Glycosylation is a key determinant in the in vivo fate of therapeutic proteins. Glycan structures code for protein properties such as clearance, protease resistance and size and all of these affect pharmacokinetics (PK) and exposure of the protein to target tissues. Clearance receptors recognize and bind to specific sugar sequences, which affect elimination rates. Proteolytic degradation is minimized by glycosylation, thereby also increasing residence time. Size is affected as well as glycosylation increases molecular hydrodynamic volume, which, in turn, is often correlated with clearance. At the same time, the protein sequence influences the resulting glycosylation pattern, i.e. the type of glycans attached, which means that glycosylation is a function of the protein as a whole. This is evident in the case of antibodies as well as other proteins such as cytokines where differences in glycosylation can result in major changes in clearance, distribution, and immunogenicity, which in turn, affects clinical efficacy. Glycosylation of therapeutic proteins is thus a key attribute of drug substance and a critical determinant of clinical performance and must be well controlled and understood to design next generation of biopharmaceuticals with improved PK profiles, reduced dosing frequency, and/or tissue specific targeting.
Glycosylation is one of the first steps in the metabolism and elimination of a drug. Glycosylation modification can have a major impact on pharmacokinetics, which is the process by which the drug concentration changes in the body over time. The changes in the concentration of the drug in the plasma and tissue over time is a result of multiple rate determining events, all of which may be affected by glycosylation. For example, terminal sialic acid, the "shield" against hepatic uptake by the asialoglycoprotein receptor, is also the determinant of isoelectric point (pI) and the electrostatic affinity of a protein for capillary endothelium and lymphatic drainage. The high mannose motifs that facilitate targeting of therapeutic proteins to the mannose receptor and degradation can also modulate the solubility and aggregation of the protein and thus the filtration at the glomerular endothelium. As a result, glyco-engineering to optimize the desired pharmacokinetic behavior may need to account for possible concomitant changes in stability, aggregation, and immunogenicity. As glycosylation analysis tools are being used to inform and even directly integrate into pharmacokinetic models, a growing ability to anticipate the influence of particular sugar structures on the duration of systemic exposure of biologics will allow for the rational design of drugs with half-lives more appropriately matched to their use.
Pharmacokinetics (PK) can be defined as the study of what the body does to a drug in terms of movement through the body. PK can be broken down into four processes: absorption, distribution, metabolism, and excretion. For protein drugs, these PK processes are different from small molecule drugs. Proteins are most often delivered intravenously (IV) or subcutaneously (SC) since absorption through the gastrointestinal tract is an important limitation for oral administration. Only small amounts of macromolecules can cross cell membranes through absorption in the GI tract. For the distribution of protein drugs, they are either present in the vasculature, the interstitium, or inside cells. Movement from the vasculature to the interstitium or cells is limited by several factors. Large proteins, for example, are not able to be filtered through the kidney and are dependent on the lymphatic system for their movement into the interstitium. Protein metabolism for protein drugs is not by the liver but rather by peptidases that degrade the drug into smaller peptides and amino acids. Elimination of peptides and amino acids is renal or by endocytosis. The elimination half-life for protein drugs can range from hours to weeks and is the limiting factor in how frequently a protein drug needs to be administered. This half-life is also regulated by other processes such as the neonatal Fc receptor (FcRn) which is a salvage pathway for IgG. In general, the FcRn and protection from reticuloendothelial uptake can help extend the half-life of protein drugs.
Glycosylation is an enzymatic multi-step process in which sugars are added to asparagine (N-linked glycosylation) or serine/threonine (O-linked glycosylation) residues on proteins, shifting the protein's structure from linear polypeptide chain to branched glycoprotein. Glycosylation is nonuniform; there are multiple glycoforms possible for each asparagine or serine/threonine on a protein, each with differing sugars, branching patterns, and caps. Additionally, the glycoform used will vary by cell line for expression: mammalian cell lines express more complex-type N-glycans with sialic acids and fucose, yeast cell lines produce more high-mannose N-glycans, and bacterial cell lines generally do not glycosylate at all. Glycosylation can impact a protein's function: glycan forms being sialylated, the resulting glycoforms are not recognized by liver clearance receptors, and the protein has a longer plasma half-life; glycan lacking fucose improves antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC); and glycan terminated with mannose instead of galactose increases binding to macrophage mannose receptors, which can be exploited for better targeting of these immune cells. O-linked glycosylation is less prevalent in antibodies, but is important for some therapeutic proteins like mucins and fusion proteins, as it affects protease resistance and antigenicity of protein epitopes. It can be challenging to ensure a consistent glycoform in therapeutic protein production, and each glycoform can have a different effect on pharmacokinetics, stability, and immunogenicity of the therapeutic protein, with a complete understanding of how a protein's glycosylation will affect these characteristics being an active area of research.
Absorption of a drug into the systemic circulation after administration is another pharmacokinetic process that can be affected by glycosylation. This aspect of glycosylation has not been as well defined as its role in drug clearance. The absorption of subcutaneously injected drugs, which has become an increasingly common route for monoclonal antibody therapeutics, must first diffuse through the interstitial space rich in glycosaminoglycans and collagen, then enter lymphatic capillaries. Glycosylation at the protein surface can affect its diffusion by changing its size, charge, and/or local stability. Sialic acid residues on the terminal position of glycans increase negative charge which can decrease interaction with surrounding tissue, resulting in faster absorption, while high mannose glycans can bind to local dendritic cells through the mannose receptor, resulting in slower absorption. Increased fucosylation can decrease bioavailability by reducing the flexibility of the protein, with afucosylated proteins showing altered conformational dynamics and tissue permeability. A mixture of glycoforms can lead to a drug product that has different absorption rates and less predictable concentration vs time profile, as well as changes in time to onset of therapeutic effects. Engineering consistent glycan structures to create a uniform population of glycoforms can improve absorption rates. The interaction between glycan structure and the physiology of the interstitial space can also impact the formation of a drug depot at the site of injection. Protein aggregates may form and take time to dissolve and diffuse into the bloodstream. Charge and steric bulk from glycosylation can decrease or increase this lag time.
The rate of absorption of glycoprotein drugs is dependent on the physiochemical environment that surrounds the protein due to the attached carbohydrate groups on the surface. For drugs that are given via routes outside of the vasculature and absorbed via passive diffusion, the absorption rate can depend on the glycosylation present. Glycoproteins injected into the subcutaneous space encounter a concentration gradient between the drug depot and the circulation. The mobility of these proteins through the subcutaneous space extracellular matrix can determine the absorption rate as well as the overall exposure. Sialic acid modified glycoproteins diffuse more readily through the subcutaneous space due to electrostatic repulsion of the negatively charged proteoglycans that comprise the matrix, decreasing the local matrix binding and resulting in an increase in lymphatic transport. Removal of the sialic acid results in glycoproteins with terminal galactose residues that may transiently bind to asialoglycoprotein receptors on resident macrophages, creating a local depot effect that extends the half-life of absorption. Glycans on glycoproteins may sterically shield the protein itself from extracellular peptidases that would otherwise clip the protein and thereby increase the stability of the drug through the absorption phase. Absorption of mannose-terminated glycoproteins are also mediated by the density of mannose receptors present on the surface of the lymphatic endothelial cells, with the capability of either increasing the transcytosis or decreasing local absorption of the protein drug. Temperature-induced glycosylation can also influence absorption with O-GlcNAc modifications to the protein drug that are sensitive to heat stress and can change the conformation of the protein at the microenvironment of the injection site and thereby increase or decrease the diffusion coefficient. Glycoprotein therapeutics that are engineered to contain additional glycosylation sites can be more massive in size and thereby limit lymphatic transport due to an increase in molecular size while still being above the renal threshold for clearance.
Bioavailability, the extent and rate at which a drug reaches the systemic circulation, is the result of the interplay between the process of absorption and local metabolism. Glycosylation, as an overall property, has an important impact on both aspects. The effects of the terminal glycan structure on the degradation of therapeutic glycoproteins by proteases are well understood. Addition of terminal sialic acid (SA) residues results in a net negative charge of the glycoprotein, which protects the protein from proteolytic attack by extracellular proteases. Addition of fucose to the innermost N-acetylglucosamine (core fucosylation) reduces local proteolytic degradation, presumably by increasing resistance to shear forces at the injection site. The heterogeneity of glycoproteins can also have a significant impact on bioavailability. Low-molecular weight, terminal SA-deficient glycoforms can be recognized by asialoglycoprotein receptors on hepatocytes and rapidly removed from circulation. This receptor-mediated uptake can lead to significant batch-to-batch variation, and minimizing asialoglycoprotein receptor targeting, through production of a homogeneous, fully SA-terminated glycoprotein, has a significant positive impact on bioavailability. Hyper-branched glycans, which bulk up a protein, may be able to better avoid uptake by the lymphatics, and thus improve bioavailability. Hyper-branched N-glycans also exclude the protein from the kidney filtration space, once it is absorbed, and thereby improve its bioavailability. In the case of oral delivery of glycoproteins, although stability in the gut is a challenge, N-glycans were recently reported to interact with mucins and transiently trap them in the mucin layer, thereby creating a time-window for absorption that the unglycosylated mutant cannot use. Bioavailability is also a function of tissue distribution and therefore depends on the overall glycan charge. The negatively charged sialylated glycoforms have less vascular permeability and are more often found in plasma than the neutral high-mannose forms. The different charge states of the glycans may also affect the interstitial transport of the proteins.
Distribution of therapeutic proteins in body tissues is another PK aspect of glycosylation. In addition to altering the overall exposure of a drug, the pattern of glycosylation can change the distribution of a protein to different organs. Negatively charged sialic acid residues in glycans, for example, can repel a protein from basement membranes, and mannose-rich glycans can be bound by mannose receptors and sequestered in macrophage-rich organs. This can occur at many barriers, such as blood-brain barrier, placental trophoblast, and synovial intima, all of which have unique lectins that either allow or prevent protein distribution. Additionally, heterogeneity of glycans in recombinant proteins can lead to a multi-phasic distribution if trace glycans direct the protein to different organs. Therefore, glycan structure can be used to manipulate distribution, whether one would like to sequester a protein such as an immunocytokine in the vasculature to reduce side effects or enhance distribution to tissues to improve efficacy by engineering glycans to be either recognized or ignored by the body's lectin network.
Glycans can be used to help control the distribution of a protein in an organism. Steric exclusion and electrostatic repulsion from glycosaminoglycans of basement membranes limits extravascular movement of many sialylated therapeutic glycoproteins, keeping them in the plasma and giving them longer half-lives (often a desired effect, as with monoclonal antibodies). On the other hand, liver asialoglycoprotein receptors bind truncated and desialylated proteins, causing their retention in and rapid clearance by the liver, and restricting their delivery to other organs. High mannose residues are a retention signal, with macrophages in liver, spleen and lymph nodes expressing mannose receptors and internalizing high mannose-containing proteins. Glycans may also have an impact on where proteins end up at the cellular level. The level of fucosylation also influences tissue distribution, since afucosylated proteins tend to have a less extended hinge region which can influence movement through the interstitial space and, thus, enhance tumor penetration. O-linked glycans on mucin-like fusion proteins can dramatically increase hydrodynamic volume and thereby cause steric exclusion effects that protect antibodies from filtration at the kidney while simultaneously restricting transcapillary transport. Glycans with Lewis epitopes can bind to selectins expressed on activated endothelium, causing proteins to become distributed in inflamed tissues. Glycans on proteins can also change distribution over time as a result of circulation in blood, where sialidases can slowly remove terminal sialic acids, changing them from proteins with longer half-lives to ones that are rapidly cleared. Access to the central nervous system (CNS) is limited by both the tight junctions of the blood brain barrier and the efflux mechanisms that transport drugs out of the CNS. Glycans have to be designed carefully in order for the protein to be transported into the CNS. Glycans that are small and neutral appear to cross more readily than highly sialylated or branched glycans.
Penetration is a separate pharmacokinetic aspect where glycosylation acts as a gatekeeper to transvascular flux and diffusion through the interstitium of tissues. A net negative charge on glycans results in electrostatic repulsion from the anionic glycocalyx of vascular endothelial cells, causing a decrease in paracellular diffusion and, potentially, an increase in convective flux through larger gaps between inflamed or tumor endothelial cells. The negative charge effect of sialylation is more pronounced in dense extracellular matrix where sulfated glycosaminoglycans in basement membranes can bind to positively charged protein residues, while sialylation will continue to provide electrostatic repulsion, effectively programming a "stealth" phenotype to minimize off-target tissue penetration. Glycan size and branching also sterically influence penetration with dense, compact high-mannose glycans diffusing more easily through the dense, narrow interstitium of solid tumors compared with the larger, more bulky, complex-type N-glycans commonly found on serum antibodies. This may provide a rationale for a tumor-targeting advantage of afucosylated antibodies. In contrast, the large size and extensive hydration shell of densely branched, sialylated glycans increase the effective molecular radius of a protein, limiting diffusion into areas with dense extracellular matrix such as cartilage, the central nervous system, or tumors with poor stroma where diffusion of large molecules is restricted. Core fucosylation affects molecular rigidity with afucosylated antibodies being more flexible at the hinge regions. This may promote conformational changes needed for proteins to navigate through the extracellular matrix and stroma, potentially improving penetration into solid tumors. O-glycosylation on fusion proteins also alters mucosal penetration where dense O-glycosylation on a domain can mimic mucin structure, increasing adherence of the protein to mucosal surfaces for local drug delivery, and preventing systemic absorption—a strategy used in orally administered biologics that are engineered to remain in the gut and not be absorbed into the bloodstream. In the lung, glycans affect alveolar epithelial transport with mannose-terminated glycoproteins binding to receptors on alveolar macrophages, sequestering them in lung tissue, whereas sialylated glycoproteins are transported more readily across alveolar epithelia into the bloodstream.
As a rheostat, glycosylation finely tunes the residence time of proteins, with mechanisms that increase the stability and/or decrease recognition by receptors involved in clearance. The glycocalyx (glycan shield) protects peptide bonds against serum proteases, but may also provide clearance signals that are triggered when recognition motifs are unmasked. The half-life of a protein is then a competition between shielding and targeting, which can be tuned by changing glycan structures. Terminal sialic acids can extend half-life by preventing recognition by asialoglycoprotein receptors, while their progressive removal by circulating neuraminidases in the case of inflammation, lead to a time-dependent clearance mechanism. High-mannose glycans are recognized and rapidly cleared by mannose receptors in the liver, while complex-type glycans with bisecting GlcNAc are not. Pharmacokinetics are thus not only determined by the initial glycan structures written in the blueprint, but also by their metabolic stability in the bloodstream. Glyco-engineering is now seen as a predictive tool, in which the expected gain in half-life is matched to the addition or removal of specific sugar nucleotides in the cell, or to the knockout of host-cell genes, with the possible caveat of unmasking other clearance pathways and thus affecting dose linearity.
Fig. 1 N-glycans influence the clearance of therapeutic proteins based (A) on size and (B) receptor-mediated clearance.1,5
The mechanism of half-life prolongation by glycosylation results from three converging mechanisms: the increase in molecular size reduces the rate of glomerular filtration, steric hindrance reduces the protease accessibility, and negative charges on the glycan reduce non-specific binding to other tissues. The size exclusion limit of glomerular basement membrane (GBM) filtration is around 70 kDa, and glycosylation has increased the size of many therapeutics beyond this cut-off, so that clearance is now routed through slower receptor-mediated mechanisms rather than through rapid renal clearance. Darbepoetin alfa, for example, has had additional N-glycosylation sites engineered to increase the sialic acid content of the protein; it has a much longer half-life than its aglycosylated counterpart, at the expense of some affinity for its receptor. In addition to the increase in size, the presence of glycans also protect from serum proteases; for example, the oligosaccharide on Asn-297 of IgG fills the space near the lower hinge region and thus reduces the accessibility of the Fc to the metalloproteinases that cleave the hinge region. On the other hand, glycan heterogeneity, which often involves the presence of glycoforms that lack the terminal sialic acids, can lead to shorter half-life: these minor glycoforms are rapidly cleared by hepatic receptors, leading to a multi-exponential decay curve, which results in high peak and trough concentrations and thus makes dosing difficult. Synthesizing glycoforms that are homogeneous and fully sialylated, on the other hand, can provide a more stable plasma concentration. The metabolic stability of the glycan also plays an important role: as proteins circulate through areas of inflammation that have high sialidase expression, their glycans can become gradually desialylated, effectively turning the long-acting version of the protein into its rapidly cleared counterpart.
The two major clearance routes for glycoproteins are the kidney and the liver, which respond to glycans through different receptors. Glycosylation can prevent renal clearance by increasing the molecular weight above the renal filtration limit, but the charge added by sialic acid can also affect reabsorption of the glycoprotein from the kidney tubules. Sialylated proteins are anionic and are therefore not taken up by the proximal tubule, but non-sialylated proteins or proteins with neutral or positively charged glycans can be reabsorbed from the tubule into the circulation, which reduces renal clearance. Hepatic clearance of glycoproteins is primarily mediated by asialoglycoprotein receptor (ASGPR) on the hepatocyte surface, which recognize terminal galactose residues that become exposed when sialic acids are removed from glycans, and trigger fast internalization and degradation in lysosomes. Glycoproteins with high-mannose glycans are not bound by ASGPR, but are instead captured by mannose receptors on resident liver macrophages known as Kupffer cells, which leads to phagocytic clearance instead of metabolic clearance by hepatocytes. Fucosylation also plays a role in hepatic clearance: glycoproteins with core fucose on the N-linked glycans can mask the terminal galactose residues and reduce binding by ASGPR. Glycoproteins with peripheral fucose on Lewis epitopes can also be bound by selectins on liver sinusoidal endothelial cells and form a vascular sink. Genetic silencing of FUT8 to prevent core fucosylation is therefore not only useful for engineering improved antibodies, but is also attractive to reduce the levels of circulating fucosylated ligands that can saturate the clearance machinery. For targeting to liver, high mannose glycans can be added to direct glycoproteins to Kupffer cells, where they are internalized for antigen presentation, and for prolonged circulation, complex sialylated glycans can be used to evade both renal and hepatic clearance pathways. It should be noted, however, that many of the glycoengineered forms studied are in fact hybrid glycans which can activate multiple receptors, resulting in non-linear pharmacokinetics that can be challenging to model for predictable dosing.
Glycosylation can both reduce and enhance the immunogenicity of a protein therapeutic. On the one hand, glycosylation can mask antigenic regions on the protein, shielding them from immune detection. On the other hand, specific glycan structures can be immunogenic and stimulate the production of anti-drug antibodies (ADA). The degree of immunogenicity of a glycoprotein is determined by the composition, structure, and heterogeneity of its glycans. Human-like glycoforms, such as those found on endogenous proteins, are less likely to be immunogenic. In contrast, non-human or abnormal glycan structures, such as plant-associated α-1,3-fucose, NGNA sialic acids, or highly branched and agalactosylated N-glycans, can be immunogenic and lead to the production of ADA. In addition, the degree of glycan heterogeneity can affect immunogenicity. A protein with multiple glycoforms can present a complex mixture of antigens to the immune system. In some cases, trace amounts of impurities that contain immunogenic epitopes can dominate the immune response and lead to ADA production. Glycans also play a role in the aggregation of proteins. Protein aggregation is known to increase the immunogenicity of protein therapeutics. Sialylated and complex-type glycans are less likely to promote protein aggregation than truncated or high-mannose glycans, because they are more hydrophilic and soluble. Aggregation can lead to the exposure of neo-epitopes, which can be recognized as foreign by the immune system. Thus, the glycosylation pattern of a protein therapeutic must be carefully designed to achieve the desired pharmacokinetics and effector function while minimizing the risk of immunogenicity. Achieving this balance is challenging and requires orthogonal analytical strategies to link glycan profiles to immune outcomes and inform rational design of less immunogenic biotherapeutics.
Intracellular signaling cascades are initiated in immune cells through glycan recognition by PRRs. Glycans serve as key recognition elements that immune cells interpret through a variety of pattern recognition receptors, such as C-type lectin receptors (CLRs) expressed on the surface of dendritic cells and macrophages. Lectins that are specific to self, altered-self, and non-self glycan signatures will then direct the activation or tolerance mechanisms that are put into place. Terminal sialic acid residues, particularly those of α-2,6-linkage, will bind to inhibitory Siglec receptors. These receptors will then negatively regulate dendritic cell maturation and T-cell activation and promote an anti-inflammatory phenotype. Exposed mannose and fucose residues, on the other hand, are recognized by activating CLRs such as DC-SIGN, leading to pro-inflammatory cytokine secretion and a strong antigen presentation response. Glycans are also capable of modulating adaptive immunity: glycosylation of the major histocompatibility complex (MHC) and T-cell receptors (TCRs) have been shown to affect peptide loading and TCR clustering, thus affecting the threshold for T-cell activation. These glycan recognition pathways can be utilized by therapeutic proteins to either activate or evade immune activation. The introduction of sialic acid residues can lead to reduced immunogenicity via Siglec engagement. In contrast, introducing mannose residues can lead to enhanced uptake of the protein by antigen presenting cells (APCs) and consequently higher ADA formation. Avidity effects based on spatial clustering of glycan structures can also impact recognition and may supersede individual receptor affinities. Therefore, it is important to understand the glycan-immune interaction for the rational design of biotherapeutics to either avoid or utilize specific immune responses depending on their intended clinical use, such as for vaccine adjuvancy or immunotherapy.
Glycosylation engineering to minimize immunogenicity utilizes a number of complementary approaches, including the elimination of non-human epitopes, and the strengthening of tolerogenic signals. Removal of plant-type α-1,3-fucose and NGNA sialic acids by changing expression systems from plant or rodent cells to humanized CHO or HEK293, removes the most frequent xenogeneic epitopes. Metabolic supplementation with sialic acid precursors induces terminal α-2,6-sialylation which promotes increased inhibitory Siglec receptor engagement, which suppress B-cell activation and ADA development. Glycosylation with high mannose glycans can also activate mannose receptors, increasing immunogenicity. Enzymatic remodeling using endoglycosidases and sialyltransferases can be used to homogenously remodel heterogeneous populations of high-mannose glycans to have a more homogenous, complex-type glycosylation, which is less immunogenic. Site-directed removal of glycans using aglycosylation mutants (e.g. N297A) which removes the Fc glycan entirely may be beneficial for antibodies that need to be neutralizing without eliciting effector functions. This, however, may decrease serum half-life and promote aggregation. More advanced strategies include bisecting GlcNAc via GnT-III overexpression, which modifies the glycan topology and decreases binding to activating Fc-γ receptors, thereby indirectly reducing ADA formation by reducing immune complex formation. The primary challenge to such modifications is maintaining the desired therapeutic effect while reducing immunogenicity, as over-sialylation can mute effector functions, and aglycosylation may promote accelerated clearance. Therefore, a rational design process must include concurrent immunogenicity screening using dendritic cell uptake assays and in vivo ADA screening to ensure glycan modifications have the intended tolerogenic profile and have not altered pharmacokinetics or potency.
Fig. 2 Simplified structure of an immunoglobulin (IgG).2,5
Glycosylation is a major molecular determinant of the pharmacokinetic and immunological behavior of therapeutic proteins, operating at levels ranging from altering proteolysis to modulating the immune response. The field has matured from an era of viewing glycans as artifacts of the production process, to glycan engineering of defined structures, to the rational engineering of glycans as tunable control elements for half-life, biodistribution, and immunogenicity. Rational glycan design now involves manipulation of metabolic pathways, glycan remodeling enzymes, and host cell types to yield defined glycoforms with predictable performance. Glycan heterogeneity has been raised to the status of critical quality attribute by regulatory agencies, and the challenge now is for developers to design robust analytical control strategies for glycan structures that correlate with therapeutic performance. Future developments are likely to include a combination of AI-based glycan prediction with continuous manufacturing approaches for real-time tuning of glycoforms, as well as personalized biologics matched to individual glycoprofiles.
Optimize the behavior of your therapeutic proteins with our expert pharmacokinetic modeling and glycoengineering services. We combine advanced glycosylation analytics-using LC-MS/MS, HILIC-HPLC, CE-MS, and site-specific glycan mapping—with predictive PK modeling to help you understand how glycan structures influence absorption, distribution, clearance, and half-life. Our integrated workflows support biopharmaceutical development from early discovery through late-stage optimization, giving you precise control over glycosylation patterns and their functional impact. Our services enable you to:
Whether your goal is extending half-life, reducing variability, improving safety, or designing next-generation biologics, our pharmacokinetic modeling and glycoengineering solutions deliver the insight and precision needed to accelerate development and strengthen therapeutic outcomes.
References:
