Physiochemical and Biochemical Factors Influencing the Pharmacokinetics of Antibody Therapeutics

Abstract

Monoclonal antibodies are increasingly being developed to treat multiple disease areas, including those related to oncology, immunology, neurology, and ophthalmology. There are multiple factors, such as charge, size, neonatal Fc receptor (FcRn) binding affinity, target affinity and biology, immunoglobulin G (IgG) subclass, degree and type of glycosylation, injection route, and injection site, that could affect the pharmacokinetics (PK) of these large macromolecular therapeutics, which in turn could have ramifications on their efficacy and safety. This minireview examines how characteristics of the antibodies could be altered to change their PK profiles. For example, it was observed that a net charge modification of at least a 1-unit shift in isoelectric point altered antibody clearance. Antibodies with enhanced affinity for FcRn at pH 6.0 display longer serum half-lives and slower clearances than wild type. Antibody fragments have different clearance rates and tissue distribution profiles than full length antibodies. Fc glycosylation is perceived to have a minimal effect on PK while that of terminal high mannose remains unclear. More investigation is warranted to determine if injection route and/or site impacts PK. Nonetheless, a better understanding of the effects of all these variations may allow for the better design of antibody therapeutics.

INTRODUCTION

Antibody therapeutics are becoming increasingly important in fighting many diseases. Antibodies are tetrameric glycoproteins composed of two heavy chains (HC) and two light chains (LC) held together by disulfide linkages (Fig. 1) (1). When properly folded into the “Y” shape motif, the constant domains C_H1 and C_L and variable domains V_H and V_L comprise the antigen binding fragment (Fab), with the variable domain containing the complementarity-determining region that is highly specific for a target antigen. The C_H2 and C_H3 domains of the heavy chains form the Fc fragment (fragment crystallizable) of the antibody and can bind to Fc receptors on cells to illicit immune effector functions or to the neonatal Fc receptor (FcRn) to protect it from degradation (2). Human immunoglobulin Gs (IgGs) are divided into four subclasses, i.e., IgG1, IgG2, IgG3, and IgG4, with structural differences in their heavy chains characterizing each subclass (1). It should be noted that IgG3 antibodies are known to be less stable than the other IgG subclasses (3) and, to date, have not been developed as therapeutic antibodies. Examining human and cynomolgus monkey (cyno) clearance values for a wide range of IgG1 antibodies with similar Fc domains demonstrates that different IgG1s have different clearance rates ranging from 0.4 to 40 mL/h (4). Antibody clearance can be driven by target expression and biology, FcRn binding, proteolytic catabolism, and cellular uptake. Furthermore, binding of antibodies to FcRn on epithelial and endothelial cells through their Fc domain regulates their metabolism and partially accounts for their relatively long serum half-lives (5,6).

Fig. 1.

A diagram of a typical IgG1 antibody along with its constituent parts. Reproduced from Mould and Sweeney (1). Originally published by BioMed Central

There are a number of other factors that can impact the pharmacokinetics (PK) of antibodies, such as antibody structural characteristics, delivery and formulation, target properties, and physiological parameters. Research on how manipulating antibody charge, size, valence, FcRn-binding affinity, and degree and type of glycosylation can influence their metabolism, clearance, penetration, and lymphatic absorption has been conducted (7–15). Furthermore, a need exists to further optimize drug delivery methods through formulation, dose, injection volume, and injection site (16–19). Antibody target expression levels and affinity as well as the local physiology of the target-expressing tissue, such as vascularity, blood flow, interstitial pressure, degree of necrosis, and/or fibrosis, may also have an impact on antibody PK (20–23). This minireview focuses only on the physiochemical and biochemical factors affecting antibody PK and aims to provide a better understanding about how antibody charge, size/valence, FcRn affinity, degree and type of glycosylation, and delivery route could affect its PK. It is a summary of a presentation given at an AAPS Focus Group Workshop (ADME of Protein Therapeutics Introductory Workshop: Scientific, Technical Concepts and Case Studies, August 15, 2011, Buffalo, NY, USA).

CHARGE AND ISOELECTRIC POINT (pI)

Charge is a crucial determinant of the interactions of biological drug material with biosystems (11). Therefore, there is a strong rationale that charge modifications will alter the electrostatic (or nonspecific) interactions with negatively charged components of the cell surface (i.e., sialic acids and glycosaminoglycans). These in turn can play a role in dictating both serum and tissue PK of antibodies. Chemical modifications of antibodies can modulate their net charge and/or hydrophobicity resulting in noticeable changes in biological activity. Typical pI values for antibodies fall in the range of 8–9 (11). Cationized proteins have a high propensity to adhere to anionic sites of cell surfaces (i.e., heparan sulfate proteoglycans and phospholipids), so the chemical modification of antibodies or antibody fragments with cationized residues that increase their pI by >1 units has resulted in increased plasma clearance, with a higher disposition to normal tissues and higher target tissue uptake (Table I) (24,25). Conversely, modification of Fabs with anionic groups causing a reduction in pI by 1–2 units was shown to result in decreased blood clearance and tissue accumulation relative to the unmodified Fab (26). It has been observed that a neutral charge modification that decreased an antibody’s pI by more than 1 unit can exhibit faster blood clearance compared to its unmodified counterpart with less activity in normal tissues and improved localization in the target (27). More recently, a study has shown that antibodies with higher pI values also tended to exhibit faster systemic clearance rates and lower subcutaneous bioavailabilities in both human and minipig than antibodies with lower pI values (28).

Table I.

Serum Exposure (AUC) and Tissue Uptake (%ID/g) of a Radiolabeled Antibody and its Cationized Counterpart

Parameter	[111In]-native 528 MAb	[111In]-cationized 528 MAb
AUC (60 min) (%ID·min/mL)	4,180 ± 210	3976 ± 192
AUCss (%ID·min/mL)	25,887 ± 3,380	17,770 ± 2,301
Liver uptake (%ID/g)	11.0 ± 0.5	18.6 ± 0.8
Kidney uptake (%ID/g)	0	4.8 ± 0.4
Heart uptake (%ID/g)	0	0
Lung uptake (%ID/g)	2.5 ± 1.6	4.9 ± 0.5

The tissue uptake has been blood background corrected, so the values represent the extravascular tissue radioactivity levels. Reprinted (adapted) with permission from Lee and Pardridge (24). Copyright 2003 American Cancer Society

AUC area under the concentration–time curve, AUC _ss AUC at steady state, %ID/g percent injected dose per gram of tissue

To further elucidate the relationship between the pharmacokinetics and the pI of antibodies, Igawa et al. (29) generated antibody variants with modified pI values using site-directed mutagenesis. Specific substitutions were chosen within the surface residues of the heavy chain variable region to modify the pI without affecting the antigen-binding property of the antibody. Variants with pI values of 1–2 units lower than wild type were shown to display longer half-lives and clearance rates, whether they were administered intravenously or subcutaneously into mice. These results suggest that altering the pI through modification of the variable region could offer an alternative to Fc engineering.

As part of many posttranslational events, antibodies undergo chemical or enzymatic degradation via several different mechanisms, including oxidation, deamidation, isomerization, and fragmentation, which result in the formation of several charge variants (12). The presence of charge variants in all antibody preparations is consistent, but the percentage of each variant differs per antibody. However, until recently, the impact of those variations on PK was largely unknown. In a recent study, acidic and basic antibody variants of an IgG1 monoclonal antibody with a range of pI values between 8.7 and 9.1 were isolated and administered intravenously and subcutaneously to rats (12). There was no significant difference in the observed serum PK profile, demonstrating that pI differences among charge variants were not large enough to result in PK changes and corroborating that at least a net difference of one pI unit is needed to exert any influence on systemic PK and tissue distribution.

SIZE AND VALENCE

Antibody fragments, such as single-chain Fv, diabody, triabody, Fab, F(ab′)₂, and full length antibodies, ranging in size from 30 to 150 kDa and valence from one to three binding sites (9,10) can be derived via molecular engineering or controlled enzymatic digestion. While retaining their antigen-binding capabilities, these fragments displayed different in vivo serum and tissue PK. The smallest fragments not only cleared the fastest but were also shown to have much higher tumor/organ ratios compared to their larger counterparts. The small fragments also reached their peak tumor levels earlier than the full length antibody.

FcRn BINDING AFFINITY

The role of FcRn in prolonging the half-life of serum IgG has been well characterized (30). It has been noted that murine FcRn binds with high affinity to IgG from different species, including human, whereas human FcRn is more selective. This receptor is expressed within intestinal epithelial cells, endothelial cells, and circulating monocytes and at the maternal–fetal barrier, although species differences in expression have been observed (30). Generally, fluid phase uptake internalizes serum IgG into intracellular endosomes. IgG binds FcRn via pH-dependent electrostatic interactions at the endosomal acidic pH (<6.5) and is recycled and released extracellularly into systemic circulation at the more neutral pH 7.4 (30). This interaction with FcRn protects antibodies from lysosomal degradation. The effect of altering the affinity for FcRn either at pH 6 or 7.4 or even at both pH values has been investigated. In both cynos and mice, increasing binding affinity to FcRn at pH 6 with negligible change to binding at pH 7.4 resulted in decreased serum clearance and increased half-life compared to that of the wild-type antibody (Fig. 2) (7,8). In contrast, increasing the binding to FcRn at both pH 6 and 7.4 exhibited very similar PK to that of the wild-type antibody, indicating that there was no improvement in PK with increased affinity at pH 7.4 (7,8). Despite the different mechanisms underlying the ADME determinants following intravenous and subcutaneous dosing, the same trend was observed regardless of injection route. These data suggest that while it is important to increase binding affinity at pH 6.0, it is also critical to preserve the efficient pH-dependent release of IgG at pH 7.4 when engineering antibodies with a slower clearance and longer half-life than the wild type. Interestingly, prolonged serum PK of high affinity FcRn antibody variants has been obtained in mice expressing human FcRn (30); however, this benefit has yet to be observed in the clinic (31).

Fig. 2.

Serum concentrations following 20 mg/kg i.v. administration in cynos of an antibody (WT) and its two FcRn variants, N434A and N434W. Both variants had improved binding to FcRn at pH 6.0, but N434W also had increased binding to FcRn at pH 7.4. N434A cleared slower from the serum than the wild-type antibody and N434W had the same clearance. Reproduced with permission from Yeung et al. (7). Copyright 2009. The American Association of Immunologists, Inc

GLYCOSYLATION

In general, glycans in the Fc domain are important for specific effector functions. All human IgGs contain carbohydrates at Asn²⁹⁷ in the C_H2 domain of the Fc region (32). These N-glycans are structurally heterogeneous containing a combination of terminal galactose and N-acetylglucosamine (GlcNAc) residues. The G0 structure is the most basic form of the N-glycan structures, terminating in two GlcNAc units (Fig. 3). The G1 and G2 designations refer to the modification of the G0 structure with either one or two galactose units, respectively. The GlcNAc unit conjugated directly to Asn²⁹⁷ is linked to a fucose moiety (called the “core fucose”) that is in the Fc gamma receptor (FcγR) binding region, and the absence of this fucose (G0-F) has been shown to improve binding of the antibody to FcγRIIB and FcγRIIIA in addition to enhancing ADCC activity (33). Additionally, there are high mannose, but low galactose and GlcNAc containing glycans, designated Man5, Man8, and Man9, which have been observed and have shown differential effects on FcγR binding and ADCC or CDC activity (33).

Fig. 3.

Schematic structures of various glycan oligosaccharides and their designations. The individual glycans are denoted as follows: galactose (open diamond), GlcNac (closed circle), mannose (open square), and fucose (F, open triange). Reproduced from Kanda et al. (33), by permission of Oxford University Press

The role of glycans in modulating clearance is not well understood (34). Glycans are buried in the interior surface of the Fc domain, making them inaccessible to asialoglycoprotein and mannose receptors that could recognize the moieties and clear the antibody (13). Consistently, differences in murine PK were not observed between a wild-type and aglycosylated chimeric IgG1 (13). The effect of terminal mannose (Man5-8) on PK has also been examined. In a more recent study, high mannose (Man5) showed increased clearance in humans (15). However, earlier studies with omalizumab (IgG1) showed that Fc glycan structure, including high mannose, did not alter the antibody’s clearance in mice, but this could be due to the fact that only a small percentage of the antibody contained high mannose (14). Furthermore, the degree of Fv sialylation did not appear to impact antibody PK (35) nor did complete desialylation (36). There is limited published information about the impact of glycosylation modifications on antibody tissue PK (37).

DELIVERY ROUTES AND SITES

Conflicting results have been reported on the effects of the delivery route and site of administration on PK. A comparison between the bioavailability of Alefacept (Fc fusion protein) following 0.075 mg/kg intramuscular and subcutaneous administration resulted in higher bioavailability with the intramuscular dosing (38). However, in a later bioequivalence study at a flat 15 mg dose in healthy volunteers, nearly identical exposures following intramuscular and subcutaneous administration were observed, demonstrating bioequivalence between the two routes. Conversely, within an injection route, no effect was seen in PK following subcutaneous injection in the upper arm, abdomen, or thigh of 100 mg of golimumab in healthy male volunteers, suggesting that injection site has little impact on PK (39), although more research in this area is warranted. Indeed, in a recent review by McDonald and colleagues (40), injection site was identified as a possible factor influencing the bioavailabilities of molecular therapeutics.

CONCLUSIONS AND FUTURE PERSPECTIVE

In summary, studies have been conducted evaluating the various effects of charge, size, glycosylation, and FcRn affinity at different pH values on antibody PK and tissue distribution. However, more investigation is needed to elucidate how systemic PK is altered by charged amino acid substitutions in the variable or constant regions of the antibody, different IgG subclasses, and different injection routes and sites. Furthermore, very little is known about the PK drivers of the subcutaneous space, and there is a particular need for a systematic evaluation of charge alteration to address such issues as saturation of subcutaneous electrostatic binding sites. Another area that remains unclear is if FcRn in the lymphatic and subcutaneous space could be saturated following administration of high concentration formulations. The effect of glycosylation on subcutaneous PK is also not well understood. Although a great deal has been done to examine the many factors governing antibody PK, more work in this area is needed.

Abbreviations

Cyno

Cynomolgus monkey

Fab

Antigen binding fragment

Fc

Fragment crystallizable

FcγR

Fc gamma receptor

FcRn

Neonatal Fc receptor

GlcNAc

N-acetylglucosamine

HC

Heavy chain

IgG

Immunoglobulin G

LC

Light chain

pI

Isoelectric point

PK

Pharmacokinetics