remnant inactive and uncontaminated by any off-target activity. Somewhere in between, designs exist that are potentially cleavable but relatively slow: a best-of-both-worlds solution with a nod to worst-case scenario crosstalk. The next two sections summarise the strategic considerations behind these two main approaches, and how chemical design follows from the decision to allow or to forbid bio-interference.
Acid-labile linkers such as hydrazones, acetal ethers, and β-thiopropionate esters contain a hidden electrophilicity that is kinetically inert at physiological pH but becomes reactive once the conjugate is delivered to the increasingly acidic endosomal pathway. The challenge is to adjust the pKa such that cleavage is essentially quantitative before lysosomal proteases can attack the peptide linker; otherwise, the payload is released into a digestive environment from which it will not escape. Redox-sensitive disulfides and selenides count a different currency, being sensitive to the cytosolic glutathione excess as a nucleophilic stimulus for disassociation. Since the extracellular space is oxidizing, the bond is preserved during blood circulation, but once the conjugate reaches the reducing cytosol, the linker unzips at a faster rate than competing esterases can hydrolyze the payload, thus ensuring the release of the drug in its full active form. A new generation of cyclized disulfides further reduces the reduction threshold by introducing ring strain which allows cleavage to take place even in tumors where glutathione is only mildly overexpressed. The beauty of these architectures is their capacity to delegate the trigger to endogenous biochemistry, thus removing the need for exogenous activators while still retaining a high spatial resolution between blood and cytosol.
Non-cleavable linkers give up on environmental triggers for metabolic determinism. The hope is that the whole package is shunted off to the lysosome, whereupon the cathepsins will chop the peptide backbone into amino-acid–drug conjugates that still have pharmacologic expression. Amide and thioether linkages are preferred because they are resilient both to hydrolysis and disulfide exchange; they do not come undone even if the conjugate is sequestered for long periods in recycling endosomes. The unstated (but critically important) assumption is that the remaining linker–drug remnant is small and polar enough to leak back across the lysosomal membrane rather than getting trapped in subcellular store sites. If that assumption is incorrect, the effect is a stealth intoxication: the payload gets sequestered in lysosomes where it slowly compromises the organelle's membrane integrity and provokes cathepsin-initiated apoptosis in off-target cells. To guard against this, a miniaturized self-immolative feature is being added by more designers, and revealed only after the peptide is proteolytically clipped.
Responsive linkers are moving away from simple-input switches to multiple gated circuits that respond to the combined influence of pH, redox, and enzyme stimuli in Boolean logic operations. The current goal is to incorporate a rudimentary logic gate into the spacer such that release upon cleavage occurs only when a specified combination of environmental signals is present (e.g. both acidic pH and high matrix metalloproteinase), which would limit the likelihood of premature drug release. Such an advancement necessitates molecular subunits that are capable of reversibly shielding electrophiles, sequestering a built-in ring strain, or tuning their own electron transport properties in response to minute biochemical cues. Two motifs that meet this criterion are discussed in the following sections: tandem-cleavable linkers that are dependent on sequential triggers, and photo- or ultrasound-responsive spacers that relinquish temporal control to external physicians.
Dual-responsive linkers graft two chemically orthogonal labile bonds into a single scaffold, arranging them so that the rupture of the first bond unveils a nucleophilic handle that accelerates scission of the second. A typical embodiment couples an acid-labile hydrazone with a downstream disulfide; endosomal acidification initiates the first cut, exposing a thiol whose intramolecular attack snaps the disulfide within minutes, liberating the drug before the construct reaches lysosomes. The kinetic beauty lies in the temporal amplification: the initial trigger operates at a relatively mild pH threshold, yet the downstream redox cleavage proceeds faster than physiologic disulfide exchange, ensuring that payload release is completed while the conjugate is still sorting through early endosomes. Sequential-cleavage strategies extend this concept into enzymatic space, embedding a peptide sequence recognized by tumor-associated proteases upstream of a self-immolative carbonate. Only after the protease has clipped the recognition motif does the carbonate become electrophilic enough to undergo 1,6-elimination, releasing carbon dioxide and the unmodified drug. The dual requirement for both proteolytic and chemical activation collapses the release probability in healthy tissue where either cue may be present individually but rarely together, thereby sharpening therapeutic specificity without invoking exotic chemistry.
An obvious advantage of external energy modalities is their spatial precision relative to endogenous triggers. Ortho-nitrobenzyl carbamates, which absorb UV or visible light, undergo rapid photoisomerization with concomitant expulsion of carbon dioxide and the free drug on the nanosecond timescale after illumination. The chromophore can be red-shifted into the tissue-transparent window by the addition of electron-donating methoxy substituents, allowing activation through centimeters of skin or mucosa. Ultrasound-sensitive linkers function through an entirely different physical mechanism: a perfluorocarbon microcavity embedded within the spacer undergoes vaporization by acoustic cavitation, with attendant mechanical shear forces sufficient to cleave an adjacent carbonate or boronate ester. Because ultrasound can be focused to millimeter resolution, drug release can be restricted to a single tumor lobule without affecting neighboring parenchyma. The clinical appeal is the ability to titrate dose in real time: the oncologist can raise acoustic intensity until imaging reveals therapeutic saturation, and then terminate illumination to avoid overdosing. Present obstacles to clinical translation include the oxygen-dependent quenching of photo-excited states and the serum protein shielding of microcavities, yet iterative chemical shielding and cross-linking of perfluorocarbons is gradually pushing the activation threshold into clinically accessible parameters.
The choice of linker has a very big effect on pharmacokinetic/pharmacodynamic (PK/PD) properties. It often goes unnoticed during the early stages of discovery, as linker hydrolysis and payload release takes place after target engagement. However, it should be kept in mind that linker chemistry is the silent force that will ultimately turn the PK exposure into the observed PD effect, and will also determine the shape of the toxicological footnotes the regulator will be poring over. A too-labile linker may increase the maximum tolerated dose (MTD) through payload accumulation in off-target tissues, and a too-stable linker may result in excellent plasma stability but no intracellular drug activity, resulting in attempts to dose-escalate to therapeutic levels which defeats the entire purpose of targeted drug delivery. The pharmacological properties of the spacer, then, determine both the ceiling of the efficacy curve and the gradient of the dose-limiting toxicity profile, thereby narrowing or expanding the therapeutic window. This is also the case because the linker often has more effects downstream of the target, and its impact is therefore often not noticed in the early screening assays that only measure target affinity but emerges once you get into in vivo studies when there is an enormous cost attached to redesign. In the following two sections, we discuss two areas in which linker choices have the most impact: (i) the PK–PD profile of the drug, and (ii) the long-term immunological effects of linker metabolites.
The linker also alters the apparent volume of distribution and the apparent potency, both of which determine the exposure–response surface. If a cleavable ester is cleaved in the plasma before reaching the target tissue, the central compartment concentration of the free drug will increase. This will cause an early peak in Cmax and may saturate receptors, while still potentially damaging healthy tissue. If a linker is cleaved after deep penetration has occurred, the time of appearance of the active drug will be delayed, shifting the PD curve to the right and increasing the necessary loading dose to maintain the same AUC at the site of action. All of these temporal shifts can be further complicated if the linker–drug fragment itself has partial agonism or antagonism. This off-target activity is more likely to be seen if the PK sampling reflects total drug levels and not the linker-modified species. This "phantom" activity can be interpreted as target resistance leading to unnecessary dose escalation until off-target toxicities become apparent. To combat this, the linker must be designed rationally, with an integrated model that links cleavage kinetics and spatial drug distribution to maximize efficacy.
In addition, when the linker is eventually removed, the short-lived conjugate surface can become decorated with neo-epitopes that can violate immune tolerance. Maleimide-thioether adducts are subject to retro-Michael reactions in vivo, which release a protein-bound maleamic acid whose haptenic capability has been associated with the formation of anti-drug antibodies that mediate accelerated blood clearance on repeat dosing. The retro-haptenization products of ortho-nitrobenzyl photolysis, which can be oxidized in vivo to form a reactive quinoneimine, have also been shown to be capable of modifying serum albumin and triggering delayed hypersensitivity. These risks may not be apparent in routine toxicology panels, as the conjugate is given acutely to immunologically naïve animals; it is only after chronic dosing in higher species that the immune system has a chance to mount a significant response. Lysosomal accumulation of linker metabolites can also be a concern, as the hydrophobicity of these species can perturb lipid trafficking and cause phospholipidosis, a chronic vacuolation that is histologically subtle but functionally corrosive to long-term organ function. For these reasons, modern linker optimization also often includes in-silico alerts for reactive metabolite formation, in-vitro hapten assays, and extended immunogenicity studies that acknowledge the insufficiency of the conventional three-month window, as the silent chemistry of the spacer can speak loudly once the patient's immune memory has been engaged.
We design and optimize cleavable and non-cleavable linkers that ensure payload release with precision. Our scientists work with disulfide bonds, click chemistry, and bioorthogonal methods to create linkers that maximize therapeutic effect while maintaining safety. Partner with us for advanced linker and conjugation solutions.
1. Why is linker chemistry critical for PDC performance?
The linker determines when and where the drug payload is released, directly affecting therapeutic efficacy and safety.
2. What functional groups are used in peptide conjugation?
Amine, carboxyl, thiol, and hydroxyl groups are commonly used to attach payloads via covalent bonds.
3. What is the difference between cleavable and non-cleavable linkers?
Cleavable linkers release drugs inside target cells, while non-cleavable linkers keep the payload attached until peptide degradation.
4. How do disulfide linkers function in PDCs?
Disulfide linkers break in reductive environments like tumor cells, ensuring payload release at the disease site.
5. What is click chemistry in linker design?
Click chemistry refers to bioorthogonal reactions that provide fast, specific, and efficient conjugation between peptides and drugs.
6. How does linker choice affect safety?
Poor linker design may cause premature release, leading to systemic toxicity. Optimized linkers improve precision and reduce side effects.
References
