Half-Life Extension of Biopharmaceuticals Using Chemical Methods: Alternatives To Pegylation

DOI: 10.1002/cmdc.
201600374 Reviews
Half-Life Extension of Biopharmaceuticals using Chemical

Methods: Alternatives to PEGylation
Søren B. van Witteloostuijn,[a, b] Søren L. Pedersen,[b] and Knud J. Jensen*[a]
ChemMedChem 2016, 11, 1 – 23 1 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim &
These are not the final page numbers! ÞÞ

Reviews
Peptides and proteins constitute a vast pool of excellent drug include PEGylation, fusion to unstructured polypeptide-based
candidates. Evolution has equipped these molecules with su- PEG mimetics, conjugation of large polysaccharides, native-like
perior drug-like properties such as high specificity and poten- glycosylation, lipidation, fusion to albumin or the Fc domain of
cy. However, native peptides and proteins suffer from an inade- IgG, and derivatization with bio-orthogonal moieties that
quate pharmacokinetic profile, and their outstanding pharma- direct self-assembly. This review provides an overview of avail-
cological potential can only be realized if this issue is ad- able conjugation chemistries, biophysical properties, and
dressed during drug development. To overcome this challenge, safety data associated with these concepts. Moreover, the ef-
a variety of half-life extension techniques relying on covalent fects of these modifications on peptide and protein pharmaco-
chemical modification have been developed. These methods kinetics are demonstrated through key examples.
1. Introduction this modification may lead to hypersensitivity reactions,[4] anti-

body formation,[5] and PEG bioaccumulation[6] (see below).
Since the approval of recombinant human insulin in 1982 by Herein we provide an overview of the most important chem-
the US Food and Drug Administration (FDA), biopharmaceuti- ical modifications to provide half-life extension of biopharma-
cals have become increasingly important from both a medical ceuticals, which are alternatives to PEGylation. The primary
and an economic perspective. Currently, approximately 250 bi- focus is on the conjugation of large carbohydrate polymers,
opharmaceutical drug products are approved and available on native O- and N-glycosylation, lipidation, and experimental
the US and European markets for the treatment of a variety of methods that direct nano-scale self-assembly. The chemical at-
diseases, including several cancer forms, inflammatory disease, tachment of unstructured polypeptides, albumin conjugation,
diabetes, obesity and heart disease.[1] The total annual revenue and Fc fusion is also mentioned (Table 1). The emphasis is on
generated by biopharmaceuticals has been steadily increasing, chemical moieties, the biophysical properties they direct, and
reaching USD 140 billion in 2013[1] and predicted to reach USD their biological compatibility. The impact of each half-life ex-
239 billion in 2015.[2] tension method is illustrated through selected examples.
The vast majority of biopharmaceuticals are based on en- The focus of this review is on alternatives to PEGylation;
dogenous peptides and proteins, such as hormones, enzymes, chemical modification by PEGylation has previously been ex-
growth factors, interferons, and antibodies. Inherently, these tensively reviewed.[7] Noncovalent drug formulation is not cov-
natural ligands possess superior drug properties such as highly ered in this review.
selective receptor binding and potent receptor activation. This
minimizes off-target side effects and allows administration of
very low drug doses, thereby improving the drug safety profile
2. PEGylation
and increasing patient compliance. Nevertheless, direct use of
most native polypeptides as drugs is hampered by their poor The attachment of poly(ethylene glycol) (PEG) moieties to pep-
pharmacokinetic (PK) properties, owing to their rapid metabo- tides and proteins is a well-established and efficient method
lism, proteolytic degradation and, especially in the case of pep- for improving their pharmacokinetic properties. PEG polymers
tides and smaller proteins, susceptibility to renal clearance. are highly hydrophilic and provide with their associated water
These obstacles result in very short systemic half-lives, typically molecules a hydration shell, which expands the hydrodynamic
in the range of minutes to few hours.[3] This challenge has volume of a polypeptide. This makes the PEGylated molecule
prompted the development of several half-life extension strat- less susceptible to renal clearance as well as to protease degra-
egies to improve the PK properties of native peptides and pro- dation, and can decrease the immunogenicity of native pep-
teins. tides and proteins.[7a] The concept of PEGylation was intro-
Chemical modification of peptides and proteins has become duced in the 1970s by Davis and co-workers,[8] and today 12
an essential route to enhance the PK profile of biopharmaceut- PEGylated peptide- and protein-based biopharmaceuticals are
icals. Arguably, the most common and so far most commercial- available on the American and European markets.[9]
ly successful approach has been PEGylation, that is, the cova- PEG polymers are produced synthetically and can be either
lent attachment of polyethylene glycol (PEG) chains. However, linear or branched, and the end group(s) may be either the
there is growing concern about the safety of PEGylation, as standard hydroxy group or a methoxy group (denoted mPEG)
(Figure 1). Typically, commercially available PEG polymers used
for chemical modification of biopharmaceuticals range up to
[a] S. B. van Witteloostuijn, Prof. K. J. Jensen 50 kDa. PEGylation of polypeptides may be obtained with dif-
Department of Chemistry, University of Copenhagen, Thorvaldsensvej 40, ferent strategies.[7b, c, 10] Usually, the N-terminal amino group or
1871 Frederiksberg C (Denmark)
the Ne-amine of a Lys residue is used as nucleophile in either
E-mail: kjj@chem.ku.dk
alkylation or acylation reactions with an activated PEG species.
[b] S. B. van Witteloostuijn, Dr. S. L. Pedersen
Gubra ApS, Hørsholm Kongevej 11B, 2970 Hørsholm (Denmark) An example of an alkylating reagent is mPEG-propionaldehyde,
The ORCID identification number(s) for the author(s) of this article can which reacts with the amino groups of polypeptides to form
be found under http://dx.doi.org/10.1002/cmdc.201600374. an imine (Figure 2 A). Subsequent reduction with NaBH3CN
& ChemMedChem 2016, 11, 1 – 23 www.chemmedchem.org 2 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Reviews
Søren B. van Witteloostuijn obtained

his MSc degree in Biology–Biotech-
nology from the University of Copen-
hagen in 2013. He currently holds a po-
sition as research scientist and indus-
trial PhD student with Gubra ApS and
the University of Copenhagen under
the supervision of Dr. Søren L. Peder-
sen and Prof. Knud J. Jensen. His re-
Figure 1. Different types of PEG structures; mPEG: methoxy-PEG. Adapted
search interests include the design and
from Harris & Chess 2003.[7a]
synthesis of peptide-based drugs, bio-
physical characterization of peptide
oligomers, and development of novel principles for half-life exten-
sion of biopharmaceuticals. provides the secondary amine.[7b, 10] Overall, this process is a re-
ductive amination.
Active succinimidyl esters of PEG-carboxylic acids (PEG-NHS)
are the most widely used reagents for acylation reactions.[10]
Dr. Søren L. Pedersen obtained his PhD Acylation of amines with PEG-NHS yield a polypeptide-PEG
in bioorganic chemistry in 2010 with conjugate held together by an amide bond (Figure 2 B). PEGy-
Prof. Knud J. Jensen, University of Co- lation of amines is generally feasible and occur under mild
penhagen, in collaboration with Rheo- conditions. However, a potential problem with targeting Lys is
science. Subsequently, he was a post- the high prevalence of accessible Lys residues in native pep-
doctoral researcher with Prof. Knud J. tides and proteins, possibly leading to a mixture of products
Jensen at the Lundbeck Foundation and batch variation.
Center for Biomembranes in Nanome- An alternative is to use the nucleophilic thiol functionality of
dicine at the University of Copenha- a free Cys. If no free Cys is present, it may be introduced
gen. In 2011, he joined Gubra ApS, chemically or recombinantly to enable site-specific PEGylation.
where he is now a senior scientist and Typical and widely used reagents for this purpose are PEG-mal-
group leader. His research group is en- eimides.[7c] Under neutral to slightly basic conditions, Michael
gaged in peptide drug discovery within diabetes and obesity, and addition of the thiol to the maleimide results in the formation
target discovery in specialized regions of the gut as well as the of a thioether (Figure 2 C). The resulting thioether is, however,
brain. not completely stable under aqueous conditions or in vivo, po-
tentially leading to deconjugation of a thiol-reactive linker.[11]
Maleimide exchange and succinimide ring hydrolysis in the
conjugated linker have been reported for several conjuga-
Knud J. Jensen obtained a PhD degree tes.[11b, 12] PEG-maleimide can exchange with albumin, cysteine,
in synthetic bioorganic chemistry from or glutathione in vivo, giving the non-modified biopharma-
the University of Copenhagen in 1992
ceutical, or undergo succinimide ring hydrolysis. To avoid
with Prof. Morten Meldal, after which these challenges, PEG-haloacetamides may be employed for
he did postdoctoral research with Prof.
Cys modification (Figure 2 D).
George Barany, University of Minneso- The PEGylation technology has been highly successful, as re-
ta. He became assistant professor at flected by the large number of PEGylated drugs available on
the Technical University of Denmark in the market and in development. However, because PEGs are
1997 and associate professor at KVL in synthetic polymers, many commercially available reagents are
Copenhagen in 2001, which in 2007
provided as polydisperse mixtures, a feature that inevitably is
became part of the University of Co- reflected in the PEGylated drug product. PEGs with narrow and
penhagen. The same year he was pro-
well-characterized dispersities are thus preferred. Moreover,
moted to full professor in nanobioscience. He is now a full profes- the safety of PEG is still a matter of debate.[2, 13] Data suggests
sor at the Department of Chemistry, University of Copenhagen and that administration of PEGylated drugs may cause both hyper-
Head of the section for Chemical Biology. He is the co-author of sensitivity reactions as well as antibody formation.[4, 5] Other
more than 130 peer-reviewed publications, has edited two books concerns include bioaccumulation of PEG and the fate of this
on the design and synthesis of peptides, and has co-authored nu-
synthetic polymer after injection.[6] These issues have led to
merous book chapters and proceedings. He was awarded the the investigation and development of alternatives to PEGyla-
Zervas Award of the European Peptide Society in 2012. His re-
tion for extending the half-life of biopharmaceuticals.
search covers a broad range of topics at the interface between
synthetic chemistry, medicinal chemistry, biology, biophysics, and
nanobioscience.
ChemMedChem 2016, 11, 1 – 23 www.chemmedchem.org 3 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim &

Reviews
Table 1. Currently approved peptide- and protein-based biopharmaceuticals that are PEGylated, lipidated, or fused to albumin or the Fc domain of IgG.[a]
INN[b] Brand name t1/2 extension Indication Approval

method year[c]
Pegadamase Adagen PEGylation Severe combined immunodeficiency disease (SCID) 1990
Pegaspargase Oncaspar PEGylation Leukemia 1994
Peginterferon-a2b PegIntron PEGylation Hepatitis C 2000
Peginterferon-a2a Pegasys PEGylation Hepatitis C 2001
Pegvisomant Somavert PEGylation Acromegaly 2002
Pegfilgrastim Neulasta PEGylation Neutropenia 2002
Methoxy polyethylene glycol-epoetin b Mircera PEGylation Anemia associated with chronic kidney disease 2007
Certolizumab pegol Cimzia PEGylation Rheumatoid arthritis, Crohn’s disease 2008
Pegloticase Krystexxa PEGylation Chronic gout 2010
Lipegfilgrastim Lonquex PEGylation Neutropenia 2013
Peginterferon-a2a Plegridy PEGylation Relapsing multiple sclerosis 2014
Antihemophilic factor (recombinant) Adynovate PEGylation Hemophilia A 2016
Insulin detemir Levemir Lipidation Type 1 diabetes 2004
Liraglutide Victoza Lipidation Type 2 diabetes 2009
Insulin degludec Tresiba Lipidation Type 1 diabetes 2013
Liraglutide Saxenda Lipidation Obesity 2014
Albiglutide Tanzeum, Eperzan Albumin fusion Type 2 diabetes 2014
Etanercept Enbrel Fc fusion Rheumatoid arthritis, psoriasis 1998
Abatacept Orencia Fc fusion Rheumatoid arthritis 2005
Rilonacept Arcalyst Fc fusion Cryopyrin-associated periodic syndromes (CAPS) 2008
Romiplostim NPlate Fc fusion Immune thrombocytopenic purpura (ITC) 2008
Aflibercept Eylea Fc fusion Wet macular degeneration 2011
Belatacept Nujolix Fc fusion Kidney transplant rejection 2011
Ziv-aflibercept Zaltrap Fc fusion Colorectal cancer 2012
Antihemophilic factor (recombinant) Eloctate Fc fusion Hemophilia A 2014
Coagulation Factor IX (recombinant) Alprolix Fc fusion Hemophilia B 2014
Dulaglutide Trulicity Fc fusion Type 2 diabetes 2014
[a] All fusion proteins listed are obtained by recombinant expression. [b] International nonproprietary name. [c] Approval year reflects first approval in
either USA or Europe.
Figure 2. Examples of PEGylation chemistry. A) mPEG-propionaldehyde reacts with an amino group to form an imine, which is reduced to form a secondary
amine. B) PEG-NHS reacts with an amino group to form an amide bond. C) PEG-maleimides react with free Cys to provide a thioether linkage. D) PEG iodo-
acetamides also react with free Cys to form a thioether bond. R: peptide or protein.
3. Recombinant PEG Mimetics

structured conformation when attached to proteins.[14] The hy-
As an alternative to the synthetic PEG polymers, a number of drophilic properties ensure solvation of the polymers, which
recombinant ‘PEG mimetics’ for half-life extension has been re- decreases the risk of aggregation during storage and allows
ported. These polypeptides are specific amino acid sequences high-dose formulation of the fusion protein. The absence of
that are attached to biopharmaceuticals as fusion proteins ob- hydrophobic residues also minimizes unspecific interactions
tained by recombinant DNA technology. In general, these poly- with biological membranes and hydrophobic cell components.
mers are hydrophilic or even charged, and they adopt an un- Moreover, the unstructured nature of these polypeptides pro-

Reviews
vides conformationally flexible polymers with an expanded hy- 3.2 Other PEG mimetics
drodynamic volume, which retards glomerular filtration and
renal clearance of the active fusion protein.[14] As such, these In addition to XTEN, a variety of other unstructured polypep-
polymers may be regarded as biocompatible alternative to tides have been reported for optimizing the PK profile of bio-
PEG. pharmaceuticals. However, these technologies all rely on re-
combinant expression to provide a fusion protein with extend-
ed half-life. Examples include elastin-like polypeptide (Phase-
3.1 XTEN
Bio, www.phasebio.com),[19] gelatin-like protein,[20] and Pro-Ala-
XTEN is an 83.5 kDa recombinant polypeptide developed by Ser (PASylation; XL-Protein, www.xl-protein.com).[21] Several
Amunix Inc. The XTEN sequence is a nonrepetitive polymer drug candidates using these technologies are currently in clini-
consisting only of the amino acids Ala, Asp, Gly, Pro, Ser, and cal development. For a detailed account, the reader is referred
Thr. This amino acid composition provides a biopolymer that is to the overview by Binder & Skerra.[14]
unstructured, nonimmunogenic, highly soluble, and chemically
stable.[15] XTENylated biopharmaceuticals were originally pro-
duced as fusion proteins of the XTEN sequence and the phar- 4. Carbohydrate Polymers
maceutically active peptide or protein using recombinant DNA
4.1 Dextran conjugation
technology. The most advanced drug candidate using the
XTEN technology is somavaratan (VRS-317, Amunix/Versatis), Conjugation of dextran was one of the first alternatives to
which is currently in phase III clinical development for pediatric PEGylation, with early efforts dating back to the 1970s.[22] Dex-
growth hormone deficiency (GHD).[16] tran polymers are obtained from bacteria such as L. mesenter-
In a proof-of-concept study, the PK properties of a recombi- oides and are d-glucose polymers linked by a(1–6) glycosidic
nant fusion protein consisting of XTEN and the GLP-1 analogue linkages and a small extent of a(1–3) bonds (~ 95 % a(1–6) and
exenatide (named E-XTEN) were evaluated in cynomolgus 5 % a(1–3) in the case of L. mesenteroides) (Figure 3). The a(1–
monkeys. Fusion to XTEN prolonged the terminal (biological) 3) linkages serve as attachment points for primarily mono- or
half-life of exenatide from 0.48 to 60 h following subcutaneous disaccharide units, giving rise to a low degree of molecular
(s.c.) administration, corresponding to a 125-fold increase, em- branching.[23] Dextran preparations have been used as plasma
phasizing the potential of the technology.[15] volume expanders since the 1950s, and today two dextran so-
Although construction of XTEN fusion proteins initially was lutions (containing 10 % of 40 kDa dextran and 6 % of 70 kDa
an entirely recombinant process, it is now also possible to dextran, respectively) are approved in the US for this pur-
make XTEN conjugates by chemical means. Chemical XTENyla- pose.[24] For research purposes, dextrans with mean molecular
tion uses special versions of the XTEN polymer containing for weight (MMW) in the range of 4 to 200 kDa and low
example, amines, thiols, or dibenzocyclooctyne (DBCO) groups, polydispersity ( 2) are commercially available.[24]
which react selectively with succinimidyl esters, maleimides, or
azides, respectively. These reactive groups may then be intro-
duced into peptides by chemical methods to allow subsequent
conjugation. Examples include the glucagon-like peptide 2
(GLP-2) analogue GLP2-2G-XTEN developed by Amunix, which
consists of an XTEN variant and the GLP-2 analogue teduglu-
tide assembled by maleimide–thiol chemistry. The resulting
conjugate has a markedly decreased in vitro potency, as re-
flected in a > 50-fold decrease in relative in vitro activity rela-
tive to unmodified teduglutide. However, s.c. administration to
Sprague–Dawley rats revealed a half-life extension to 38.5 h,
resulting in a largely increased drug exposure.[17]
Chemical XTENylation has also been used to produce multi-
valent peptide-XTEN conjugates otherwise inaccessible by re-
combinant methods. Ding et al. have published conjugates of
XTEN and the HIV fusion inhibitor T-20 (enfuvirtide, Fuzeon) as-
Figure 3. Example of dextran architecture. Degree of cross-linking varies
sembled by maleimide–thiol chemistry. The conjugates contain among different dextran preparations.
multiple copies of T-20 attached to a single XTEN polymer, but
instead of traditional end-to-end fusion, the T-20 molecules are
attached at various positions along the XTEN sequence
through the side-chains of Cys residues. Investigation of the PK In addition to unmodified dextran, various synthetic dextran
properties of the most active conjugate (named T-20-XTEN-4) derivatives such as carboxymethyl-dextran (CMD), diethylami-
following s.c. administration to Sprague–Dawley rats revealed noethyl-dextran (DEAED), glycosylated versions of CMD such
an elimination half-life of 55.7 h, corresponding to a 20-fold in- as galactose-CMD (Gal-CMD) and mannose-CMD (Man-CMD),
crease relative to the reported half-life of unmodified T-20.[18] carboxymethyl benzylamide dextran (DCMB), carboxymethyl

Reviews
sulfate dextran (DCMSu), and carboxymethyl benzylamide sul- threshold for renal clearance of dextran is ~ 40 kDa, as indicat-
fate dextran (DCMBSu) have also been reported.[25] ed by studies of both unconjugated dextran as well as fluores-
Historically, methods for coupling proteins to dextran have cein–dextran conjugates.[33] Studies of carboxypeptidase–dex-
focused on obtaining dextran conjugates with multiple copies tran conjugates in tumor-bearing Balc/c mice showed that
per dextran moiety rather than site-specific incorporation. First elimination of dextran is size-dependent, with conjugates con-
attempts involved cyanogen bromide (BrCN) activation of dex- taining large dextran moieties exhibiting the longest half-
tran hydroxy groups with subsequent aminolysis by protein life.[34] This notion was further supported by PK studies of fluo-
amino groups.[26] However, large product diversity combined rescein–dextran conjugates in rats.[24, 33b] Gel filtration studies of
with the apparent instability of the protein–dextran linkage as the fluorescein–dextran conjugates demonstrated a linear rela-
well as the toxicity of the BrCN reagent made this method un- tionship between dextran size and column retention time,
attractive.[23, 27] with larger dextrans displaying shorter retention times.[24, 33b] In
Periodate oxidation to form dextran dialdehydes[28] followed addition, conjugation of dextran has been shown to decrease
by reductive amination to yield a secondary amine linkage has in vitro enzymatic degradation of catalase[32c] and to reduce
been used successfully for synthesizing protein–dextran conju- the immunogenicity of xenogeneic antibodies.[35]
gates (Figure 4 A).[25b, 29] Takakura et al. prepared conjugates of Since 1975, dextran has been labeled as ‘Generally Regarded
dextran and soybean trypsin inhibitor (STI) by six different As Safe’ (GRAS) by the FDA,[36] testifying to the overall safety of
chemical methods, including periodate oxidation and cyano- these polymers. Dextran is metabolized in vivo by dextranases,
gen bromide activation. Of all the tested methods, periodate which are found predominantly in the liver and spleen.[23]
oxidation provided both high reaction yield, high degree of The ability of dextran and other colloid volume substitutes
molar substitution and high activity of the modified enzyme, to induce anaphylactoid reactions was investigated by Ring &
indicating that periodate oxidation is the preferred method for Messmer and the incidence of severe reactions following dex-
synthesizing protein–dextran conjugates.[30] A similar conclu- tran administration was found to be 0.008 %.[37] This was similar
sion was made by Yasuda et al. from studies of dextran-conju- to the incidence observed following administration of hydroxy-
gated uricase.[31] However, although periodate oxidation of ethyl starch (HES, see below) (0.006 %) and slightly raised rela-
dextran may be controlled to some extent,[28a] inherent draw- tive to plasma protein solutions (0.003 %).[37] In a similar study
backs associated with this method include instability of the di- conducted by Laxenaire et al., the incidence of anaphylactoid
aldehyde intermediate, product heterogeneity due to the pres- reactions following dextran administration was found to be
ence of multiple ring-opening sites, and cross-linking of the 0.273 %, corresponding to a slightly increased incidence rela-
target protein if this contains multiple amino groups.[2, 23] tive to HES (0.0058 %) and albumin (0.099 %).[38] However, the
Dextran conjugation via a spacer has been extensively ex- very low incidence rates observed in both studies emphasize
plored for small molecules,[23] but has also been applied to the low risk of developing anaphylactoid reactions in response
peptides[22] and enzymes.[32] Introduction of spacers such as e- to dextran infusion.
aminocaproic acid at the reducing end has enabled site-specif- Although dextran-reactive antibodies have been implicated
ic conjugation of dextran to the N-terminus of catalase sub- in dextran-induced anaphylactoid reactions,[39] there are, to the
units (Figure 4 B).[32c] Lastly, dextran conjugation to proteins has best of our knowledge, no published reports of antibodies
also been obtained by an enzymatic approach using microbial toward dextran conjugates in human subjects. However, Sep-
transglutaminase to attach dextran with hexylenediamine at pl et al. were able to generate antibodies against a conjugate
the reducing end to surface Gln residues of catalase.[32d] of 10 kDa dextran and chicken serum albumin (CSA) after im-
Similar to PEGylation, half-life extension by dextran conjuga- munization of mice, and the immune response of the conju-
tion is driven by the increase in hydrodynamic radius, which gate was stronger than the response obtained after immuniza-
decreases the renal clearance of the target compound. The tion with plain dextran.[40] Importantly, immunogenicity was
Figure 4. Dextran conjugation by A) periodate oxidation and subsequent reductive amination, and B) attachment of e-aminocaproic acid at the reducing end
followed by amide bond formation. Only monoconjugated product is shown; R: peptide or protein.

Reviews
most pronounced for CSA–dextran conjugates with small dex- creased half-life (from ~ 20 min to 3.6 h) in a H2O2 inactivation
trans (1 and 4 kDa, respectively), but was lower for a conjugate assay. As in the case of catalase, the PK profile of the SOD–dex-
containing a large 40 kDa dextran moiety.[41] tran conjugate was markedly improved over wild-type SOD.
A long-acting somatostatin–dextran conjugate has been pre- The conjugate showed a 150-fold longer elimination half-life
pared for cancer treatment. Using reductive amination, soma- (from 4.8 min to 12 h), a 9.7-fold increase in exposure, and
tostatin was coupled to a 70 kDa dextran moiety activated a 9.8-fold decrease in clearance (from 62.5 to 6.4 U h 1) follow-
using sodium periodate to yield a conjugate with a 4:1 soma- ing i.v. administration to rats. The improved PK profile and the
tostatin/dextran ratio. The somatostatin–dextran conjugate increased resistance toward H2O2 inactivation further translated
showed receptor binding affinity in the low nanomolar range, into improved pharmacological properties of the SOD–dextran
which was similar to octreotide (a somatostatin analogue). Fol- conjugate, as visualized by a carrageenan-induced paw edema
lowing s.c. administration to mice, the conjugate displayed an test.[32b]
extended PK profile with delayed absorption (maximal concen- Other examples include dextran conjugates of asparagi-
tration being reached 24 h post-injection) and prolonged half- nase,[45] carboxypeptidase G2,[34, 46] and hemoglobin,[47] respec-
life relative to octreotide (27 h for conjugate,[42] compared to tively, as well as CMD-insulin conjugates.[48]
less than 2 h in man for octreotide).[43] The conjugate has been
evaluated in phase I clinical trials and was found to be well-tol-
4.2 HESylation
erated.[44]
Villalonga and co-workers have reported the synthesis of Hydroxyethyl starch (HES) is a modified natural polymer ob-
catalase–dextran conjugates. Catalase is a tetrameric enzyme tained by controlled hydroxyethylation of the plant polysac-
important for neutralization of reactive oxygen species that charide amylopectin. Amylopectin is a polymer of d-glucose
have been associated with a variety of diseases. First, a cata- containing primarily a-1,4 glycosidic bonds, but also a lower
lase–dextran conjugate was prepared using a 5 kDa dextran abundance of a-1,6 linkages, leading to a naturally branched
with a hexyleneamine installed at the reducing end. Coupling carbohydrate (Figure 5).[49] Hydroxyethylation of the starch pre-
of dextran to Asp and Glu residues situated on the surface of cursor serves two purposes: first, to increase the water solubili-
catalase was performed at pH 6 under aqueous conditions and ty by increasing water-binding capacity and decreasing viscosi-
resulted in a product containing an average of four dextrans ty, and second, to prevent immediate degradation by plasma
per catalase molecule, corresponding to one dextran chain per a-amylase and subsequent renal excretion.[49] Hydroxyethyla-
catalase subunit. The specific activity of the catalase–dextran tion predominantly occurs at C2 and C6, and to a lesser extent
conjugate was slightly increased over unmodified catalase at C3.[50]
(from 14.6 U g 1 to 15.5 U g 1). Moreover, intravenous (i.v.) ad-
ministration of the conjugate to lean rats illustrated a markedly
improved PK profile, with a 20-fold longer elimination half-life
(from 0.8 to 16 h), a 171-fold increased exposure (from 1.8 to
307 U h mL 1), and a 176-fold decreased clearance (from 120 to
0.68 mL min 1).[32a]
Second, Villalonga and co-workers prepared a dextran–cata-
lase conjugate using a dextran variety activated with an e-ami-
nocaproic acid spacer. By using this spacer, site-specific inser-
tion of dextran at the N-terminus of each catalase subunit was
achieved. Although catalase activity was decreased by 23 %,
this catalase–dextran variant displayed improved thermostabili-
ty as well as decreased susceptibility to trypsin digestion (func-
tional half-life extension from 40 min to 2 h as compared with Figure 5. Example of HES architecture. Degree of cross-linking as well as the
native catalase). Intravenous administration of the dextran–cat- amount and location of hydroxyethyl substituents vary among different HES
preparations.
alase conjugate to rats revealed a seven-fold extension of half-
life relative to unmodified catalase (from 0.7 to 5.1 h) as well
as a 104-fold increase in exposure (from 1.9 to 197 U h mL 1), The first use of HES for pharmacological purposes dates
and a 100-fold decrease in clearance (from 110 to back to the 1970s, when this modified polysaccharide was in-
1.1 mL min 1).[32c] troduced by Fresenius Kabi as a plasma volume substitute.
Villalonga and co-workers have also reported the synthesis Commercially available HES products are characterized by
of a conjugate of superoxide dismutase (SOD) and dextran. three parameters: mean molecular weight (MMW); molar sub-
Like catalase, SOD is an enzymatic scavenger of reactive stitution (MS), calculated as the average number of hydroxy-
oxygen species, specifically the superoxide anion. The dextran ethyl groups per glucose unit; and the C2/C6 ratio, which de-
conjugate was prepared using hexyleneamine chemistry and scribes the substitution pattern at C2 relative to C6 (Table 2).
provided a SOD–dextran conjugate containing an average of HES preparations are labeled according to mean molecular
4.4 dextran units per SOD subunit. The conjugate retained weight and molar substitution (e.g., HES 450/0.7, which has
81 % of initial enzymatic activity, but showed a 10.8-fold in- MMW = 450 kDa and MS = 0.7) (Table 2).[50]

Reviews
The conjugation of HES moieties increases the hydrodynam-
Table 2. Common HES preparations and their parameters.
ic volume of the modified molecule, thereby making it less
Name MMW [kDa][a] MS[b] C2/C6 ratio susceptible to renal clearance.[49] The most relevant parameters
HES 450/0.7[156] 450 0.7 5:1 for describing the PK properties of the various HES polymers,
HES 200/0.62[49] 200 0.62 9:1 in addition to size, are molar substitution and C2/C6 ratio.
HES 200/0.5[157] 200 0.5 5:1 High molar substitution, reflecting a high degree of hydroxy-
HES 70/0.5[157] 70 0.5 3:1
ethylation, decreases enzymatic degradation by a-amylase via
HES 130/0.4[156] 130 0.4 9:1
HES 130/0.4[156] 130 0.42 6:1 steric factors, resulting in prolonged elimination half-life after
both single and multiple doses.[50] In addition, a high C2/C6
[a] MMW: mean molecular weight. [b] MS: molar substitution.
ratio further decreases susceptibility to enzymatic degradation.
This is especially pronounced for HES preparations of similar
Interestingly, correct choice of analytical method is essential molar substitution.[50, 57] However, it should be noted that HES
for accurate size determination of HES polymers. Using size ex- polymers with high MMW are more effective providers of steric
clusion chromatography coupled to low-angle laser light scat- shielding, as indicated by studies on HES-shielded polyplex-
tering (SEC-LALLS), Lederer et al. showed that commercially es.[54] HESylation has also been shown to increase protein sta-
available HES 450/0.7 has a MMW of ~ 740 kDa, corresponding bility by increasing the melting temperature and the enthalpy
to a 64 % increase from the nominal value determined by con- of melting as well as decreasing stress-induced aggregation.[56]
ventional light scattering.[51] HES is ultimately metabolized in vivo by a-amylase in
Chemical coupling of HES to proteins was originally explored plasma and excreted by the kidneys as smaller fragments. In-
in the 1970s, typically as a random conjugation approach intracellular degradation of HES by lysosomal enzymes such as
volving BrCN activation of available alcohols followed by ami- a-glucosidase has also been suggested,[58] but the exact mech-
nolysis to yield an isourea or carbamate linkage.[49, 52] However, anism remains unknown. Because of the close resemblance be-
the latter approach often results in cross-linked products, and tween the semisynthetic HES polymer and the natural glucose
random conjugation has now been superseded by site-specific polymers known to the human body (e.g., glycogen and
incorporation. This approach utilizes that the HES polymer only starch), HES is generally considered nonimmunogenic. This
contains one available hemiacetal (i.e., the reducing end), notion was emphasized by Dieterich et al., who investigated
which may be oxidized using iodine,[49] KMnO4,[53] or NaClO[53] antibody formation in 1004 patients in response to infusion of
to yield the terminal aldonic acid. The aldonic acid can be acti- various HES preparations.[59] Only one patient (corresponding
vated either by dehydration to produce the corresponding lac- to 0.1 %) developed low-titer IgM antibodies, corresponding to
tone or by formation of an active NHS-ester (Figure 6 A).[49] In a very low incidence. Importantly, the antibodies were specifi-
both cases, subsequent aminolysis results in the formation of cally directed toward the hydroxyethyl groups, with no ob-
a polypeptide–HES conjugate linked by an amide bond. served cross-reactivity toward amylose, a natural component
An alternative to oxidation exploits that hemiacetals are in of starch.[59] In rabbits, however, Richter & de Belder were able
equilibrium with the corresponding open-chain aldehyde, to generate high-titer anti-HES antibodies after immunization
which enables HESylation via reductive amination using with a bovine serum albumin–HES (BSA–HES) conjugate. No
NaBH3CN (Figure 6 B).[49, 54] The aldehyde may also be obtained cross-reactivity toward amylopectin or glycogen was observed
by oxidation of the hemiacetal with NaIO4.[55] Finally, the alde- for the anti-HES antibodies.[52]
hyde functionality may be introduced via a spacer at the re- As indicated by the previously mentioned studies by Ring &
ducing end.[56] Messmer and Laxenaire et al., the risk of developing anaphy-
Figure 6. Examples of HESylation chemistry. A) The hemiacetal of the reducing end of the polymer chain may be oxidized to yield the corresponding aldonic
acid, which then can be activated by dehydration to form the lactone or by formation of the NHS ester. B) The hemiacetal exists in an equilibrium with the
open-chain aldehyde, which enables HESylation via reductive amination. R: peptide or protein; TSTU: N,N,N’,N’-tetramethyl-O-(N-succinimidyl)uronium tetra-
fluoroborate; DIEA: N,N-diisopropylethylamine.

Reviews
lactoid reactions following HES infusion is very low (< 0.1 %) mer.[63] The conjugate was deemed equipotent to erythropoie-
and does not exceed the risk associated with dextran infu- tin (EPO), as reflected by similar half-maximal effective concen-
sion.[37, 38] tration (EC50) values in survival assays conducted in EPO-re-
Despite the effective enzymatic degradation of infused HES, sponsive cell lines.[63] Notably, the conjugate was also found to
transient bioaccumulation of the polymer is known to occur. In be equipotent to both EPO and Aranesp in stimulating eryth-
a recent meta-analysis of HES bioaccumulation conducted by ropoiesis from bone marrow cells in vitro.[63]
Wiedermann & Joannidis, several safety concerns were raised. A long-acting factor VIII (FVIII) for prophylactic treatment of
Based on data from 37 human studies (635 patients in total), hemophilia A was prepared by site-specific covalent attach-
the authors documented accumulation in a wide range of ment of HES to B-domain deleted rFVIII. Addition of up to
tissue types, including skin, kidney, liver, bone marrow, lymph 700 kDa HES provided FVIII–HES conjugates with fully pre-
nodes, spleen, lung, pancreas, intestine, muscle, trophoblasts, served in vitro activity and retained ability to bind von Wille-
as well as placental stroma. Tissue uptake was very rapid (as brand factor with high affinity.[64] Moreover, the FVIII–HES con-
early as 30 min after infusion), and storage increased in a dose- jugate exhibited acute efficacy comparable to rFVIII in a hemo-
dependent manner. HES deposits could be detected in the philia A mouse-tail bleeding model, while also demonstrating
skin and in the kidneys for as long as eight and ten years, re- prolonged exposure as a result of extended half-life.[64]
spectively, after administration. Other adverse events were pru- More recently, a long-acting HESylated version of anakinra
ritus and renal failure, and cases of liver dysfunction and bone (recombinant human interleukin 1 receptor antagonist, rhIL-
marrow suppression were also associated with HES administra- 1ra) was reported. Unmodified anakinra is approved for the
tion.[60] treatment of rheumatoid arthritis, but has a short-acting PK
HES has also been associated with negative effects on the profile. A HESylated analogue, anakinra–HES, was obtained by
blood coagulation system. In a study of 30 patients with cere- conjugation of 85 kDa HES at the N-terminus. The HES moiety
brovascular disease, administration of three different HES prep- carried a propionaldehyde spacer at the reducing end, en-
arations (HES 200/0.62, HES 200/0.5, HES 40/0.5, n = 10 for abling conjugation by reductive amination. Biophysical charac-
each group) for ten days resulted in a decrease in platelet terization of the anakinra–HES conjugate confirmed the ex-
volume, with the more substituted HES 200/0.62 causing the pected increase in molar mass (from 16.6 to 105.5 kDa) and hy-
largest decrease.[61] A further comparison of HES 200/0.62 and drodynamic radius (from 4.36 to 14.73 nm). The secondary
HES 200/0.5 in a ten-day hemodilution study revealed that in- structure of anakinra was unaffected by HESylation, and the
fusion of HES200/0.62 led to acquired von Willebrand syn- conjugate retained a dissociation constant (KD) in the nanomo-
drome as well as large decreases in FVIII function (72.2 %), von lar affinity range toward the interleukin-1 receptor, as shown
Willebrand ristocetin cofactor (61.3 %), and von Willebrand by surface plasmon resonance (SPR) and microscale thermo-
factor antigen (64 %). In contrast, no adverse effects were ob- phoresis (MST) experiments. In addition, differential scanning
served for HES 200/0.5. The discrepancy was attributed to the calorimetry (DSC) revealed that HESylation increased protein
slower degradation of HES 200/0.62, which upon repeated ad- stability, as illustrated by an increase in melting point (from
ministration leads to an in vivo build-up of high molecular 58.0 to 62.8 8C) and in enthalpy of melting (from 58.4 to
weight HES fragments and causes negative effects on blood 100.8 kcal mol 1). The conjugate exhibited no stress-induced
coagulation.[62] aggregation during a seven day shaking test, whereas aggre-
The first attempt to use HES for improving PK properties gation of unmodified anakinra was observed already after
dates back to 1989, when the polymer was used to prolong 30 min. Intravenous injection of anakinra–HES to rats revealed
the half-life of deferoxamine, a siderophore isolated from S. pi- a markedly improved PK profile of the conjugate, with a 45-
losus. The deferoxamine–HES conjugate was prepared by initial fold decrease in clearance (from 403.3 to 9.1 mL kg 1 h 1),
activation of HES (obtained from HES 450/0.7) with sodium pe- a 6.5-fold increase in half-life (from 1.7 to 10.8 h) and a 45-fold
riodate to provide the aldehyde, followed by reductive amina- increase in exposure (from 12.3 to 548.8 ng mL 1).[56]
tion with NaBH3CN to yield a substituted HES polymer contain-
ing 10–20 % (w/w) deferoxamine.[55] Intravenous administration
of the deferoxamine–HES conjugate to male Swiss-Webster
4.3 Polysialylation
mice showed that the half-life of the conjugate was increased
15-fold relative to unmodified deferoxamine (from 5.5 to Conjugation of polysialic acid (PSA) to drugs in order to im-
84 min), while the ability to induce urinary excretion of excess prove their pharmacokinetic properties (polysialylation) was
iron was fully maintained.[55] first suggested and explored by Gregoriadis and co-workers in
HES was also used to prepare a conjugate of AGE M400, 1993.[65] Strictly speaking, the term ‘sialic acid’ refers not to
a novel dimeric erythropoietin mimetic peptide (EMP) with op- a single chemical entity, but rather to an entire group of nine-
timized in vitro performance. The conjugate, named carbon monosaccharides, with the most important examples
AGEM400(HES), was assembled by reacting the free thiol of the being 5-N-acetylneuraminic acid (Neu5Ac), 5-N-glycolylneura-
C-terminal Cys of the AGEM400 sequence with a maleimide- minic acid (Neu5Gc), and 2-keto-3-deoxynonulosonic acid
functionalized HES polymer (MW = 130 20 kDa, obtained (Kdn).[66] However, the only observed PSA variant in humans is
from HES 200/0.5) under aqueous conditions to provide conju- colominic acid (CA), the linear a-2,8-linked homopolymer of
gates containing an average of five peptides per HES poly- Neu5Ac (Figure 7).[67]

Reviews
often site-specific, as free Cys residues occur at a low frequen-
cy in natural polypeptides or have been specifically engineered
into the peptide or protein.
In addition to these chemical methods, a chemo-enzymatic
method using the bacterial enzymes C. jejuni Cst-II sialyltrans-
ferase and N. meningitides a2,8-polysialyltransferase has also
Figure 7. Structure of colominic acid, the linear polymer of Neu5Ac used for
been described for the production of polysialylated proteins.[72]
polysialylation.
Similar to PEG and the PEG mimetics dextran and HES, the
driving force behind the improved PK properties of polysialy-
CA is a naturally occurring polysaccharide found mainly as lated compounds is thought to be an increase in hydrodynam-
an extracellular glycan of the neural cell adhesion molecule ic volume, resulting in decreased renal clearance as well as
(NCAM) protein, which is known to be very important in cell– shielding from enzymatic degradation and antibody recogni-
cell adhesion and synaptic plasticity. NMR studies have shown tion.[65, 71e] A notable physicochemical feature of PSA is the per-
that the ‘basal conformational unit’ of CA is an extended heli- manent negative charges. It may be speculated that this fea-
cal structure with n = 9.[66, 68] For pharmaceutical purposes, CA ture decreases glomerular filtration even further because of
is used almost exclusively. The technology is currently being electric repulsion by the negatively charged surfaces of glo-
commercialized as PolyXen by Xenetic Biosciences (formerly merular capillaries.[73]
Lipoxen PLC, www.xeneticbio.com), and several polysialylated It has been proposed that the increase in proteolytic stabili-
drug candidates are now in development, including EreproXen ty occurs by a different mechanism than observed with PEGyla-
(polysialylated EPO) in phase II/III clinical trials.[69] tion.[71a] Catalase is a very large protein with a molecular
For the production of PSA-drug conjugates, an activated weight of 240 kDa. It was observed that a catalase–PSA conju-
form of CA containing a terminal aldehyde moiety is used. This gate containing four 10 kDa CA units was less prone to degra-
species may be obtained by mild NaIO4 oxidation of the C7–C8 dation by trypsin and chymotrypsin than native catalase. The
diol system found in the non-reducing end of the polysaccha- negatively charged CA polymer may interact with positively
ride to yield an aldehyde functionality. Conjugation may then charged amino acids residues of the enzyme, thereby forming
be performed by nucleophilic attack of an amino group to an extensive hydrophilic layer able to shield against protea-
yield a Schiff base, followed by reductive amination under ses.[71a]
aqueous conditions using excess of oxidized colominic acid in Interestingly, PSA has been shown to act as a ‘lyo-protec-
the presence of NaBH3CN (Figure 8 A).[70] The reductive amina- tant’. A catalase–PSA conjugate was shown to retain 35–40 %
tion strategy has so far been the most widely used method for of its initial activity after lyophilization, compared with only
producing polysialylated drugs.[71] However, because peptides 20 % for the corresponding native enzyme.[71a] The exact mech-
and proteins often contain multiple amino groups in the form anism behind this observation is unknown, but it was speculat-
of N-terminal amino group(s) and Lys Ne-amines, this approach ed that during the freeze-drying process, the extreme hydro-
may lead to a mixture of products. philicity of the PSA polymer is able to maintain an aqueous mi-
Alternatively, polysialylation may also be obtained by the croenvironment around the enzyme necessary for maintaining
chemoselective maleimide–thiol conjugation reaction. A malei- the active conformation.[71a]
mide-activated PSA species can be prepared by reacting the al- In contrast to the synthetic PEG and semi-synthetic HES
dehyde functionality of oxidized PSA with N-[b-maleimidopro- polymers, PSAs are truly natural polymers. Colominic acid is
pionic acid] hydrazide to produce a maleimide-functionalized biodegradable and the ultimate degradation product, Neu5Ac,
PSA hydrazone. Subsequent conjugation to the thiol function- is not known to be toxic.[65] Moreover, PSAs are T-cell inde-
ality of an available Cys may then be performed under slightly pendent antigens, and immunization of mice with colominic
basic aqueous conditions (Figure 8 B).[71e] This approach is acid resulted in no detectable antibody formation, suggesting
Figure 8. Chemical introduction of PSA. A) The non-reducing end of CA may be oxidized to yield an aldehyde, which may then undergo reductive amination,
enabling conjugation via the N-terminal amino group(s) and Lys Ne amines. B) CA aldehyde may also be further modified by incorporation of a maleimide,
which may then be selectively attached to the thiol functionality of a free Cys. R: peptide or protein.

Reviews
that this PSA variant is nonimmunogenic.[65, 74] Immunization was only slightly decreased relative to unmodified Fab (2.7- to
with PSA-protein conjugates, however, have been shown to 3.7-fold or 1.3- to 3.4-fold, depending on the antibody used for
elicit an immune response, but in the case of colominic acid, ELISA detection).[71d] Intravenous administration of the Fab–
this only leads to generation of low-affinity antibodies that cir- PSA immunoconjugates to mice showed a 4.2- to 5.2-fold in-
culate at low concentration.[65] These circumstances all contrib- crease in bioavailability and a 1.8- to 3.3-fold increase in total
ute to create an attractive safety profile for PSA. tumor exposure compared to native Fab, with a higher degree
Early proof-of-concept for polysialylation was obtained when of polysialylation providing the biggest improvements.[71d]
Gregoriadis and co-workers showed that conjugation of Site-specific polysialylation of the single-chain variable frag-
a 24 kDa CA (82 Neu5Ac units) to fluorescein increased the ment (scFv) MFE-23, a well-described anti-carcinoembryonic
half-life of this otherwise rapidly excreted compound. The fluo- antigen, was obtained using the chemoselective maleimide–
rescein–PSA conjugate exhibited a half-life of 20 h following thiol conjugation reaction. An engineered version of MFE-23
i.v. administration to mice, corresponding to the half-life of the containing a terminal Cys residue was produced by recombi-
PSA polymer itself.[65] nant expression and polysialylated using a maleimide-function-
The concept was later expanded to also include modulation alized PSA hydrazine, which was produced from 11 kDa CA al-
of protein pharmacokinetics. In 1996, Fernandes & Gregoriadis dehyde and N-[b-maleimidopropionic acid] hydrazide.[71e] The
presented a modified version of catalase containing an average MFE-23-Cys–PSA immunoconjugate retained its antigen affinity
3.8 molecules of 10 kDa CA per molecule enzyme produced by compared with both native MFE-23 and MFE-23-Cys, and s.c.
the reductive amination approach. Compared with the native administration of MFE-23-Cys–PSA to mice revealed a 3.6-fold
enzyme, the catalase–PSA conjugate exhibited increased stabil- prolongation of elimination half-life (from 4.3 to 15.6 h) and an
ity toward degradation by the endopeptidases trypsin and chy- 9.8-fold increase in bioavailability compared to native MFE-23,
motrypsin (illustrated by 95 % retained activity of the conju- which was accompanied by an increase in tumor exposure and
gate compared with only 15 % retained activity of the native uptake.[71e]
enzyme after incubation with chymotrypsin for 3 h at 37 8C).[71a] Recently, a polysialylated version of the anorectic gut pep-
In addition, polysialylated versions of asparaginase were re- tide oxyntomodulin (Oxm) with improved protease resistance
ported with ~ 85 % retained enzymatic functionality and im- was published.[71f] The Na-substituted Oxm–PSA conjugate was
proved plasma stability relative to the native enzyme. The produced by reductive amination using a commercially avail-
lowest degree of polysialylation provided the largest increase able 14.1 kDa oxidized PSA.[71f] Oxm is a well-known substrate
in stability, with 87 % retained activity being observed for of for the enzyme dipeptidyl peptidase IV (DPP-4), which partly
a conjugate with an average of 4.2 CA units per asparaginase explains the very short in vivo half-life of native Oxm. Incuba-
molecule compared with only 18 % retained activity of the tion of the Oxm–PSA conjugate with DPP-4 at a concentration
native enzyme after 6 h of incubation at 37 8C in diluted of 2.9 U mL 1 over a period of 20 h revealed no enzymatic deg-
mouse plasma.[71b] Intravenous injection of asparaginase–PSA radation, suggesting that N-terminal polysialylation of Oxm
to mice revealed an improvement in functional and terminal confers complete resistance toward DPP-4.[71f]
half-life proportional to the degree of PSA conjugation, with Also recently, conjugates of PSA and butyrylcholinesterase
a conjugate containing an average of 8.1 CA units per aspara- (BChE), a bioscavenger useful for treatment of organophospho-
ginase molecule exhibiting a half-life of almost 40 h.[71b, 75] rus poisoning, was reported.[71g] A commercially available
Moreover, measurement of antibody titers following immuniza- 27 kDa mono-oxidized CA polymer containing a terminal
tion of mice with asparaginase–PSA conjugate suggested Neu5Ac unit with a C7-keto group (named CAO27) was conju-
a lower immunogenicity compared to the native enzyme.[75] gated to recombinant human BChE (rhBChE) by reductive ami-
Two polysialylated versions of insulin, modified with 22 and nation, resulting in a polysialylated product (rhBChE-CAO27)
39 kDa CA chains, respectively, have also been described.[71c] with a protein/PSA ratio of 1:6 and full catalytic activity relative
The insulin–PSA conjugates were assembled by the reductive to unmodified rhBChE.[71g] Intravenous administration of
amination approach using excess of CA aldehyde, yielding rhBChE-CAO27 to mice revealed a 6.3-fold increase in mean
compounds with PSA/insulin molar ratios of 1.60–1.74 and residence time (from 3.7 to 23.3 h) and a 5.6-fold increase in
2.37–2.45 for the 22 kDa and 39 kDa CA variant, respectively. elimination half-life (from 3 to 16.6 h) in comparison with un-
Subcutaneous administration of insulin–PSA to mice showed modified rhBChE.[71g]
that polysialylated insulin had a prolonged efficacy relative to A polysialylated version of the 52 kDa glycoprotein a-1-anti-
unmodified insulin, with the 22 kDa CA variant having a glu- trypsin (A1AT), a serine protease inhibitor targeting the
cose lowering effect for 9 h, compared with 6 h for the 39 kDa enzyme neutrophil elastase, has also been described.[72] Native
CA variant and 3 h for native insulin.[71c] A1AT is currently used as a therapeutic protein useful in treat-
Conjugation of PSA has also been used to modulate anti- ment of patients with congenital A1AT deficiency, a disorder
body pharmacokinetics. Polysialylation of a Fab antibody frag- associated with lung emphysema, chronic obstructive lung dis-
ment specific for the placental-like alkaline phosphatase (PLAP) ease (COPD), as well as other disorders.[76] Using their chemo-
antigen was performed by reductive amination using either enzymatic method (see above), Lindhout et al. attached six
11 kDa or 22 kDa periodate-oxidized CA, providing Fab–PSA PSA undecamers to the three available biantennary sialylation
immunoconjugates with a varying degree of molar substitu- sites of A1AT. The bioactivity of the resulting product (PSA-
tion.[71d] Antigen affinity of the obtained immunoconjugates A1AT) was examined using an enzyme activity assay measuring

Reviews
the inhibition of neutrophil elastase, with PSA-A1AT exhibiting long-term use in various medical applications, HA is regarded
retained bioactivity compared to native A1AT.[72] Intravenous as biocompatible, nontoxic, and nonimmunogenic.[85] The poly-
administration of PSA-A1AT to mice revealed an 18-fold greater mer is degraded in vivo by enzymes known as hyaluronidases
bioavailability (measured as area under the curve) and a 5.4- in a highly controlled manner, which also involves receptor-
fold increase in elimination half-life (from 5 h to 27 h).[72] Inter- mediated cellular uptake.[79] This controlled degradation pat-
estingly, a version of A1AT containing very short PSA oligomers tern has suggested a potential use of HA for drug delivery, pro-
with only 1–3 repeats of Neu5Ac (named diSA-A1AT) exhibited viding sustained release of the active component upon HA
an approximate 2-fold increase in bioactivity, a 6.5-fold in- degradation in vivo.[86]
crease in bioavailability as well as a 2.6-fold increase in elimina- Numerous options exist for chemical modification of HA,
tion half-life (from 5 to 13 h) relative to unmodified A1AT.[72] and these have been used for preparing both cross-linked HA
These data indicate that introduction of only very few sialic derivatives as well as several drug-HA conjugates.[79, 85] For the
acid residues may have dramatic effects on the pharmacokinet- purpose of preparing protein-HA conjugates, preferred meth-
ic properties of biopharmaceuticals. ods have been either direct coupling of protein amino groups
to the COOH of GlcA or partial periodate oxidation of 2-OH
and 3-OH of GlcA to yield the dialdehyde followed by reduc-
4.4 HAylation
tive amination.[78] However, both of these strategies are prone
As in the case of PSA, hyaluronic acid (HA) is a negatively to result in cross-linking of HA and the target protein and may
charged linear polysaccharide. HA consists of alternating d-glu- also lead to significant product heterogeneity.[78, 86a]
curonic acid (GlcA) and N-acetyl-d-glucosamine (GlcNAc) units To address this limitation, strategies to avoid cross-linking
(repeating disaccharide: [b(1-4)-GlcA-b(1-3)-GlcNAc]n (Figure 9)). and produce multivalent HA conjugates have been developed
It is a naturally occurring biopolymer found primarily in the ex- (Figure 10). Pasut and co-workers attached a 4-aminobutyral-
tracellular matrix of connective tissues, with highest abun- dehyde diethyl acetal as spacer to the COOH groups of GlcA
dance in the synovial fluid of joints, the dermis of the skin, and (Figure 10 A), creating 200 kDa HA polymers with varying
the vitreous humor of the eye.[77] The HA polymer serves multi- degree of acetal derivatization. The acetal may be hydrolyzed
ple physiological purposes, including tissue viscoelasticity and under mildly acidic conditions to yield the corresponding alde-
repair, control of tissue hydration, as well as cell proliferation, hyde, which may then be used for conjugation of protein
differentiation, and migration.[78] These features have prompted amino groups by reductive amination.[86a]
the use of HA for various medical applications such as visco- Initially, the HA-acetal was used to prepare multivalent HA
supplementation for treating joint disease, wound healing, conjugates of trypsin, RNase A, and insulin.[86a] By varying the
ophthalmic surgery, and cosmetic applications.[79] pH value, the conjugation process could be directed toward
Lys residues or be selective for the N-terminus. HA conjugation
increased the hydrodynamic radius of both trypsin and
RNase A, but also decreased the enzymatic activity relative to
the unmodified enzymes, possibly due to steric hindrance.
However, thermal stability was increased in the case of the N-
terminal trypsin-HA conjugate.[86a] HAylation of insulin provided
a conjugate with a prolonged and enhanced glucose-lowering
effect after administration to diabetic rats, suggesting that
Figure 9. Chemical structure of the repeating GlcA-GalNAc disaccharide that
makes up HA. HAylation of insulin increases the half-life and facilitates forma-
tion of a soluble depot.[86a]
Pasut and co-workers have subsequently employed the HA-
Under normal physiological conditions, a HA polymer con- acetal for synthesizing a conjugate of HA and salmon calcito-
sists of 20 000–25 000 disaccharide units, corresponding to nin (sCT-HA). The HA polymer was site-selectively attached to
a molar weight of up to 10 MDa and a length of 2–25 mm.[80] the N-terminus of sCT by reductive amination, and the result-
For conjugation chemistry, however, polymers of maximal 100– ing conjugate displayed an extended pharmacokinetic profile
300 kDa are used to avoid viscosity problems.[2] HA variants of and chondro-protective effects in the ALCT rabbit model of os-
appropriate size and quality for medical applications may be teoarthritis.[87] Also using the HA-acetal, the same authors pre-
obtained by fermentation using S. equi as expression host,[81] pared a conjugate of the interferon a2a (INFa2a) and HA with
or alternatively by degradation of high-MW HA polymers[82] or increased bioavailability due to a HAylation-induced depot
by chemoenzymatic synthesis using P. multocida hyaluronic effect and improved antitumor activity in an ovarian cancer
acid synthase (PmHAS).[83] xenograft mouse model.[88]
Conjugation of large HA polymers increases the hydrody- In another approach, Hahn and co-workers used a methacry-
namic radius of the modified therapeutic, thereby decreasing lated version of 200 kDa HA to obtain multivalent conjugates
renal clearance and prolonging plasma half-life. In addition, of HA and CWRYMVm, a peptide agonist of the formyl peptide
the hydrodynamic nature of HA combined with its ability to receptor like 1 (FPRL1) receptor.[89] Aminoethyl methacrylate
retain water can shield conjugated proteins from enzymatic HA (AMEA-HA) was synthesized by coupling AEMA to the
degradation.[80, 84] Due to its presence in the human body and COOH functionality of the GlcA residues of HA (Figure 10 B).

Reviews
Figure 10. Representative GlcA-GlcNAc disaccharides from HA derivatives used for controlled preparation of multivalent HA conjugates: A) HA acetal, B) HA
methacrylate, C) HA vinyl sulfone, and D) HA dialdehyde.
AMEA-HA was subsequently coupled to the Cys residue of formed under mildly acidic conditions to target only the N-ter-
CWRYMVm by Michael addition of the thiol to the methacry- minus of INFa. HAylation increased the hydrodynamic volume
late double bond. By changing the amount of AMEA used for while retaining the secondary structure of INFa, as indicated
preparing the AMEA-HA polymers, the average number of pep- by gel permeation chromatography (GPC) and circular dichro-
tides per HA conjugate could be controlled in the range from ism (CD) spectroscopy, respectively. Compared with unconju-
5 to 23. Compared to unmodified CWRYMVm, the multivalent gated INFa, a HA conjugate containing six INFa molecules ex-
HA conjugates had increased hydrodynamic volume and exhibited decreased biological activity in an anti-proliferation
hibited increased stability in diluted serum, with no degrada- Daudi cell assay, but displayed improved serum stability (half-
tion observed after 96 h of incubation. The multivalent life of > 120 h, corresponding to a 3-fold increase). Following
CWRYMVm-HA conjugates displayed partially decreased bio- i.v. administration to rats, INFa-HA exhibited prolonged circula-
logical activity in terms of FPR L1 receptor activation, but full tion and could be detected for 50–110 h (depending on the al-
activity was recovered upon treatment with hyaluronidases.[89] dehyde content of the polymer), whereas native INFa was
Using a related strategy, Hahn and co-workers have also pre- cleared within 24 h.[91]
pared multivalent conjugates of exendin-4 (Ex4) and HA using
a 100 kDa HA variant containing a vinyl sulfone functionality at
5. N- and O-Glycosylation
the COOH position (VS-HA) (Figure 10 C).[90] Ex4 was prepared
with an additional C-terminal Cys residues and efficiently con- Protein glycosylation is one of the most important and most
jugated to VS-HA by Michael addition under slightly alkaline complex post-translational modifications.[92] It has been esti-
conditions to yield multivalent Ex4-HA conjugates with a vary- mated that more than 50 % of all proteins may in fact be gly-
ing number of peptides per HA polymer. The conjugates have coproteins, emphasizing the ubiquity of glycosylation.[93] Glyco-
an increased hydrodynamic volume and a markedly prolonged sylation of peptides and proteins has many functions, includ-
half-life when incubated with human serum (half-life of > 96 h, ing modulation of protein physicochemical properties, control
corresponding to a 20-fold increase over unconjugated Ex4). of protein folding, stabilization of protein conformation, pro-
The in vivo efficacy of the Ex4-HA conjugates was evaluated in tein interactions and cellular recognition, protection from pro-
diabetic db/db mice. Compared to unmodified Ex4, the hypo- teolytic degradation, regulation of protein transport and locali-
glycemic effect of Ex4-HA was similar in an intraperitoneal glu- zation, modification of immunological properties, and media-
cose tolerance test (IPGTT) but prolonged following adminis- tion of protein-protein and cell–cell interactions.[94] Many of
tration of a single s.c. dose, with the effect being maintained these effects directly provide half-life extension of a glycosyla-
for up to three days.[90] ted peptide or protein.
Although the periodate oxidation strategy is associated with In N-glycosylation, a glycan is attached by a glycosidic
product heterogeneity, it has been used to prepare multivalent amide from GlcNAc to the side-chain of an Asn residue in an
conjugates of HA and interferon a (INFa) in a controlled fash- Asn-X-Ser/Thr consensus sequence, with X being any amino
ion.[91] Using NaIO4, Kahn and co-workers obtained a partly oxi- acid except Pro.[95] N-linked glycans share a common pentame-
dized 100 kDa HA polymer (Figure 10 D) and used it to prepare ric core motif (Man3GlcNAc2, where Man is mannose) and are
INFa-HA conjugates with varying INFa content (2–9 copies per derived from a common 14-mer oligosaccharide precursor
HA polymer). Conjugation by reductive amination was per- (Glc3Man9GlcNAc2, where Glc is glucose), which is further trim-

Reviews
med to provide three main architectures known as the high- using Ba/F3 cells, indicating correct disulfide bond formation
mannose, complex, and hybrid type oligosaccharide structu- and protein folding.[100]
res.[95a, 96]
In O-glycosylation, a glycan is linked by an O-glycosidic
6. Lipidation
bond to side-chain hydroxy groups, most commonly Ser or
Thr, but apparently all types of side-chain hydroxy groups can Lipidation of peptides and proteins is a well-known and impor-
be found glycosylated. The glycans in the widely occurring tant posttranslational modification, with examples including S-
mucin-type O-glycosylation are linked by an a-glycosidic bond prenylation (either farnesylation or geranylgeranylation), N-
from N-acetylgalactosamine (GalNAc) to the hydroxy group of myristoylation, S-palmitoylation, N-palmitoylation, O-octanoyla-
either Ser or Thr. Mucin-type glycans can have a wide variety tion, and cholesteroylation.[101] A special case is the glycosyl-
of structures.[97] phosphatidylinositol (GPI) anchor, a glycolipid which besides
In nature, glycoproteins are obtained by enzymatic co- or a phospholipid tail also contains a carbohydrate core struc-
post-translational modification of a polypeptide precursor. ture.[102] In nature, covalent attachment of lipids to peptides
Many glycosylated biopharmaceuticals are produced by re- and proteins is mediated by enzymes and serves to increase
combinant techniques using expression hosts that to some the hydrophobicity of the modified protein, often with the
extend can mimic the human glycosylation pattern.[94b] Howev- purpose of aiding anchoring to the cell membrane. Lipidation
er, glycosylated biopharmaceuticals are often obtained as a het- as post-translational modification has been described else-
erogeneous mixture of several glycoforms. More recently, where and is not discussed further.[103]
chemically well-defined glycosylated biopharmaceuticals have Acylation of peptides and proteins with long-chain fatty
also been obtained by synthetic or semisynthetic methods. acids has become a well-described and validated approach to
Kent and co-workers have prepared a monodisperse, polymer- improve the half-life of several native peptides and small pro-
modified, synthetic erythropoiesis protein (SEP) as a homogene- teins. The concept was originally explored by scientists at
ous alternative to the multiple glycoforms obtained by re- Novo Nordisk and Eli Lilly in the 1990s for the generation of
combinant EPO production. To mimic the physicochemical novel insulin analogues with a protracted mode of action,[104]
properties of wild-type EPO glycosylation, SEP contains two and the approach has since been adapted by others. For an
PEG-like, branched, negatively charged polymers installed at overview of selected lipid constructs used for half-life exten-
engineered Lys(Ne-levulinyl) residues using oxime ligation. SEP sion of biopharmaceuticals, please see Figure 11.
was constructed from four peptide fragments synthesized by Lipidation of biopharmaceuticals is typically obtained by N-
solid-phase peptide synthesis (SPPS). The full SEP sequence acylation with a fatty acid, either directly or via a spacer, to
was assembled using native chemical ligation (NCL) and exhib- produce a stable amide bond. While S-prenylation and S-palmi-
ited full hematopoietic activity in mice. Compared with re- toylation of peptides and proteins are abundant in nature and
combinant EPO, the half-life of SEP was increased approxi- may also be obtained by means of chemical synthesis,[101b]
mately 2-fold (from 5.1 to 9.5 h) following i.v. administration to these lipidation forms are rarely used in the context of bio-
rats. This benefit was attributed to the use of PEG-like poly- pharmaceuticals. S-prenylation is not known to facilitate half-
mers instead of native glycosylation.[98] life extension, and the thioester linkage obtained by S-palmi-
Danishefsky and co-workers synthesized human EPO as toylation is less stable under physiological conditions.
a single, native-like glycoform. The glycoform of choice had Importantly, the conditions for lipidation by N-acylation
three N-linked chitobiose disaccharides at Asn24, Asn38, and depend on whether it is a peptide or protein that is being
Asn83, as well as an O-linked glycan at Ser126. The 166-residue modified. If the target is a protein, lipidation can almost exclu-
EPO sequence was prepared as four glycosylated fragments sively be performed in aqueous solution because of the insta-
synthesized by solid-phase methods and assembled by itera- bility of most proteins toward organic solvents. However, the
tive ligations. Following protein folding, the fully glycosylated patent literature describes methods for lipidation of coagula-
EPO exhibited clear erythropoietic activity in a cell proliferation tion factor FVIIa (406 AA, 50 kDa) using hydroxypropyl-b-cyclo-
assay using cord blood CD34 + cells.[99] dextrin (HPCD) as a lipid solubilizer under aqueous condi-
The semisynthesis of the immunomodulatory glycoprotein tions,[105] as well as lipidation of human growth hormone (hGH,
interleukin-6 (IL6) was recently reported by Unverzagt and co- 190 AA, 22 kDa) by enzymatic approaches.[106]
workers. The 183-residue IL6 protein was prepared from three If the target is a peptide or a small protein like insulin
peptide fragments: a short glycopeptide thioester (correspond- (51 AA, 5.8 kDa), lipidation may take place in solution using
ing to AAs 43–48) prepared by chemical synthesis, and a thio- polar aprotic solvents, as peptides are often stable to organic
ester fragment (AAs 1–42) and a Cys-peptide fragment (AAs conditions. A successful approach has been to employ NHS-ac-
49–183) obtained by recombinant methods using intein splic- tivated fatty acid constructs for acylation of available amino
ing and cleavage of a SUMO tag, respectively. Two N-glycosy- groups.[107] However, this approach often requires a two-step
lated versions of IL6 (containing either GalNAc or a biantennary HPLC purification process, which decreases the overall yield of
nonasaccharide at Asn44) were assembled by NCL-based reaction.[107b]
methods using a hydrazide as a latent thioester for fragment For peptides synthesized by SPPS, lipidation can also be per-
43–48. Relative to recombinant IL6, both semisynthetic IL6 ver- formed in organic solvents while the peptide is attached to
sions displayed full in vitro bioactivity in a proliferation assay the solid support. Typically, the point of lipidation will be

Reviews
Figure 11. Examples of lipid moieties used for half-life extension of biopharmaceuticals. A) Myristic acid, which is used for acylation of the LysB29 position of
insulin detemir. B) Palmitic acid linked to a g-glutamic acid (gGlu) spacer, which is inserted at the e-amine of Lys26 position of liraglutide. C) Hexadecanoic
diacid with a gGlu spacer, used for modification of the LysB29 position of insulin degludec. D) Octadecanoic diacid coupled to a spacer consisting of gGlu and
two 8-amino-3,6-dioxaoctanoic acid (OEG) units, used for acylating position Lys26 of semaglutide.
either the N-terminus or a Lys residue carrying an orthogonal chloride-mediated denaturation of insulin detemir (Levemir,
protecting group. In Fmoc-SPPS, this could be for example, al- Novo Nordisk) showed that myristoylation increases the free
lyloxycarbonyl (Alloc),[107b] 2-acetyl-5,5-dimethyl-1,3-cyclohexa- energy of unfolding by 30 %.[113]
nedionyl (Dde),[108] or (4-methoxyphenyl)diphenylmethyl Lipidation with dietary fatty acids such as myristic acid and
(Mmt).[109] During assembly of the peptide chain, the orthogo- palmitic acid is generally perceived as a safe approach to half-
nal protecting group may be selectively removed to reveal an life extension, as these compounds are known to be complete-
attachment point for anchoring of the lipid moiety. By per- ly metabolized by the human body. Whether this is also true
forming the lipidation on-resin, purification in one step is usu- for the structurally similar fatty diacids remains unknown. Gen-
ally sufficient to achieve a pure product.[107b] eration of antibodies toward lipidated biopharmaceuticals has
The ability of fatty acid conjugation to improve the half-life been reported,[114] but so far the level of antibody formation is
of biopharmaceuticals relies on noncovalent binding to albu- low and without clinical relevance. The long-term safety profile
min. Human serum albumin (HSA) is the most abundant pro- associated with more elaborate lipid motifs containing PEG-
tein of the bloodstream and contains nine different fatty acid based spacers is currently unknown.
binding sites, which may bind free fatty acids as well as fatty The first lipidated polypeptide to obtain regulatory approval
acids conjugated to larger molecules.[110] When bound to HSA, was insulin detemir in 2004. Insulin detemir is a basal insulin
lipidated peptides are sterically shielded from proteolytic deg- for treatment of type 1 diabetes (T1D). It is based on desB30
radation, and because of the large size of HSA (MW of human insulin and is derivatized with myristic acid (C14) at the
~ 66 kDa), this protein, along with any molecule associated Ne-amine of LysB29 (Figure 11 A). It has a half-life of 5–7 h fol-
with it, is protected from rapid glomerular filtration. lowing s.c. administration to healthy human subjects.[115] The
Following binding to serum albumin, lipidated biopharma- mechanism underlying the half-life extension involves both al-
ceuticals are released gradually into the circulation, thereby bumin binding (more than 95 % of circulating insulin detemir
obtaining a protracted half-life and an extended mode of is albumin-bound)[111, 112] as well as self-association. Insulin dete-
action. Early studies on fatty acid-modified insulin analogues, mir is co-formulated with phenol and Zn2 + to induce forma-
which include insulin detemir, revealed that half-life was corre- tion of insulin hexamers, but size exclusion chromatography
lated with the affinity for HSA, such that strong albumin affini- (SEC) studies have revealed that insulin detemir exists in a sta-
ty resulted in prolonged action of these analogues.[111] bilizing hexamer–dihexamer equilibrium in the s.c. depot,
Conjugation with fatty acids has in several cases been which delays the absorption. Once in the bloodstream, the
shown to elicit in vivo self-assembly of drug monomers into hexamers dissociate into monomers, which are then able to
stable oligomeric structures.[112] Such oligomers are less suscep- bind to albumin.[112a] The X-ray crystal structure indicates that
tible to proteolytic degradation and may also form a depot dihexamer formation is mediated by hydrophobic interactions
from which the active monomer is gradually released. In some of the myristic acid groups.[113]
cases, lipidation also increases the chemical stability of the In 2009, the once-daily GLP-1 analogue liraglutide (Victoza,
modified protein. For instance, a study of guanidinium hydro- Novo Nordisk) was marketed for the treatment of type 2 diabe-

Reviews
tes (T2D). In 2014, it was also approved for treatment of obesi- vitro functional assays using BHK cells.[109] In humans, semaglu-
ty under the brand name Saxenda. The peptide backbone of tide has a half-life of ~ 160 h following s.c. injection.[122] This re-
liraglutide is similar to native GLP-1 with the exception of the sults from very high albumin affinity (5.6-fold greater than lira-
Arg residue at position 34 (Lys in native GLP-1). Liraglutide is glutide as assessed by analytical ultracentrifugation under
a full agonist of the GLP-1 receptor (GLP-1R) and equipotent to near-physiological conditions) as well as resistance toward
native GLP-1, as shown by in vitro functional assays using baby DPP-4.[109] Whether semaglutide also forms oligomers under
hamster kidney (BHK) cells.[107a, 116] Lipidation with palmitic acid physiological conditions has not yet been reported.
(C16) is obtained by acylation of the Ly s26 position via a short Somapacitan (NNC0195-0092, Novo Nordisk), a long-acting
g-glutamic acid (gGlu) spacer, which is inserted to assist albu- lipidated version of human growth hormone (hGH), is in late-
min binding and to increase compound solubility (Fig- stage clinical development for the treatment of GHD. Plasma
ure 11 B).[107a] This results in a very high degree of albumin protein binding of somapacitan was investigated using SPR
binding (~ 99 %), as indicated by equilibrium dialysis studies.[117] biosensing and revealed that > 99 % of somapacitan was
Analytical ultracentrifugation studies have revealed that liraglu- bound to plasma proteins, with albumin as the primary binder
tide forms a heptamer under aqueous conditions. The ability in human plasma.[123] The chemical structure of somapacitan
of liraglutide to undergo self-assembly is thought to delay ab- has not been publicly disclosed at the time of writing.
sorption following s.c. injection and thereby contribute to the The lipidation motif used for liraglutide (C16-gGlu-) has been
markedly prolonged half-life of liraglutide (11–15 h)[118] com- used successfully for half-life extension of other biopharma-
pared to native GLP-1 (1–1.5 h).[119] ceuticals in preclinical development. Bellmann-Sickert et al.
Insulin degludec (Tresiba, Novo Nordisk) is a long-acting have reported that lipidation of radiolabelled pancreatic poly-
once-daily basal insulin approved by the European Medicines peptide with gGlu-spaced palmitic acid at Lys13 provided a 7-
Agency (EMA) in 2013 and FDA in 2015 for treatment of T1D. fold increase in half residence time after i.v. infusion to Wistar
It contains a gGlu-spaced hexadecanoic diacid at position rats.[108] Also, Dalbøge et al. reported that lipidation of neuro-
LysB29 (Figure 11 C). In addition to albumin binding, the ex- medin U (NMU) with C16-gGlu- at either the N-terminus or an
tended half-life of insulin degludec relies on in vivo oligomeri- engineered Lys at position nine resulted in a 15-fold prolonga-
zation. Insulin degludec is formulated as a dihexamer using tion of in vitro plasma stability at 37 8C.[107b]
Zn2 + and phenol as excipients. Following injection, rapid diffu- Other examples of lipidated biopharmaceuticals in various
sion of phenol away from the injection site results in a confor- stages of development include a N-palmitoylated GLP-1/glu-
mational change of the dihexamer (from intermediate T3R3 cose-dependent insulinotropic polypeptide co-agonist,[124] N-
conformation to the exposed T6 conformation; T ‘tense’, R ‘re- palmitoylated analogues of prolactin-releasing peptide, a long-
laxed’), leading to the formation of a s.c. depot containing acting N-acylated (OEG-gGlu-spaced octadecanoic diacid) ana-
large chain-like multihexamers (MW > 5 MDa), as indicated by logue of calcitonin,[125] a cholesteroylated analogue of oxynto-
SEC and dynamic light scattering (DLS).[112b] Interestingly, the modulin obtained by S-alkylation of a free Cys with cholesterol
presence of a fatty diacid seems to be critical for multihexamer via a short PEG-based spacer,[126] and insulin-327, a liver-specific
formation. Using transmission electron microscopy (TEM), it insulin analogue obtained by N-acylation at LysA22 with
was shown that the architecture of insulin degludec multihex- a OEG-gGlu-spaced icosanoic diacid.[127]
amers resembles ‘pearls on a string’,[120] with the individual Very recently, van Witteloostuijn et al. combined lipids and
hexamers being anchored to each other through the interac- carbohydrates in a series of novel neoglycolipid conjugates.
tion between Zn2 + and the terminal carboxylic acid of hexade- They consist of an open-chain carbohydrate, mucic acid, and
canoic diacid.[112b] Subsequent passive diffusion of Zn2 + releas- fatty acids of varying length. The impact of neoglycolipidation
es the bioactive monomer, which is then absorbed into the on peptide properties was evaluated from a library of neogly-
bloodstream.[112b] As a result of both albumin affinity and multi- colipidated GLP-1 analogues. Biophysical studies revealed that
hexamer formation, insulin degludec exhibits a half-life of 25 h neoglycolipidation mediated albumin binding and directed for-
in humans, with the glucose-lowering effect being preserved mation of soluble peptide oligomers of variable size, depend-
for more than 42 h after administration, as indicated by eugly- ing on the chemical composition of the neoglycolipid. More-
cemic clamp studies in patients with diabetes.[120, 121] over, the neoglycolipidated GLP-1 analogues had maintained
Semaglutide (Novo Nordisk) is a once-weekly GLP-1 analogue or even improved in vitro potency relative to native GLP-1.
in phase III clinical development for treatment of T2D and obe- This translated into potent in vivo efficacy, manifested as de-
sity. It is based on the same peptide backbone as liraglutide, creased food intake and improved glucose tolerance following
with the exception of the Ala residue at position 8, which has s.c. administration to lean mice.[128]
been changed to the nonproteinogenic amino acid 2-aminoi- Besides lipidation, a variety of other techniques also use
sobutyric acid (Aib) in order to confer resistance toward enzy- noncovalent albumin binding for half-life extension. Examples
matic degradation by DPP-4. Semaglutide is lipidated at Lys26 include albumin binding domains (ABDs),[129] albumin binding
with octadecanoic diacid using a spacer consisting of gGlu and peptides,[130] albumin-binding domain antibodies (Albu-
two 8-amino-3,6-dioxaoctanoic acid (OEG) units, representing dAbs),[131] Nanobodies,[132] and small-molecule albumin binders
one of the most optimized lipid moieties currently used for bi- such as dicoumarol.[133] An overview is provided by Sleep
opharmaceuticals (Figure 11 D).[109] Semaglutide is equipotent et al.[134]
to native GLP-1 in terms of GLP-1R activation as shown by in

Reviews
7. Fusion Proteins Using FcRn-Mediated 8. Other Chemical Modifications—Insulin as

Recycling Example
Compared with other plasma proteins, albumin and immuno- Numerous other chemical entities have also been covalently
globulin G (IgG) exhibit a remarkably long half-life of approxi- anchored to peptides and proteins. Most have seen limited if
mately three weeks.[135] This is due in part to their large size any use in biopharmaceutical development or have only been
(66 kDa and 150 kDa, respectively), which inhibits renal clear- reported recently. Instead of providing a full overview, we have
ance, but also due to a unique receptor-mediated pH-con- selected a few examples of chemically modified insulin ana-
trolled recycling process, which makes these proteins less sus- logues.
ceptible to intracellular degradation. The receptor responsible As mentioned, chemical lipidation of insulin at LysB29 has
is known as the neonatal Fc receptor (FcRn). Its presence was been a very successful strategy for developing long-acting in-
first suggested by Brambell[136] and later confirmed upon isola- sulin variants such as insulin detemir and insulin degludec.
tion by Simister & Rees.[137] FcRn was first recognized for its Upon complexation with two Zn2 + ions, insulin readily forms
role in IgG recycling, but it was later discovered that FcRn also well-defined hexamers, which can occur in two states. The oli-
salvages albumin.[138] FcRn is situated in the cell membrane of gomeric state of new insulin variants, that is, whether they are
immune cells and binds and internalizes IgG (via the Fc preferably monomers/dimers, hexamers or multimers of hex-
domain) as well as albumin under acidic conditions. Binding to amers, determines whether they are fast-acting or slow-acting
FcRn ensures that IgG and albumin escapes lysosomal degra- in the treatment of diabetes. Insulin has only one Lys (B29),
dation and are redirected to the cell surface, where the pro- and hence only one free amine besides the two N-terminals.
teins dissociate from FcRn at physiological pH and are released Ligands may be attached by pH-controlled regioselective acyla-
back into circulation.[135] tion, as the pKa of the Ne-amine of LysB29 is ~ 11.2 whereas it is
The FcRn-mediated recycling mechanism has been used to 8.4 and 7.1 for the Na-amine of the A and B-chain, respective-
create several half-life extended conjugates of various thera- ly.[149]
peutic proteins with albumin or the Fc portion of IgG.[139] Ex-
amples include etanercept (Enbrel, TNFR75-Fc, Wyeth/
8.1 Bile acid derivatives
Amgen),[140] dulaglutide (Trulicity, GLP-1-Fc, Eli Lilly),[141] and al-
biglutide (Tanzeum/Eperzan, GLP-1-albumin, GlaxoSmithK- In a collaboration between Jonassen et al. at Novo Nordisk and
line).[142] For the development of albiglutide, an engineered al- Dodson and co-workers, several different bile acids (cholic acid
bumin variant with improved FcRn binding (Veltis, Albumedix/ and derivatives) were anchored to the LysB29 position of
Novozymes) was used.[143] desB30 human insulin and studied.[150] Insulin desB30 acylated
Although all approved albumin and Fc conjugates are ob- in position LysB29 with derivatives of cholic acid displayed
tained as recombinant fusion proteins, there are examples of a wide range of affinities for HSA, but serum albumin affinity
albumin conjugates that have been obtained by chemical con- made only a minor contribution to the long elimination times
jugation. Using maleimide–thiol chemistry, a NMU-HSA conju- observed for these analogues. The most promising analogue,
gate was prepared by coupling of a iodoacetamide-functional- LysB29-(N-(lithocholyl-g-Glu)) des(B30) human insulin (NN344),
ized NMU variant to the free Cys34 residue on the surface of showed glucose-lowering activity lasting more than 24 h in
albumin. The conjugate exhibited prolonged in vivo efficacy in pigs. NN344 forms a stable dodecameric zinc complex in the
diet-induced obese (DIO) mice and had a predicted half-life in presence of phenol, but becomes a high-MW structure com-
humans of approximately 3–4 days.[11c] posed of insulin zinc hexamers in the absence of phenol. With-
In humans, the half-life extension obtained by conjugation/ out phenol, the apparent molecular mass was > 5000 kDa. The
fusion to albumin and Fc typically amounts to several days, ex- formation of a large aggregate after injection and depot-for-
emplified by for example, etanercept (2–4 days),[144] dulaglutide mation may have contributed to the long duration of action of
(3–4 days),[145] and albiglutide (6–8 days).[146] NN334. A low but measurable affinity for albumin of the litho-
Despite albumin’s human origin, antibodies may develop cholic acid ligand may also have contributed to the prolonged
toward albumin and Fc fusion proteins. In a phase II study of action. The crystal structure revealed that the cholic acid
albiglutide in T2D patients, 5 out of 356 patients (1.4 %) devel- moiety served as a hook, connecting the insulin hexamers.[150b]
oped anti-drug antibodies, although these were mainly transi- However, NN344 appears not to have been further developed
ent, low titer, and non-neutralizing.[147] In a population of pa- in clinical trials.
tients with rheumatoid arthritis, 16 % (corresponding to 80 pa-
tients) developed antibodies toward etanercept at some point
8.2 Perfluoroalkylation
during a five year follow-up period.[148] However, in a phase I
study of dulaglutide containing 20 healthy subjects, no anti- Jensen and co-workers used a bioorthogonal strategy to ex-
drug antibodies were observed.[145] plore the self-segregating behavior of perfluoroalkyl moieties
for intermolecular protein self-assembly.[151] Highly fluorinated
molecules tend to segregate into a fluorous phase, a phenom-
enon, which is often referred to as the fluorous effect.[152] At-
taching a perfluoroalkyl group on the surface of a protein

Reviews
could potentially direct intermolecular fluorous self-assembly, 9. Summary and Outlook
representing a bio-orthogonal approach to control the oligo-
meric state of proteins in solution. Six new insulin variants car- Today, 40 years after the first work on protein PEGylation was
rying perfluoroalkyl moieties were prepared by regioselective published, the attachment of PEG remains the most successful
chemical acylation of LysB29 in desB30 human insulin at high method for chemically modifying peptides and proteins to
pH.[151] The perfluoroalkyl chain length was varied systematical- protract their half-life. As of today, 12 PEGylated peptide- and
ly to control the level of fluorous interactions between the per- protein-based biopharmaceuticals are approved in the USA
fluoroalkyl insulins. The six new insulin variants maintained sat- and Europe, which clearly outnumbers any other chemical
isfactory levels of binding to the insulin receptor. Relatively technique for half-life prolongation. However, the bioincom-
short perfluoroalkyl chains did not influence the overall quater- patibility of the large PEG polymers remains an issue. No cellu-
nary structure of insulin, whereas the longer perfluoroalkyl lar machinery is able to fully degrade PEG into metabolically
chains provided increased degrees of self-association of insulin useful constituents and, although PEG polymers can be excret-
hexamers. With even longer perfluoroalkyl chain length, ed by the kidneys and in feces, there are indications of PEG im-
a breakdown of the hexameric substructure was observed in munogenicity and bioaccumulation. This is especially concern-
favor of a new structure, apparently with local cylindrical ge- ing given that many PEGylated biopharmaceuticals are ap-
ometry. This new cylindrical geometry was observed for insu- proved for chronic treatment and thus are administered over
lins with C7F15, and C8F17 chains, and it was observed that the long periods of time. The current concerns associated with
increased molecular attraction introduced by the longer per- chronic use should not be neglected, and the search for bio-
fluoroalkyl chains also gave rise to a significant increase in the compatible PEG mimetics is an important effort toward mini-
overall length of the cylindrical structures.[151] Thus, it seems mizing the risks of long-term toxicity of biopharmaceuticals.
that with short perfluoroalkyl appendices the hexameric subu- Of the current carbohydrate-based alternatives to PEGyla-
nit core is still conserved as the dominant structure, whereas tion, dextran conjugation and HESylation resembles PEGylation
with longer perfluoroalkyl appendices it breaks down and the the most. Both techniques decreases renal excretion by in-
fluorous interactions start to override and promote intermolec- creasing the hydrodynamic radius of the target molecule.
ular self-assembly leading to larger structures. Moreover, dextran and HES carry no charge, making them
structurally similar to PEG. In contrast, the PSA and HA poly-
mers carry multiple negative charges, which confer increased
water solubility to the biopharmaceuticals and further decrease
8.3 Bipyridyl
their renal clearance by electric repulsion in addition to steric
Jensen and co-workers attached the metal ion binding ligand factors. This makes PSA and HA attractive for modification of
2,2’-bipyridyl (bipy) to LysB29 on insulins. Bipy has a high affin- for example, hydrophobic peptides with low intrinsic solubility.
ity for FeII and forms a chiral tris-bipy FeII complex, which is A similar advantage could be expected from CMD and other
colored. FeII forms a six-coordinated octahedral complex with charged dextran derivatives.
three bipyridine units and gives rise to a charge-transfer band The glucose or glucose-based constituents of dextran, HES,
at 550 nm, which enabled monitoring of the coordination reac- and HA exhibit greater chemical stability than the Neu5Ac
tion. Besides inducing self-assembly, the combination of bipyri- units of CA chains, and the options for conjugation chemistry
dine and FeII also acts as a molecular probe, as the color is also more diverse for these three polymers than for PSA. The
change from colorless to magenta is specific for formation of apparent lack of reported methods for obtaining monovalent
the trimeric compound. Bipy was first attached to an insulin protein-HA conjugates (i.e., with a 1:1 molar ratio of protein
variant, insulin X2, which has a low propensity to form the and polymer) is a potential drawback for the use of this tech-
hexamer and is mainly in the monomer/dimer form.[153] The nology in a biopharmaceutical context. However, the degree of
self-assembly through FeII chelation was reversible and provid- HA substitution can be regulated by controlled reaction condi-
ed the first insulin trimer reported in the literature as well as tions to yield multivalent protein-HA conjugates with a well-
the first insulin variant where assembly/disassembly could be defined degree of substitution.[86a, 89, 91] Also, as in the case of
followed by visual inspection. This study gave the first indica- dextran and HES, it should be possible to selectively modify
tion that the use of a metal ion binding ligand was possible. the reducing end of HA, thereby providing an opportunity for
Next, bipy was anchored to human insulin. This conjugate is ‘true’ (i.e., 1:1) site-specific HA conjugation.
able to lower blood glucose upon i.v. injection. It forms higher- One advantage of PEG over the carbohydrate-based poly-
ordered structures, combining native Zn2 + complexation with mers is that molecular properties of synthetic polymers are
abiotic Fe2 + binding. Insulin 18-mers were observed by small more easily controlled than those of polymers obtained from
angle X-ray scattering in solution, while 54-mers were ob- natural sources. Despite optimized production methods, poly-
served on surfaces by atomic force microscopy (AFM).[154] This mer heterogeneity is a challenge for dextran, HES, PSA, and
is a significant step toward construction of novel peptide and HA. From a drug development perspective, this is a potential
protein drugs, which change their oligomeric state in response problem because the molecular structure of drug products
to metal ion concentration. While the half-life extension of the must be very well-defined. Efforts to further optimize produc-
two latter strategies was not tested, it can be speculated that tion of carbohydrate polymers may still be needed to ensure
they have the potential for this.

Reviews
maximal homogeneity, thereby strengthening the case for [1] G. Walsh, Nat. Biotechnol. 2014, 32, 992 – 1000.
these biocompatible polymers. [2] G. Pasut, Polymers 2014, 6, 160 – 178.
[3] R. Kontermann in Therapeutic Peptides: Strategies to Modulate Their
While carbohydrate-based PEG mimetics are often claimed Plasma Half-Lives (Ed.: R. Kontermann), Wiley-Blackwell, Weinheim,
as nontoxic alternatives to PEG, the safety of these molecules 2012, pp. 3 – 21.
may be questioned. Generation of antibodies is only rarely ob- [4] A. Chanan-Khan, J. Szebeni, S. Savay, L. Liebes, N. M. Rafique, C. R.
served in the clinic, but bioaccumulation of dextran and HES is Alving, F. M. Muggia, Ann. Oncol. 2003, 14, 1430 – 1437.
[5] a) R. P. Garay, R. El-Gewely, J. K. Armstrong, G. Garratty, P. Richette,
known to occur, and this has been associated with unwanted Expert Opin. Drug Delivery 2012, 9, 1319 – 1323; b) N. J. Ganson, S. J.
side effects (see above). Bioaccumulation of PSA and HA is not Kelly, E. Scarlett, J. S. Sundy, M. S. Hershfield, Arthritis Res. Ther. 2006, 8,
known to occur, emphasizing the safety of these molecules. R12; c) M. S. Hershfield, N. J. Ganson, S. J. Kelly, E. L. Scarlett, D. A. Jag-
Despite substantial effort and strong preclinical evidence, no gers, J. S. Sundy, Arthritis Res. Ther. 2014, 16, R63.
[6] T. Yamaoka, Y. Tabata, Y. Ikada, J. Pharm. Sci. 1994, 83, 601 – 606.
biopharmaceutical containing a carbohydrate-based PEG mim- [7] a) J. M. Harris, R. B. Chess, Nat. Rev. Drug Discovery 2003, 2, 214 – 221;
etic has yet made it to the market. In contrast, several lipidated b) F. M. Veronese, Biomaterials 2001, 22, 405 – 417; c) G. Pasut, F. M. Ver-
peptide-based biopharmaceuticals have obtained regulatory onese, J. Controlled Release 2012, 161, 461 – 472; d) S. Jevševar, M. Kun-
approval. Early examples used dietary fatty acids, which are in- stelj, V. G. Porekar, Biotechnol. J. 2010, 5, 113 – 128.
[8] a) A. Abuchowski, T. van Es, N. C. Palczuk, F. F. Davis, J. Biol. Chem.
expensive and highly biocompatible. The use of fatty diacids 1977, 252, 3578 – 3581; b) A. Abuchowski, J. R. McCoy, N. C. Palczuk, T.
has provided drug products with optimized properties, but Van Es, F. F. Davis, J. Biol. Chem. 1977, 252, 3582 – 3586.
this often requires extra synthetic effort, and the long-term [9] a) A. Grigoletto, K. Maso, A. Mero, A. Rosato, O. Schiavon, G. Pasut, J.
safety of this kind of lipids is yet unknown. Lipidation increases Drug Delivery Sci. Technol. 2016, 32, 132 – 141; b) FDA approves modi-
fied antihemophilic factor for hemophilia A, FDA News Release:
the overall hydrophobicity of the target biopharmaceutical, www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/
and this may result in solubility problems. This issue may be ucm472643.htm (accessed October 12, 2016).
addressed by inserting a hydrophilic spacer between the lipid [10] M. J. Roberts, M. D. Bentley, J. M. Harris, Adv. Drug Delivery Rev. 2002,
and the target biopharmaceutical, as in the case of the gGlu 54, 459 – 476.
[11] a) A. D. Baldwin, K. L. Kiick, Bioconjugate Chem. 2011, 22, 1946 – 1953;
moiety used in liraglutide, insulin degludec, and semaglutide. b) B. Q. Shen, K. Y. Xu, L. N. Liu, H. Raab, S. Bhakta, M. Kenrick, K. L. Par-
As indicated by a growing number of publications, the use sons-Reponte, J. Tien, S. F. Yu, E. Mai, D. W. Li, J. Tibbitts, J. Baudys,
of lipidation for protraction of peptides is increasing within O. M. Saadi, S. J. Scales, P. J. McDonald, P. E. Hass, C. Eigenbrot, T.
drug discovery, and the methodology is also expanding into Nguyen, W. A. Solis, R. N. Fuji, K. M. Flagella, D. Patel, S. D. Spencer,
L. A. Khawlil, A. Ebens, W. L. Wong, R. Vandlen, S. Kaur, M. X. Sliwkow-
the protein field. Lipidation is a safe and efficient way of pro- ski, R. H. Scheller, P. Polakis, J. R. Junutula, Nat. Biotechnol. 2012, 30,
longing the half-life of biopharmaceuticals, and it may super- 184 – 189; c) P. Neuner, A. M. Peier, F. Talamo, P. Ingallinella, A. Lahm, G.
sede PEGylation as the preferred method for half-life extension. Barbato, A. Di Marco, K. Desai, K. Zytko, Y. Qian, X. Du, D. Ricci, E. Mon-
Lipid constructs are becoming more diverse in terms of both teagudo, R. Laufer, A. Pocai, E. Bianchi, D. J. Marsh, A. Pessi, J. Pept. Sci.
2014, 20, 7 – 19.
composition and function, illustrated by the development [12] a) S. C. Alley, D. R. Benjamin, S. C. Jeffrey, N. M. Okeley, D. L. Meyer, R. J.
from insulin detemir to semaglutide. Moreover, lipidation may Sanderson, P. D. Senter, Bioconjugate Chem. 2008, 19, 759 – 765; b) D.
also hold potential for targeted delivery, exemplified by the Lin, S. Saleh, D. C. Liebler, Chem. Res. Toxicol. 2008, 21, 2361 – 2369.
central delivery of liraglutide following peripheral administra- [13] a) K. Knop, R. Hoogenboom, D. Fischer, U. S. Schubert, Angew. Chem.
Int. Ed. 2010, 49, 6288 – 6308; Angew. Chem. 2010, 122, 6430 – 6452;
tion[155] and the liver-specificity of insulin-327.[127] b) H. Schellekens, W. E. Hennink, V. Brinks, Pharm. Res. 2013, 30, 1729 –
Designing and developing lipidated peptides is not 1734.
a straightforward task. In the development of semaglutide, [14] U. Binder, A. Skerra in Therapeutic Proteins: Strategies to Modulate Their
many related analogues exhibited an extended PK profile, but Plasma Half-Lives (Ed.: R. Kontermann), Wiley-Blackwell, Weinheim,
2012, pp. 63 – 80.
only a few withheld the high potency of native GLP-1.[109] Also, [15] V. Schellenberger, C. W. Wang, N. C. Geething, B. J. Spink, A. Campbell,
solubility problems may arise when introducing hydrophobic W. To, M. D. Scholle, Y. Yin, Y. Yao, O. Bogin, J. L. Cleland, J. Silverman,
lipid constructs, as illustrated by the lipidated NMU ana- W. P. Stemmer, Nat. Biotechnol. 2009, 27, 1186 – 1190.
logues.[107b] Consequently, there is no ‘one size fits all’ solution [16] a) J. L. Cleland, N. C. Geething, J. A. Moore, B. C. Rogers, B. J. Spink,
C. W. Wang, S. E. Alters, W. P. C. Stemmer, V. Schellenberger, J. Pharm.
for lipidation of biopharmaceuticals. For each target peptide or
Sci. 2012, 101, 2744 – 2754; b) K. C. J. Yuen, G. S. Conway, V. Popovic,
protein, a significant effort is required to confirm the optimal G. R. Merriam, T. Bailey, A. H. Hamrahian, B. M. K. Biller, M. Kipnes, J. A.
position for lipidation and to develop the optimal lipid con- Moore, E. Humphriss, G. M. Bright, J. L. Cleland, J. Clin. Endocrinol.
struct in terms of composition and length of both the spacer Metab. 2013, 98, 2595 – 2603.
[17] V. N. Podust, B. C. Sim, D. Kothari, L. Henthorn, C. Gu, C. W. Wang, B.
and the lipid moiety.
McLaughlin, V. Schellenberger, Protein Eng. Des. Sel. 2013, 26, 743 –
Anchoring chemical groups that enable controlled self-as- 753.
sembly of peptides and proteins by bioorthogonal principles [18] S. Ding, M. Song, B. C. Sim, C. Gu, V. N. Podust, C. W. Wang, B.
hold great promises for precise control of the oligomeric state McLaughlin, T. P. Shah, R. Lax, R. Gast, R. Sharan, A. Vasek, M. A. Hart-
man, C. Deniston, P. Srinivas, V. Schellenberger, Bioconjugate Chem.
of biopharmaceutical drug candidates. However, the biocom-
2014, 25, 1351 – 1359.
patibility of these moieties needs to be considered. [19] a) H. Betre, W. Liu, M. R. Zalutsky, A. Chilkoti, V. B. Kraus, L. A. Setton, J.
Controlled Release 2006, 115, 175 – 182; b) M. F. Shamji, H. Betre, V. B.
Kraus, J. Chen, A. Chilkoti, R. Pichika, K. Masuda, L. A. Setton, Arthritis
Keywords: biopharmaceuticals · drug delivery · half-life Rheum. 2007, 56, 3650 – 3661; c) M. Amiram, K. M. Luginbuhl, X. Li,
extension · peptides · proteins M. N. Feinglos, A. Chilkoti, J. Controlled Release 2013, 172, 144 – 151;
d) M. Amiram, K. M. Luginbuhl, X. H. Li, M. N. Feinglos, A. Chilkoti, Proc.
Natl. Acad. Sci. USA 2013, 110, 2792 – 2797.

Reviews
[20] Y. S. Huang, X. F. Wen, Y. L. Wu, Y. F. Wang, M. Fan, Z. Y. Yang, W. Liu, [46] R. G. Melton, C. N. Wiblin, A. Baskerville, R. L. Foster, R. F. Sherwood,
L. F. Zhou, Eur. J. Pharm. Biopharm. 2010, 74, 435 – 441. Biochem. Pharmacol. 1987, 36, 113 – 121.
[21] a) M. Schlapschy, U. Binder, C. Borger, I. Theobald, K. Wachinger, S. Kis- [47] S. C. Tam, J. Blumenstein, J. Tzefeiwong, Proc. Natl. Acad. Sci. USA
ling, D. Haller, A. Skerra, Protein Eng. Des. Sel. 2013, 26, 489 – 501; b) V. 1976, 73, 2128 – 2131.
Morath, F. Bolze, M. Schlapschy, S. Schneider, F. Sedlmayer, K. Seyfarth, [48] M. Baudyš, D. Letourneur, F. Liu, D. Mix, J. Jozefonvicz, S. W. Kim, Bio-
M. Klingenspor, A. Skerra, Mol. Pharm. 2015, 12, 1431 – 1442. conjugate Chem. 1998, 9, 176 – 183.
[22] A. E. Vasilev, G. F. Zhukova, G. A. Ravdel, R. G. L. Shchukin, Zh. Obshch. [49] T. Hey, H. Knoller, P. Vorstheim in Therapeutic Proteins: Strategies to
Khim. 1973, 43, 2529 – 2532. Modulate Their Plasma Half-Lives (Ed.: R. Kontermann), Wiley-Blackwell,
[23] C. Larsen, Adv. Drug Delivery Rev. 1989, 3, 103 – 154. Weinheim, 2012, pp. 117 – 140.
[24] R. Mehvar, J. Controlled Release 2000, 69, 1 – 25. [50] C. Jungheinrich, T. A. Neff, Clin. Pharmacokinet. 2005, 44, 681 – 699.
[25] a) Y. Takakura, T. Fujita, M. Hashida, H. Sezaki, Pharm. Res. 1990, 7, [51] K. Lederer, C. Huber, M. Dunky, J. K. Fink, H. P. Ferber, E. Nitsch, Arz-
339 – 346; b) T. Fujita, M. Nishikawa, C. Tamaki, Y. Takakura, M. Hashida, neim.-Forsch. 1985, 35, 610 – 614.
H. Sezaki, J. Pharmacol. Exp. Ther. 1992, 263, 971 – 978; c) M. Nishikawa, [52] A. W. Richter, A. B. de Belder, Int. Arch. Allergy Appl. Immunol. 1976, 52,
A. Kamijo, T. Fujita, Y. Takakura, H. Sezaki, M. Hashida, Pharm. Res. 307 – 314.
1993, 10, 1253 – 1261; d) M. C. Degat, G. Dubreucq, A. Meunier, L. [53] M. A. Abbas, S. Hameed, J. Kressler, Starch/Staerke 2013, 65, 264 – 272.
Dahri-Correia, L. Sedel, H. Petite, D. Logeart-Avramoglou, J. Biomed. [54] M. Noga, D. Edinger, R. Klager, S. V. Wegner, J. P. Spatz, E. Wagner, G.
Mater. Res. Part A 2009, 88A, 174 – 183; e) D. Logeart-Avramoglou, R. Winter, A. Besheer, Biomaterials 2013, 34, 2530 – 2538.
Huynh, F. Chaubet, L. Sedel, A. Meunier, Biochem. Pharmacol. 2002, 63, [55] P. E. Hallaway, J. W. Eaton, S. S. Panter, B. E. Hedlund, Proc. Natl. Acad.
129 – 137. Sci. USA 1989, 86, 10108 – 10112.
[26] R. Axn, J. Porath, S. Ernback, Nature 1967, 214, 1302 – 1304. [56] R. Liebner, R. Mathaes, M. Meyer, T. Hey, G. Winter, A. Besheer, Eur. J.
[27] J. F. Kennedy, J. A. Barnes, J. B. Matthews, Br. Polym. J. 1983, 15, 133 – Pharm. Biopharm. 2014, 87, 378 – 385.
138. [57] S. Schramm, C. Thyes, P. Frascarolo, T. Buclin, M. Burki, A. Fisch, M.-A.
[28] a) G. M. Lindenbaum, O. A. Mirgorodskaya, B. V. Moskvichev, Khim.- Burmeister, L. Asmis, D. R. Spahn, Anesthesiology 2007, 107, 442 – 451.
Farm. Zh. 1977, 11, 80 – 83; b) O. A. Mirgorodskaya, L. V. Poletaeva, [58] P. Kiehl, D. Metze, H. Kresse, S. Reimann, D. Kraft, A. Kapp, Br. J. Derma-
Khim.-Farm. Zh. 1985, 19, 594 – 599. tol. 1998, 138, 672 – 677.
[29] a) V. P. Torchilin, E. G. Tischenko, V. N. Smirnov, E. I. Chazov, J. Biomed. [59] H.-J. Dieterich, D. Kraft, C. Sirtl, H. Laubenthal, W. Schimetta, W. Pçlz, E.
Mater. Res. 1977, 11, 223 – 235; b) V. P. Torchilin, E. V. Il’lna, A. V. Mazaev, Gerlach, K. Peter, Anesth. Analg. 1998, 86, 1123 – 1126.
B. S. Lebedev, V. N. Smirnov, E. I. Chazov, J. Solid-Phase Biochem. 1977, [60] C. J. Wiedermann, M. Joannidis, Intensive Care Med. 2014, 40, 160 – 170.
2, 187 – 193. [61] J. Treib, A. Haass, G. Pindur, W. Treib, E. Wenzel, K. Schimrigk, Eur. J.
[30] Y. Takakura, Y. Kaneko, T. Fujita, M. Hashida, H. Maeda, H. Sezaki, J. Haematol. 1996, 56, 168 – 172.
Pharm. Sci. 1989, 78, 117 – 121. [62] J. Treib, A. Haass, G. Pindur, M. T. Grauer, E. Wenzel, K. Schimrigk, Trans-
[31] Y. Yasuda, T. Fujita, Y. Takakura, M. Hashida, H. Sezaki, Chem. Pharm. fusion 1996, 36, 450 – 455.
Bull. 1990, 38, 2053 – 2056. [63] A. Greindl, C. Kessler, B. Breuer, U. Haberl, A. Rybka, M. Emgenbroich,
[32] a) Y. Prez, A. Valdivia, H. L. Ramrez, R. Villalonga, Macromol. Rapid A. J. G. Pçtgens, H.-G. Frank, Open Hematol. J. 2010, 4, 1 – 14.
Commun. 2005, 26, 1304 – 1308; b) Y. Perez, A. Valdivia, L. Gomez, B. K. [64] J. Sommer, D. Sim, L. Leong, F. Hacket, B. Subramanyan, J. Haaning,
Simpson, R. Villalonga, Macromol. Biosci. 2005, 5, 1220 – 1225; c) R. Vil- Haemophilia 2010, 16, 39.
lalonga, A. Valdivia, Y. Perez, B. Chongo, J. Appl. Polym. Sci. 2006, 102, [65] G. Gregoriadis, B. McCormack, Z. Wang, R. Lifely, FEBS Lett. 1993, 315,
4573 – 4576; d) A. Valdivia, R. Villalonga, P. Di Pierro, Y. Perez, L. Mari- 271 – 276.
niello, L. Gomez, R. Porta, J. Biotechnol. 2006, 122, 326 – 333. [66] M. Mhlenhoff, M. Eckhardt, R. Gerardy-Schahn, Curr. Opin. Struct. Biol.
[33] a) R. L. S. Chang, I. F. Ueki, J. L. Troy, W. M. Deen, C. R. Robertson, B. M. 1998, 8, 558 – 564.
Brenner, Biophys. J. 1975, 15, 887 – 906; b) R. Mehvar, T. L. Shepard, J. [67] A. Constantinou, C. Chen, M. P. Deonarain in Therapeutic Peptides:
Pharm. Sci. 1992, 81, 908 – 912. Strategies to Modulate Their Plasma Half-Lives (Ed.: R. Kontermann),
[34] R. G. Melton, C. N. Wiblin, R. L. Foster, R. F. Sherwood, Biochem. Phar- Wiley-Blackwell, Weinheim, 2012, pp. 95 – 115.
macol. 1987, 36, 105 – 112. [68] H. Baumann, J.-R. Brisson, F. Michon, R. Pon, H. J. Jennings, Biochem-
[35] R. Fagnani, S. Halpern, M. Hagan, Nucl. Med. Commun. 1995, 16, 362 – istry 1993, 32, 4007 – 4013.
369. [69] PolyXen Technology-Based Candidates, Xenetic Biosciences: www.xene-
[36] Select Committee on GRAS Substances (SCOGS) Opinion: Dextran, US ticbio.com/product-pipeline/polyxen (accessed October 12, 2016).
Food and Drug Administration: www.fda.gov/Food/IngredientsPacka- [70] H. J. Jennings, C. Lugowski, J. Immunol. 1981, 127, 1012 – 1018.
gingLabeling/GRAS/SCOGS/ucm261271.htm (accesssed October 12, [71] a) A. Fernandes, G. Gregoriadis, Biochim. Biophys. Acta Protein Struct.
2016). Mol. Enzymol. 1996, 1293, 90 – 96; b) A. Fernandes, G. Gregoriadis, Bio-
[37] J. Ring, K. Messmer, Lancet 1977, 309, 466 – 469. chim. Biophys. Acta Protein Struct. Mol. Enzymol. 1997, 1341, 26 – 34;
[38] M. C. Laxenaire, C. Charpentier, L. Feldman, Ann. Fr. Anesth. Reanim. c) S. Jain, D. H. Hreczuk-Hirst, B. McCormack, M. Mital, A. Epenetos, P.
1994, 13, 301 – 310. Laing, G. Gregoriadis, Biochim. Biophys. Acta Gen. Subj. 2003, 1622, 42 –
[39] a) H. Hedin, W. Richter, J. Ring, Int. Arch. Allergy Appl. Immunol. 1976, 49; d) A. Constantinou, A. A. Epenetos, D. H. Hreczuk-Hirst, S. Jain, M. P.
52, 145 – 159; b) D. Kraft, H. Hedin, W. Richter, O. Scheiner, H. Rumpold, Deonarain, Bioconjugate Chem. 2008, 19, 643 – 650; e) A. Constantinou,
M. E. Devey, Allergy 1982, 37, 481 – 489. A. A. Epenetos, D. H. Hreczuk-Hirst, S. Jain, M. Wright, K. A. Chester,
[40] I. Seppl, J. Pelkonen, O. Mkel, Eur. J. Immunol. 1985, 15, 827 – 833. M. P. Deonarain, Bioconjugate Chem. 2009, 20, 924 – 931; f) I. Vorobiev,
[41] I. Seppl, O. Mkel, J. Immunol. 1989, 143, 1259 – 1264. V. Matskevich, S. Kovnir, N. Orlova, V. Knorre, S. Jain, D. Genkin, A. Ga-
[42] M. Behe, J. Du, W. Becker, T. Behr, C. Angerstein, M. Mrquez, J. Hiltu- bibov, A. Miroshnikov, Biochimie 2013, 95, 264 – 270; g) D. G. Ilyushin,
nen, S. Nilsson, A. R. Holmberg, Med. Oncol. 2001, 18, 59 – 64. I. V. Smirnov, A. A. Belogurov, Jr., I. A. Dyachenko, T. Zharmukhamedo-
[43] I. Lancranjan, C. Bruns, P. Grass, P. Jaquet, J. Jervell, P. Kendall-Taylor, va, T. I. Novozhilova, E. A. Bychikhin, M. V. Serebryakova, O. N. Kharybin,
S. W. J. Lamberts, P. Marbach, H. Ørskov, G. Pagani, M. Sheppard, L. A. N. Murashev, K. A. Anikienko, E. N. Nikolaev, N. A. Ponomarenko,
Simionescu, Metabolism 1995, 44, 18 – 26. D. D. Genkin, G. M. Blackburn, P. Masson, A. G. Gabibov, Proc. Natl.
[44] T. K. Joensuu, S. Nilsson, A. R. Holmberg, M. Marquez, M. Tenhunen, T. Acad. Sci. USA 2013, 110, 1243 – 1248.
Saarto, H. Joensuu in Signal Transduction and Communication in [72] T. Lindhout, U. Iqbal, L. M. Willis, A. N. Reid, J. Li, X. Liu, M. Moreno,
Cancer Cells, Vol. 1028 (Eds.: H. L. Bradlow, L. Castagnetta, L. Massimo, W. W. Wakarchuk, Proc. Natl. Acad. Sci. USA 2011, 108, 7397 – 7402.
K. Zaenker), New York Academy of Sciences, New York, 2004, pp. 361 – [73] a) G. Gregoriadis, A. Fernandes, M. Mital, B. McCormack, Cell. Mol. Life
374. Sci. 2000, 57, 1964 – 1969; b) B. M. Brenner, C. Baylis, W. M. Deen, Physi-
[45] a) T. E. Wileman, R. L. Foster, P. N. C. Elliott, J. Pharm. Pharmacol. 1986, ol. Rev. 1976, 56, 502 – 534.
38, 264 – 271; b) T. E. Wileman, Adv. Drug Delivery Rev. 1991, 6, 167 – [74] C. Moreno, M. R. Lifely, J. Esdaile, Infect. Immun. 1985, 47, 527 – 533.
180. [75] A. Fernandes, G. Gregoriadis, Int. J. Pharm. 2001, 217, 215 – 224.

Reviews
[76] E. Kelly, C. M. Greene, T. P. Carroll, N. G. McElvaney, S. J. O’Neill, Respir. [105] C. Behrens, P. W. Garibay, P. W. Garibay, H. S. Andersen, N. L. Johansen,
Med. 2010, 104, 763 – 772. B. Peschke, S. Bak (Novo Nordisk A/S), Int. PCT Pub. No.
[77] T. C. Laurent, J. R. Fraser, Faseb J. 1992, 6, 2397 – 2404. WO2010015668 A1, 2010.
[78] A. Mero, M. Campisi, Polymers 2014, 6, 346 – 369. [106] N. L. Johansen, H. S. Andersen, J. Buchardt, C. Behrens, L. Norskov-
[79] C. E. Schant, G. Zuber, C. Herlin, T. F. Vandamme, Carbohydr. Polym. Lauritsen, J. Su (Novo Nordisk Health Care AG), Int. PCT Pub. No.
2011, 85, 469 – 489. WO2011015649 A1, 2011.
[80] B. P. Toole, Nat. Rev. Cancer 2004, 4, 528 – 539. [107] a) L. B. Knudsen, P. F. Nielsen, P. O. Huusfeldt, N. L. Johansen, K.
[81] V. Rangaswamy, D. Jain, Biotechnol. Lett. 2008, 30, 493 – 496. Madsen, F. Z. Pedersen, H. Thøgersen, M. Wilken, H. Agersøe, J. Med.
[82] R. Stern, G. Kogan, M. J. Jedrzejas, L. Soltes, Biotechnol. Adv. 2007, 25, Chem. 2000, 43, 1664 – 1669; b) L. S. Dalbøge, S. L. Pedersen, S. B. van
537 – 557. Witteloostuijn, J. E. Rasmussen, K. T. Rigbolt, K. J. Jensen, B. Holst, N.
[83] P. L. DeAngelis, Curr. Pharm. Biotechnol. 2008, 9, 246 – 248. Vrang, J. Jelsing, J. Pept. Sci. 2015, 21, 85 – 94; c) C. Christensen, R. Se-
[84] A. J. Day, J. K. Sheehan, Curr. Opin. Struct. Biol. 2001, 11, 617 – 622. verinsen, A. K. Petersen, S. Kidal, C. U. Jessen, P. Madsen, H. Valore,
[85] E. J. Oh, K. Park, K. S. Kim, J. Kim, J. A. Yang, J. H. Kong, M. Y. Lee, A. S. T. M. Tagmose, J. L. Sorensen (Novo Nordisk A/S), Int. PCT Pub. No.
Hoffman, S. K. Hahn, J. Controlled Release 2010, 141, 2 – 12. WO2010029159 A1, 2010.
[86] a) A. Mero, M. Pasqualin, M. Campisi, D. Renier, G. Pasut, Carbohydr. [108] K. Bellmann-Sickert, C. E. Elling, A. N. Madsen, P. B. Little, K. Lundgren,
Polym. 2013, 92, 2163 – 2170; b) S. J. Kim, S. K. Hahn, M. J. Kim, D. H. L. O. Gerlach, R. Bergmann, B. Holst, T. W. Schwartz, A. G. Beck-Sicking-
Kim, Y. P. Lee, J. Controlled Release 2005, 104, 323 – 335. er, J. Med. Chem. 2011, 54, 2658 – 2667.
[87] A. Mero, M. Campisi, M. Favero, C. Barbera, C. Secchieri, J. M. Dayer, [109] J. Lau, P. Bloch, L. Schaffer, I. Pettersson, J. Spetzler, J. Kofoed, K.
M. B. Goldring, S. R. Goldring, G. Pasut, J. Controlled Release 2014, 187, Madsen, L. B. Knudsen, J. McGuire, D. B. Steensgaard, H. M. Strauss,
30 – 38. D. X. Gram, S. M. Knudsen, F. S. Nielsen, P. Thygesen, S. Reedtz-Runge,
[88] I. M. Montagner, A. Merlo, D. Carpanese, A. Dalla Piet, A. Mero, A. Gri- T. Kruse, J. Med. Chem. 2015, 58, 7370 – 7380.
goletto, A. Loregian, D. Renier, M. Campisi, P. Zanovello, G. Pasut, A. [110] G. Fanali, A. di Masi, V. Trezza, M. Marino, M. Fasano, P. Ascenzi, Mol.
Rosato, J. Controlled Release 2016, 236, 79 – 89. Aspects Med. 2012, 33, 209 – 290.
[89] E. J. Oh, J. W. Kim, J. H. Kong, S. H. Ryu, S. K. Hahn, Bioconjugate Chem. [111] J. Markussen, S. Havelund, P. Kurtzhals, A. S. Andersen, J. Halstrom, E.
2008, 19, 2401 – 2408. Hasselager, U. D. Larsen, U. Ribel, L. Schaffer, K. Vad, I. Jonassen, Diabe-
[90] J. H. Kong, E. J. Oh, S. Y. Chae, K. C. Lee, S. K. Hahn, Biomaterials 2010, tologia 1996, 39, 281 – 288.
31, 4121 – 4128. [112] a) S. Havelund, A. Plum, U. Ribel, I. Jonassen, A. Vølund, J. Markussen,
P. Kurtzhals, Pharm. Res. 2004, 21, 1498 – 1504; b) I. Jonassen, S. Have-
[91] J. A. Yang, K. Park, H. Jung, H. Kim, S. W. Hong, S. K. Yoon, S. K. Hahn,
lund, T. Hoeg-Jensen, D. B. Steensgaard, P. O. Wahlund, U. Ribel, Pharm.
Biomaterials 2011, 32, 8722 – 8729.
Res. 2012, 29, 2104 – 2014; c) D. B. Steensgaard, J. K. Thomsen, H. B.
[92] F. Fares in Therapeutic Proteins: Strategies to Modulate Their Plasma
Olsen, L. B. Knudsen, Diabetes 2008, 57, A164.
Half-Lives (Ed.: R. Kontermann), Wiley-Blackwell, Weinheim, 2012,
[113] H. B. Olsen, N. C. Kaarsholm, Biochemistry 2000, 39, 11893 – 11900.
pp. 81 – 94.
[114] a) J. B. Buse, A. Garber, J. Rosenstock, W. E. Schmidt, J. H. Brett, N. Vide-
[93] R. Apweiler, H. Hermjakob, N. Sharon, Biochim. Biophys. Acta Gen. Subj.
baek, J. Holst, M. Nauck, J. Clin. Endocrinol. Metab. 2011, 96, 1695 –
1999, 1473, 4 – 8.
1702; b) J. Vora, P. Hollander, S. C. Tamer, T. Johansen, R. Bergenstal, Di-
[94] a) H. Lis, N. Sharon, Eur. J. Biochem. 1993, 218, 1 – 27; b) H. J. Li, M.
abetologia 2012, 55, S53 – S54.
d’Anjou, Curr. Opin. Biotechnol. 2009, 20, 678 – 684.
[115] G. A. Brunner, G. Sendlhofer, A. Wutte, M. Ellmerer, B. Søgaard, A. Sie-
[95] a) B. Imperiali, S. E. O’Connor, Curr. Opin. Chem. Biol. 1999, 3, 643 – 649;
benhofer, S. Hirschberger, G. J. Krejs, T. R. Pieber, Exp. Clin. Endocrinol.
b) R. Kornfeld, S. Kornfeld, Annu. Rev. Biochem. 1985, 54, 631 – 664.
Diabetes 2000, 108, 100 – 105.
[96] G. Opdenakker, P. M. Rudd, C. P. Ponting, R. A. Dwek, Faseb J. 1993, 7,
[116] K. Madsen, L. B. Knudsen, H. Agersoe, P. F. Nielsen, H. Thøgersen, M.
1330 – 1337.
Wilken, N. L. Johansen, J. Med. Chem. 2007, 50, 6126 – 6132.
[97] P. Van den Steen, P. M. Rudd, R. A. Dwek, G. Opdenakker, Crit. Rev. Bio-
[117] A. Plum, L. B. Jensen, J. B. Kristensen, J. Pharm. Sci. 2013, 102, 2882 –
chem. Mol. Biol. 1998, 33, 151 – 208.
2888.
[98] a) G. G. Kochendoerfer, S. Y. Chen, F. Mao, S. Cressman, S. Traviglia, H. Y.
[118] B. Elbrønd, G. Jakobsen, S. Larsen, H. Agersø, L. B. Jensen, P. Rolan, J.
Shao, C. L. Hunter, D. W. Low, E. N. Cagle, M. Carnevali, V. Gueriguian, Sturis, V. Hatorp, M. Zdravkovic, Diabetes Care 2002, 25, 1398 – 1404.
P. J. Keogh, H. Porter, S. M. Stratton, M. C. Wiedeke, J. Wilken, J. Tang, [119] M. K. Gutniak, B. Linde, J. J. Holst, S. Efendic, Diabetes Care 1994, 17,
J. J. Levy, L. P. Miranda, M. M. Crnogorac, S. Kalbag, P. Botti, J. Schin- 1039 – 1044.
dler-Horvat, L. Savatski, J. W. Adamson, A. Kung, S. B. H. Kent, J. A. [120] P. Kurtzhals, T. Heise, H. M. Strauss, S. G. Bottcher, C. Granhall, H. Haahr,
Bradburne, Science 2003, 299, 884 – 887; b) S. B. H. Kent, Angew. Chem. I. Jonassen, Diabetologia 2011, 54, S426.
Int. Ed. 2013, 52, 11988 – 11996; Angew. Chem. 2013, 125, 12208 – [121] a) T. Heise, U. Hovelmann, L. Nosek, S. G. Bottcher, C. Granhall, H.
12217. Haahr, Diabetologia 2011, 54, S425; b) T. Heise, L. Nosek, S. G. Bottcher,
[99] P. Wang, S. W. Dong, J. A. Brailsford, K. Iyer, S. D. Townsend, Q. Zhang, H. Hastrup, H. Haahr, Diabetes Obes. Metab. 2012, 14, 944 – 950.
R. C. Hendrickson, J. Shieh, M. A. S. Moore, S. J. Danishefsky, Angew. [122] a) Abstracts of the 48th EASD Annual Meeting of the European Associa-
Chem. Int. Ed. 2012, 51, 11576 – 11584; Angew. Chem. 2012, 124, tion for the Study of Diabetes: M. A. Nauck, J. R. Petrie, G. Sesti, E. Man-
11744 – 11752. nucci, J. P. Courreges, S. Atkin, M. During, C. B. Jensen, S. Heller, Diabe-
[100] A. Reif, S. Siebenhaar, A. Troster, M. Schmalzlein, C. Lechner, P. Velisetty, tologia 2012, 55 (Suppl. 1), S7; b) C. Kapitza, L. Nosek, L. Jensen, H.
K. Gottwald, C. Pohner, I. Boos, V. Schubert, S. Rose-John, C. Unverzagt, Hartvig, C. B. Jensen, A. Flint, J. Clin. Pharmacol. 2015, 55, 497 – 504.
Angew. Chem. Int. Ed. 2014, 53, 12125 – 12131; Angew. Chem. 2014, [123] P. Thygesen, G. Konradsen, C. B. Schjødt, P. F. Nielsen, Endocr. Rev.
126, 12321 – 12327. 2016, SAT-034.
[101] a) R. K. Mann, P. A. Beachy, Annu. Rev. Biochem. 2004, 73, 891 – 923; [124] B. Finan, T. Ma, N. Ottaway, T. D. Muller, K. M. Habegger, K. M. Heppner,
b) D. Kadereit, J. Kuhlmann, H. Waldmann, ChemBioChem 2000, 1, H. Kirchner, J. Holland, J. Hembree, C. Raver, S. H. Lockie, D. L. Smiley,
144 – 169. V. Gelfanov, B. Yang, S. Hofmann, D. Bruemmer, D. J. Drucker, P. T.
[102] M. G. Paulick, C. R. Bertozzi, Biochemistry 2008, 47, 6991 – 7000. Pfluger, D. Perez-Tilve, J. Gidda, L. Vignati, L. Zhang, J. B. Hauptman, M.
[103] a) F. L. Zhang, P. J. Casey, Annu. Rev. Biochem. 1996, 65, 241 – 269; b) P. Lau, M. Brecheisen, S. Uhles, W. Riboulet, E. Hainaut, E. Sebokova, K.
Orlean, A. K. Menon, J. Lipid Res. 2007, 48, 993 – 1011; c) M. D. Resh, Conde-Knape, A. Konkar, R. D. DiMarchi, M. H. Tschçp, Sci. Transl. Med.
Biochim. Biophys. Acta Mol. Cell Res. 1999, 1451, 1 – 16. 2013, 5, 209ra151.
[104] a) P. Kurtzhals, S. Havelund, I. Jonassen, B. Kiehr, U. D. Larsen, U. Ribel, [125] C. Nilsson, T. K. Hansen, C. Rosenquist, B. Hartmann, J. T. Kodra, J. F.
J. Markussen, Biochem. J. 1995, 312, 725 – 731; b) S. R. Myers, F. E. Lau, T. R. Clausen, K. Raun, A. Sams, Eur. J. Pharmacol. 2016, 773, 24 –
Yakubu-Madus, W. T. Johnson, J. E. Baker, T. S. Cusick, V. K. Williams, 31.
F. C. Tinsley, A. Kriauciunas, J. Manetta, V. J. Chen, Diabetes 1997, 46, [126] A. Pocai, P. E. Carrington, J. R. Adams, M. Wright, G. Eiermann, L. Zhu,
637 – 642. X. Du, A. Petrov, M. E. Lassman, G. Jiang, F. Liu, C. Miller, L. M. Tota, G.

Reviews
Zhou, X. Zhang, M. M. Sountis, A. Santoprete, E. Capito, G. G. Chicchi, medix: albumedix.com/drug-delivery/veltis/ (accessed October 12,
N. Thornberry, E. Bianchi, A. Pessi, D. J. Marsh, R. SinhaRoy, Diabetes 2016).
2009, 58, 2258 – 2266. [144] J. M. Korth-Bradley, A. S. Rubin, R. K. Hanna, D. K. Simcoe, M. E. Leb-
[127] a) D. S. Edgerton, M. C. Moore, J. J. Winnick, M. Scott, B. Farmer, H. sack, Ann. Pharmacother. 2000, 34, 161 – 164.
Naver, C. B. Jeppesen, P. Madsen, T. B. Kjeldsen, E. Nishimura, C. L. [145] P. Barrington, J. Y. Chien, F. Tibaldi, H. D. Showalter, K. Schneck, B. Ellis,
Brand, A. D. Cherrington, Diabetes 2014, 63, 3946 – 3954; b) A. N. Diabetes Obes. Metab. 2011, 13, 434 – 438.
Zaykov, J. P. Mayer, R. D. DiMarchi, Nat. Rev. Drug Discovery 2016, 15, [146] M. A. Bush, J. E. Matthews, E. H. De Boever, R. L. Dobbins, R. J. Hodge,
425 – 439. S. E. Walker, M. C. Holland, M. Gutierrez, M. W. Stewart, Diabetes Obes.
[128] S. B. van Witteloostuijn, K. Mannerstedt, P. Wismann, E. M. Bech, M. B. Metab. 2009, 11, 498 – 505.
Thygesen, N. Vrang, J. Jelsing, K. J. Jensen, S. L. Pedersen, accepted for [147] J. Rosenstock, J. Reusch, M. Bush, F. Yang, M. Stewart, Albiglutide
publication in Mol. Pharm. Study Group, Diabetes Care 2009, 32, 1880 – 1886.
[129] A. Jonsson, J. Dogan, N. Herne, L. Abrahmsen, P. A. Nygren, Protein [148] B. Haraoui, J. P. Pelletier, J. Martel-Pelletier, Curr. Rheumatol. Rep. 2007,
Eng. Des. Sel. 2008, 21, 515 – 527. 9, 265 – 267.
[130] M. S. Dennis, M. Zhang, Y. G. Meng, M. Kadkhodayan, D. Kirchhofer, D. [149] J. M. Gao, M. Mrksich, F. A. Gomez, G. M. Whitesides, Anal. Chem. 1995,
Combs, L. A. Damico, J. Biol. Chem. 2002, 277, 35035 – 35043. 67, 3093 – 3100.
[131] a) C. Herring, O. Schon in Therapeutic Proteins: Strategies to Modulate [150] a) I. Jonassen, S. Havelund, U. Ribel, A. Plum, M. Loftager, T. Hoeg-
Their Plasma Half-Lives (Ed.: R. Kontermann), Wiley-Blackwell, Wein- Jensen, A. Volund, J. Markussen, Pharm. Res. 2006, 23, 49 – 55; b) J. L.
heim, 2012, pp. 249 – 268; b) L. J. Holt, A. Basran, K. Jones, J. Chorlton, Whittingham, I. Jonassen, S. Havelund, S. M. Roberts, E. J. Dodson, C. S.
L. S. Jespers, N. D. Brewis, I. M. Tomlinson, Protein Eng. Des. Sel. 2008, Verma, A. J. Wilkinson, G. G. Dodson, Biochemistry 2004, 43, 5987 –
21, 283 – 288. 5995.
[132] B. M. Tijink, T. Laeremans, M. Budde, M. S. V. Walsum, T. Dreier, H. J. de [151] L. Malik, J. Nygaard, R. Hoiberg-Nielsen, L. Arleth, T. Hoeg-Jensen, K. J.
Haard, C. R. Leemans, G. van Dongen, Mol. Cancer Ther. 2008, 7, 2288 – Jensen, Langmuir 2012, 28, 593 – 603.
2297. [152] a) E. Neil, G. Marsh, Chem. Biol. 2000, 7, R153 – R157; b) M. Cametti, B.
[133] J. Han, L. D. Sun, Y. Y. Chu, Z. Li, D. D. Huang, X. Y. Zhu, H. Qian, W. L. Crousse, P. Metrangolo, R. Milani, G. Resnati, Chem. Soc. Rev. 2012, 41,
Huang, J. Med. Chem. 2013, 56, 9955 – 9968. 31 – 42.
[134] D. Sleep, J. Cameron, L. R. Evans, Biochim. Biophys. Acta Gen. Subj. [153] H. K. Munch, S. T. Heide, N. J. Christensen, T. Hoeg-Jensen, P. W. Thulstr-
2013, 1830, 5526 – 5534. up, K. J. Jensen, Chem. Eur. J. 2011, 17, 7198 – 7204.
[135] J. Kim in Therapeutic Proteins: Strategies to Modulate Their Plasma Half- [154] H. K. Munch, J. Nygaard, N. J. Christensen, C. Engelbrekt, M. Oster-
Lives (Ed.: R. Kontermann), Wiley-Blackwell, Weinheim, 2012, pp. 143 – gaard, T. Porsgaard, T. Hoeg-Jensen, J. D. Zhang, L. Arleth, P. W. Thulstr-
155. up, K. J. Jensen, Angew. Chem. Int. Ed. 2016, 55, 2378 – 2381; Angew.
[136] a) F. W. Brambell, W. A. Hemmings, I. G. Morris, Nature 1964, 203, 1352; Chem. 2016, 128, 2424 – 2427.
b) F. W. Brambell, Lancet 1966, 288, 1087. [155] A. Secher, J. Jelsing, A. F. Baquero, J. Hecksher-Sorensen, M. A. Cowley,
[137] N. E. Simister, A. R. Rees, Eur. J. Immunol. 1985, 15, 733 – 738. L. S. Dalbøge, G. Hansen, K. L. Grove, C. Pyke, K. Raun, L. Schaffer, M.
[138] C. Chaudhury, S. Mehnaz, J. M. Robinson, W. L. Hayton, D. K. Pearl, Tang-Christensen, S. Verma, B. M. Witgen, N. Vrang, L. Bjerre Knudsen,
D. C. Roopenian, C. L. Anderson, J. Exp. Med. 2003, 197, 315 – 322. J. Clin. Invest. 2014, 124, 4473 – 4488.
[139] W. R. Strohl, Biodrugs 2015, 29, 215 – 239. [156] R. Bellmann, C. Feistritzer, C. J. Wiedermann, Clin. Pharmacokinet. 2012,
[140] B. Jarvis, D. Faulds, Drugs 1999, 57, 945 – 966. 51, 225 – 236.
[141] M. Sanford, Drugs 2014, 74, 2097 – 2103. [157] M. Jamnicki, T. Bombeli, B. Seifert, A. Zollinger, V. Camenzind, T. Pasch,
[142] R. M. Poole, M. L. Nowlan, Drugs 2014, 74, 929 – 938. D. R. Spahn, Anesthesiology 2000, 93, 1231 – 1237.
[143] a) J. T. Andersen, B. Dalhus, D. Viuff, B. T. Ravn, K. S. Gunnarsen, A.
Plumridge, K. Bunting, F. Antunes, R. Williamson, S. Athwal, E. Allan, L. Received: July 21, 2016
Evans, M. Bjoras, S. Kjaerulff, D. Sleep, I. Sandlie, J. Cameron, J. Biol. Revised: September 24, 2016
Chem. 2014, 289, 13492 – 13502; b) Veltis Optimized Drug Dosing, Albu- Published online on && &&, 0000

REVIEWS
An alternative direction: Despite the S. B. van Witteloostuijn, S. L. Pedersen,
commercial success of PEGylation, new K. J. Jensen*
strategies for extending the half-lives of
&& – &&
biopharmaceuticals are warranted. In
this review, we describe the available al- Half-Life Extension of
ternatives with an emphasis on conju- Biopharmaceuticals using Chemical
gation chemistry, biophysical properties, Methods: Alternatives to PEGylation
and biological compatibility.

Half-Life Extension of Biopharmaceuticals Using Chemical Methods: Alternatives To Pegylation

Uploaded by

Copyright:

Available Formats

Half-Life Extension of Biopharmaceuticals Using Chemical Methods: Alternatives To Pegylation

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Half-Life Extension of Biopharmaceuticals Using Chemical Methods: Alternatives To Pegylation

Uploaded by

Copyright:

Available Formats

DOI: 10.1002/cmdc.

Half-Life Extension of Biopharmaceuticals using Chemical

These are not the final page numbers! ÞÞ

1. Introduction this modification may lead to hypersensitivity reactions,[4] anti-

ÝÝ These are not the final page numbers!

Søren B. van Witteloostuijn obtained

These are not the final page numbers! ÞÞ

INN[b] Brand name t1/2 extension Indication Approval

3. Recombinant PEG Mimetics

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

7. Fusion Proteins Using FcRn-Mediated 8. Other Chemical Modifications—Insulin as

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

ÝÝ These are not the final page numbers!

These are not the final page numbers! ÞÞ

You might also like