Building and Characterizing Antibody-Targeted Lipidic Nanotherapeutics
Immunoliposomes provide a complementary, and in many instances advantageous, drug delivery strategy to antibody-drug conjugates. Their high carrying capacity of 20,000-150,000 drug molecules/liposome, allows for the use of a significantly broader range of moderate-to-high potency small molecule drugs when compared to the comparably few subnanomolar potency maytansinoid- and auristatin-based immunoconjugates. The multivalent display of 5-100 antibody fragments/liposome results in an avidity effect that can make use of even moderate affinity antibodies, as well as a cross-linking of cell surface receptors to induce the internalization required for intracellular drug release and subsequent activity. The underlying liposomal drug must be effectively engineered for long circulating pharmacokinetics and stable in vivo drug retention in order to allow for the drug to be efficiently delivered to the target tissue and take advantage of the site-specific bioavailability provided for by the targeting arm. In this chapter, we describe the rationale for engineering stable immunoliposome-based therapeutics, methods required for preparation of immunoliposomes, as well as for their physicochemical and in vivo characterization.
Building and Characterizing
Dmitri B. Kirpotin,* Charles O. Noble,* Mark E. Hayes,*
Zhaohua Huang,* Tad Kornaga,* Yu Zhou,
Ulrik B. Nielsen,*
James D. Marks,
and Daryl C. Drummond*
2. Preparation of Immunoliposomes
2.1. Drug-loading aid preparation
2.2. Lipid hydration and liposome sizing using extrusion
2.3. Gradient generation and drug encapsulation
2.4. Antibody selection
2.5. Antibody fragment generation and conjugation
2.6. Engineering antibody molecules having C-terminal cysteine
2.7. Thiol-reactive lipopolymer linkers
2.8. Preparing the protein for conjugation
2.9. Antibody-lipopolymer conjugation
2.10. “Insertion” of the antibody-lipopolymer conjugate into the
3. Physicochemical Characterization of Immunoliposomes
3.1. Particle size
3.2. Drug encapsulation efficiency
3.3. Drug stability
3.4. Lipid stability
3.5. Antibody association rate
4. In Vitro and In Vivo Characterization of Immunoliposomal
4.1. Optimization of antibody density on liposomal carriers
vivo pharmacokinetic and
Methods in Enzymology, Volume 502
2012 Elsevier Inc.
ISSN 0076-6879, DOI: 10.1016/B978-0-12-416039-2.00007-0 All rights reserved.
* Merrimack Pharmaceuticals, Cambridge, Massachusetts, USA
Department of Anesthesia and Perioperative Care, University of California, San Francisco, California, USA
Immunol iposomes provide a complementary, and in many instances advanta-
geous, drug delive ry strategy to antibody-drug conjugates. Their high carrying
capacity of 20,000–150,000 drug molecules/liposome, al lows for the use of a
significantly broader range of moderate-to-high potency small molecule drugs
whe n compared to the comparably few subnanomolar potency maytansinoid-
and auristatin-based immunoconjugates. The multivalent display of 5–100
antibody fragments/liposome results in an avidity effect that can make use
of even m oderate affinity antibodies, as well as a cross-linking of cell surface
recepto rs to induce the internalization required for intracellular drug release
and subsequent activity. The underlying liposomal drug must be effectively
engineered for long circulating pharmacokinet ics and stable in vivo drug
retention in order to allow for the drug to b e efficiently delivered to the target
tissue and take advantage of the site-specific bioavailabilit y provided for by
the targeting arm. In this ch apter, we describe the rationale for engineering
stable immunoliposome-based therapeutics, methods required for p repara-
tion of immunoliposomes, as well as for their physicochemical and in vivo
A liposome is a ves icle having one or more lipid bilaye rs enclosing an
aqueous interior. As used herein, “ immunoliposome” means a liposome
that has antibody molecules displayed on its outer surface. Among various
uses of immunoliposomes, the focus of this chapter is t he use of immu-
noliposomes for the targeted delivery of pharmaceuticals. Immunolipo-
somes targe ted to cance r-overexpre ssing cell s urface epitopes such as
ErbB2 (Kirpotin et al., 2006; Nielsen et al., 2002; Park et al., 2002),
EGFR (Mamot et al., 2005), GD2 (Pagn an et al., 2000; Raffaghello
et al., 2003), and CD19 (Sapra and Allen, 2002) have been shown to
improve the efficacy of encapsulated therapeutics. The improved efficacy
for vascula rly accessible ta rgets, such as in hema tologic cancers or on the
surface of a ngiogenic b lood ves sels supporting the tumors results in part
from an altered deposition and increased accumulation at t he target.
However, for solid tumors, we have shown tha t gross distribution of
immunoliposomes to the tumor is not significantl y affected by molecular
targeting (Kirpotin et al., 2006),andthusislimitedbytheextravasation
of liposomes across the vasculature and into the tumor. The improved
efficacy in solid tumors likely r esults from the differential microdistribu-
tion of immunoliposomes within th e tu mor, including internalization
into cancer cells (Kirpotin et al., 2006) where they can be processed by
intracellular enz ymes. Indeed, there have now been multiple studies
demonstrating the requirement for internalization on improved a nticancer
140 Dmitri B. Kirpotin et al.
activity (Kirpotin et al., 2006; Noble et al., 2004; Park et al., 2002; Sapra
and Allen, 2002; Sugano et al., 2000).
Antibody-targeted lipidic nanocarriers, such as immunoliposomes, offer
multiple levels of targeting and significant flexibility in the specific delivery
of entrapped therapeutic agents. This flexibility is essential in adapting the
immunoliposome for use with a specific drug, different routes of adminis-
tration, or specific application. The various components or characteristics of
the lipidic nanocarrier that can be modified are shown in Fig. 7.1. The lipid
matrix forms the membrane barrier that protects the entrapped drug from
premature interaction with the biological milieu, allowing the drug to take
on the pharmacokinetic and distribution characteristics of the carrier when
properly designed. The lipid composition can vary in membrane fluidity at
physiological temperature, surface charge density, cholesterol content, and
degree of pegylation. The relative importance and complex interdepen-
dence of these various components is the subject of multiple reviews (Allen
et al., 2006; Drummond et al., 1999, 2008) and will not be discussed further
Gradient-based drug-loading methods for weakly basic small molecule
drugs (Drummond et al., 2006, 2008; Fenske and Cullis, 2005; Haran et al.,
1993) and condensation with lipophilic amines of complex nucleic acid-
based therapeutics (Hayes et al., 2006 ; Pagnan et al., 2000)canprovidefor
highly stabilized lipidic nanocarrier formulations that resist pr emature
release of the entrapped drug whil e in the general circulation, an impor-
tant property for immunotargeted formulations. Unlike small molecule
immunoconjugates where the drug is covalently conjugated directly to the
targeting antibody, in immunoliposomes, the antibody-targeting ligand
and therapeutic are associated indirectly through the carrier (Fig. 7.1;
Noble et al., 2004). Thus, drug retention in the circulation is critical to
for encapsulating and stabilizing some more challengi ng w eakly basic small
Antibody-targeting ligands provide the specificity for increasing delivery
to target tissue either at a tissue or cellular level. Significant engineering
must be applied to ensure target specificity, lack of immunogenicity, acces-
sibility, and conjugate and particle uniformity (Noble et al., 2004; Sapra
et al., 2005). The ability to induce internalization is a key attribute of
antibody-targeting ligands used in the construction of immunoliposomes
(Kirpotin et al., 1997, 2006; Sapra and Allen, 2002). Internalization provides
access to intracellular enzymes and acidic pH that can be used to release
the entrapped therapeutic from the confines of the carrier, thus improving
bioavailability. Some of the methods for selecting internalizing antibodies,
conjugating them to lipopolymer anchors, and incorporating them into
liposomes will be described in this chapter.
Figure 7.1 (A) Suggested design of a therapeutic immunoliposome. The lipids form a
vesicle 75–120 nm in diameter, having a single bilayer membrane that encloses internal
aqueous space. The drug is stably encapsulated within the internal space in a precipi-
tated form. The membrane surfaces are grafted with the hydrophilic polymer polyeth-
ylene glycol (PEG) by inclusion of a lipopolymer PEG-DSPE. Some PEG chains are
conjugated to the antigen-binding antibody fragment (e.g., scFv) at their distal termini.
(B) Structures of the lipid components: DSPC, distearoylphosphatidylcholine; Chol,
-DSPE, poly(ethylene glycol)-derivatized phosphatidylethanol-
amine (n ¼ 45); PEG
-DSG, poly(ethylene glycol)-derivatized 1,2-distearoylgly-
cerol (n ¼ 45). Conjugation to distearoylglycerol through an ether bond results in a
neutral charge at the membrane interface, whereas conjugation through ethanolamine
results in a negatively charged phosphate at the interface.
142 Dmitri B. Kirpotin et al.
2. Preparation of Immunoliposomes
There are a series of steps required in the preparation of immunolipo-
somes. In our preferred process, the underlying highly stable liposomal drug
formulation is prepared independently from an antibody-lipopolymer con-
jugate, followed by incorporation of the conjugate into the liposomal drug
to form the final immunoliposome. The initial step in producing stable
liposomes is the preparation of the intracellular drug-loading aid to facilitate
the loading and subsequent in vivo stabilization of the drug. Liposomes are
then formed through a combination of lipid hydration, sizing, and gradient
generation by removal of the external drug-loading aid. The drug is
encapsulated at high efficiency, typically > 95%, utilizing an electrochemi-
cal gradient. Internalizing and highly selective antibodies are identified
using high throughput screening methods and subsequently engineered to
have a single C-terminal cysteine. Antibody conjugates manufactured sepa-
rately are prepared through reaction of the reactive thiol on this C-terminal
cysteine with a maleimide-terminated lipopolymer. The resulting conjugate
is finally inserted into the outer monolayer of the liposomes, converting
a previously inert liposomal drug into an active immunoliposome. The
detailed steps involved in the preparation of an exemplary liposomal drug
are given below.
2.1. Drug-loading aid preparation
A variety of salts can be used to aid remote loading of drugs into liposomes.
The method of preparation for each salt solution depends on commercial
availability. Many of the salts that are commonly used for remote-loading
such as manganese sulfate, ammonium sulfate, ammonium citrate, and citric
acid are commercially available and can be prepared by simply dissolving
in water. If the loading agent is not available with the desired counterion,
but is supplied in the acidic form, a solution can be formed and simply
titrated with the corresponding amine. Often the agent is supplied with an
alternate counterion such as Na
due to ease of
preparation or stability concerns of acids; this is common with poly- or
oligoanionic compounds. Our lab regularly uses sucrose octasulfate, which
is typically supplied as the sodium salt, as the intraliposomal trapping agent. A
weak binding di- or tri-alkylammonium salt serves as the final drug-loading
aid, and is formed through a combination of ion-exchange chromatography
(IEX) and subsequent titration with the appropriate amine (Fig. 7.2). An ion
exchange column (e.g., Dowex 50Wx8-200, Dow Chemical Co.) is used to
acidify the sucrose octasulfate. The column is first conditioned by treatment
with 2 vols. of 1 N NaOH, followed by 2 vols. of ddH
O, and finally 3 vols.
of 3 N HCl. In order to avoid excessive dilution, the sodium salt of sucrose
octasulfate is added to the column at maximum solubility, and elution is
monitored using an inline conductivity meter to allow for batching of high
concentration sucrose octasulfate fractions.
The acidic form of the agent should be quickly titrated with the appro-
priate weakly basic amine as degradation is often associated with extended
time at low pH. Upon titration, the charge of the anionic loading agent is
now countered with the positive charge of a protonated amine. The basic
amines our group prefers are di- or trialkyl-substituted amines, including
triethylamine, diethylamine, and 2-diethylaminoethanol. The resulting
solution should be characterized to ensure complete exchange of the initial
cationic species; the use of an ion-specific electrode is a relatively straight-
forward technique. Additional pharmaceutically acceptable substituted
ammonium salts are reviewed in Stahl and Wermuth (2002). Regardless
of the method used for preparing the loading agent, the pH for the solution
should be adjusted near 6.5 to minimize hydrolysis of the liposomal lipids
during the process of liposome formation. A variety of factors should
determine the concentration of the drug-loading agent. The drug-loading
capacity is generally proportional to the concentration of the drug-loading
agent. However, as the osmolality increases with concentration, care must
be taken to avoid creating an excessive osmotic gradient across the liposome
bilayer that can result in rupture of the liposome during drug loading or in
the presence of plasma (Mui et al., 1993, 1994).
2.2. Lipid hydration and liposome sizing using extrusion
The formation of liposomes can be accomplished through a range of meth-
ods. One commonality in preparing liposomes is the hydration of the lipids
in an aqueous solution. The lipids can be hydrated from a solid or from an
organic solution. Hydration of the lipids from a solid is more thorough if the
mixture is repeatedly frozen and thawed which forces water in between
the lipid membranes. If the lipids are hydrated from a solid, and more than
one lipid is used, the solid lipids can be initially dissolved in an organic
Figure 7.2 Production of diethylammonium sucrose octasulfate. Sodium sucrose octa-
sulfate is brought into free acid form by ion exchange on Dowex 50Wx8-200 resin in
hydrogen form; sucrose octasulfuric acid is neutralized with neat diethylamine to pH 6–7.
144 Dmitri B. Kirpotin et al.
solvent (e.g., chloroform/methanol, 9:1, vol:vol) to ensure proper mixing
of the different lipids, followed by solvent removal through rotary evapora-
tion and lyophilization. Ethanol injection involves hydration by mixing an
ethanol solution of lipids with the aqueous hydrating buffer. Most lipids will
require heating the ethanol to completely dissolve the lipids, in which case
the aqueous buffer should be heated to the same temperature before mixing.
Our lab prefers to use the ethanol injection method for hydrating lipids
as this method eliminates the need for toxic solvents, and results in the
formation of large (500 nm) multilamellar vesicles rather than significantly
larger multilamellar structures.
The formation of small (50–150 nm) unilamellar vesicles (liposomes)
requires additional processing of the hydrated lipid sheets or large multi-
lamellar vesicles to reach the appropriate size. Careful control over particle
size is a critical parameter for immunoliposomes, as it affects clearance rates
from the circulation, extravasation rates into solid tumors and efflux rates
from the same tumors, drug encapsulation and retention, and filterability
required for terminal sterilization. The two most common procedures for
sizing liposomes are sonication and extrusion through defined pore mem-
branes, such as polycarbonate membranes manufactured using track-etching
technology (Nuclepore(R), Whatman, USA). Unlike common depth fil-
ters, the membrane-spanning channels in polycarbonate track-etched filters
are round, smooth-walled, and of well-defined, uniform diameter
(Fig. 7.3). Generally the liposomes are formed by extrusion through poly-
carbonate membranes having a pore size of 80–100 nm with a total of 10–13
passes through the pores, with the average liposome diameter and polydis-
persity being reduced with each pass. The solution can either be passed
through single membrane multiple times, or through a stack of membranes
using a lower number of passes. Stacked membranes will typically require
more pressure to complete the extrusion process. The temperature of the
extruded solution should be maintained above the phase transition of
the particular lipid mixture comprising the liposomes, which for many
lipid compositions requires a heat-jacketed extruder at 60–70
2.3. Gradient generation and drug encapsulation
Active loading of the drug into the liposomes requires removal of the
loading aid from the liposome exterior in order to generate a gradient across
the liposome membrane (Fig. 7.4). The extraliposomal loading agent can be
removed using size-exclusion chromatography (SEC), dialysis, IEX, or a
combination of these approaches. SEC and dialysis techniques take advan-
tage of the considerable size difference between the liposomes and the
loading agents. Size discriminating separation methods are less effective
with large polymeric drug-loading agents and the utility of ion exchange
is more pronounced. Sucrose octasulfate, having eight negative charges and
a larger hydrodynamic radius than sulfate, presents some challenge for
separation by size exclusion or dialysis separation when compared to more
common salts such as ammonium sulfate or citric acid. We typically use a
Sepharose CL-2B or CL-4B column when employing SEC and a 500-kDa
molecular weight cutoff pore membrane for diafiltration. Diafiltration, partic-
ularly hollow fiber tangential flow filtration (TFF), is very practical for larger
scale separations. The solution used for replacing the loading agent should
contain both an isotonicity agent (sucrose, dextrose, saline) and an appropriate
buffer. The final osmolality (300–600 mmol/kg for most applications) of the
Figure 7.3 Forming unilamellar liposomes of the defined size by extrusion through a
defined pore (track-etched) membrane filter. (A) Schematic representation of the
extrusion process. Multilamellar liposomes (top) are fragmented into smaller, unila-
mellar vesicles (bottom) by forcing them through the pores under pressure. (B) Uni-
formity of size is reached by repeated passages of the liposomes through the membrane.
Scattering intensity-based size distribution of the liposomes obtained after the passage
number shown at the legend.
146 Dmitri B. Kirpotin et al.
extraliposomal solution should be in a range that will not cause the lipo-
somes to burst when the temperature is raised above the lipids’ phase
transition temperature during drug loading. The solubility of the drug
being loaded should be considered when choosing the isotonicity agent,
as some drugs will have low solubility in ionic salts. The pH of the liposome
solution depends upon the optimum for loading each specific drug.
Stable LAQ824 - SOS complex
Figure 7.4 Stable loading of a weakly basic substance into the liposome assisted by a
small molecule polyanion with high charge density. A weakly basic HDAC inhibitor,
LAQ824, in its neutral form crosses the liposome membrane, becomes protonated, and
tightly binds to sucrose octasulfate anion (SOS
) forming a poorly soluble salt depos-
ited as a gel or a precipitate within the liposome. Deprotonated triethylammonium
(triethylamine) is membrane-permeable and leaves the liposome maintaining the bal-
ance of charge.
Drug loading is accomplished by introducing the free unencapsulated
drug to the liposomes at temperature above the phase transition temperature
of the lipids. The drug and lipid solutions can be heated before or after
mixing depending on the particular situation. The loading time, tempera-
ture, pH, and drug to lipid ratio (drug payload) are all factors that can be
optimized in an attempt to determine the most efficient drug-loading
protocol. Our lab has determined that general conditions including a pH
between 5.0 and 7.0, time of 30 min, and temperature of 60–65
C results in
efficient loading for most of the more common liposomal cancer agents
including camptothecins, anthracyclines, vinca alkaloids, and a weakly basic
HDAC inhibitor (Drummond et al., 2005b, 2009, 2010; Mamot et al., 2005;
Noble et al., 2009). Cooling the solution will stop the loading process.
Similar to the drug-loading agent, removal of the unencapsulated drug from
the liposomes can be accomplished using SEC, dialysis, IEX, or a combina-
tion of these approaches. Highly efficient drug loading can mitigate the
requirement for removal of the unencapsulated drug.
2.4. Antibody selection
There are multiple requirements of the antibody component, including
specificity, lack of immunogenicity, binding affinity, and ability to induce
internalization (Noble et al., 2004; Sapra and Allen, 2003). We have focused
on antibody fragments selected from fully human antibody libraries to
eliminate Fc-receptor-mediated liposome clearance resulting from conju-
gation of full IgG molecules (Harding et al., 1997; Noble et al., 2004) and
the generation of an immune response that would compromise multiple
administrations if using nonhuman antibodies. Phage- or yeast-display
libraries have been used to screen for specific, tight binding, and highly
internalizing antibodies (Becerril et al., 1999; Noble et al., 2004; Poul et al.,
2000; Zhou and Marks, 2009; Zhou et al., 2010). Selection directly on cell
lines has allowed us to screen antibodies directly for their ability to induce
internalization. In this screening protocol, the library is first depleted on
control cells to deplete the library of nonspecific binders. The depleted
library is then incubated with cancer cells overexpressing the receptor of
interest, followed by washing to remove nonbinders, incubation at 37
to allow for receptor-mediated endocytosis, and stripping of cell surface-
bound phage from the surface using multiple low pH (i.e., 2.5) glycine
washes. The cells are then trypsinized and lysed using 100 mM triethylamine
(with subsequent neutralization using 0.33 M tris) to recover the internaliz-
ing phage, and finally amplified for additional rounds of selection. Typically
three rounds of selection are performed.
Antibody panels that were previously assembled based on selection for
binding can also later be screened for internalization using a CHElated
Ligand-induced Internalization Assay (CHELIA) that detects the amount
148 Dmitri B. Kirpotin et al.
of immunoliposomes on the cell surface and/or inside the cell (Fig. 7.5;
Nielsen et al., 2006). This assay has the advantage of measuring the internal-
ization potential of antibodies in the format (i.e., immunoliposomes)
that they are too eventually to be used in, as well as use with nonpurified
and low concentrations of antibodies, and avoids the necessity for preparing
stable immunoconjugates of a large panel of prospective antibodies. In
this method, hexahistidine (His
)-tagged antibodies are conjugated to fluo-
rescent liposomes through a Ni(II)-activated nitrilotriacetic (NTA) lipid
(Huang et al., 2006, 2009; Nielsen et al., 2006). These liposomes are then
I II III
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Figure 7.5 (A) Chelated liposome-antibody (CHELIA) assay for liposome-internalizing
antibody ligands. I. Fluorescently labeled Ni-NTA-immunoliposomes, liposomes bearing
antibodies attached via His-tag, are panned on the live cells and allowed to adsorb on
the antigens on the cell surface. II. The cells are incubated at 37
C to allow internaliza-
tion of the liposomes linked to internalizable antibodies: at this point, the total uptake
of the liposomes (surface-bound and internalized) can be quantified, for example,
by flow cytometry. III. The cells are treated with imidazole or EDTA to dissociate
His-tag22Ni-NTA bond of the extracellular immunoliposomes, removing them from
the cells. Now the flow cytometry assay will quantify only the cells with internalized
liposomes. (B) Using CHELIA assay to screen anti-EGFR Fabs according to their quality
as targeting ligands for EGFR-targeted immunoliposomes. Twenty-one clone of His-
tagged Fabs were produced by the display library selection for the binding to EGFR ECD
and panned over A431 cells in the presence of the fluorescently labeled PEGylated
liposomes containing Ni-NTA-PEG-DSPE (1 mol% of the phospholipid). Relative cel-
lular uptake of the liposomes across different clones was assessed by flow cytometry.
MFS—mean cell fluorescence (in arbitrary units).
added to cells and allowed to internalize for 2–4 h, after which cell-
associated fluorescence is measured either before or after disassociation
of cell surface-bound antibodies with an imidazole buffer to disrupt the
–NTA bond. The fluorescence of the pre-imidazole wash provides a
measurement of total cell-associated liposomes, and post-imidazole wash,
a measurement of internalized liposomes. Care should be taken in not
overinterpreting the results. Because the screening is typically completed
at one or two concentrations of antibodies, there is the chance that the
optimum antibody density on the liposome surface is not utilized during the
screening for many of the antibodies. Thus, this initial screen is simply used
as a quick screen to identify potential hits that can be further studied as
covalent conjugates. An example of this is shown in Fig. 7.5B where a panel
of 21 antibodies was screened using the CHELIA protocol, with four hits
being identified for further characterization. We also typically perform this
screening on a series of four cell lines overexpressing the receptor and one to
two cell lines with low levels of the receptor to help ensure that the selected
antibodies have broad applicability and specificity.
We have previously used monovalent NTA-derivatized lipids (Nielsen
et al., 2006) but have recently moved to the use of novel trivalent-NTA
lipids developed by Huang et al. (2006, 2009) due to the three orders of
magnitude higher equilibrium dissociation constant (K
) for His
mono-NTA when compared to the tris-NTA lipid (10 mM vs. 10 nM).
The lipophilic fluorescent dye DiIC18(5)-DS is incorporated into the
membrane of the liposomes at a concentration of 0.3 mol% and has excita-
tion and emission wavelengths of 650 and 670 nm, respectively. The tris-
NTA lipid is incorporated at a concentration of 0.5 mol%. A typical lipid
composition for the CHELIA-detectable liposomes would also include
hydrogenated soy phosphatidylcholine (HSPC), cholesterol, and PEG-
DSPE in a 3:2:0.015, mol:mol:mol ratio.
For the screening protocol, a solution containing the NTA liposomes
(50 mM PL), 10–50 mg Fab
/ml, and 50 mM NiSO
is freshly prepared. Four
cell lines known to overexpress the receptor of interest and two control cell
lines are trypsinized, counted, and aliquoted at a concentration of 1 10
cells/well and 100 ml in polypropylene V-bottom-shaped 96-well plates.
The cells are pelleted by centrifugation at 1200 rpm for 5 min and at 4
and the media subsequently removed by aspiration and replaced with 100 ml
of the above immunoliposomes. Control Ni
-NTA liposomes without
antibody are also added as a control. The plates are then incubated on a
shaker at 37
C and 5% CO
for 2–4 h. For one set of cells, the unassociated
material is removed by washing with nondisassociating phosphate-buffered
saline and the cell-associated fluorescence determined by flow cytometry
using the FL4 channel (FACSCalibur, Becton Dickinson, USA). For a
second set of cells, the cells are washed with 250 mM imidazole buffer
post liposome incubation and the cell-associated fluorescence determined
150 Dmitri B. Kirpotin et al.
by flow cytometry as a measure of the total cell-internalized immunolipo-
somes. For screens where material is limited, only internalized samples can
be measured (Fig. 7.5B). However, measuring both total and internalized
provides a more detailed understanding of both the extent and efficiency of
2.5. Antibody fragment generation and conjugation
The bond between the liposome and the antibody molecule is the crucial
element of the immunoliposome design. In order to be effective as a vehicle
for antibody-targeted delivery of its payload, the liposome-antibody bond
must fulfill a number of requirements: (1) Stability over the period of time
from making the immunoliposome and its interaction with the target cell
or tissue, including stability during storage in the vial and while in vivo in
the circulation, (2) compatibility with the liposome payload, (3) minimal
effects on the antigen-binding properties of the antibody on the liposome,
including the effects from any polymer coating, (4) minimal effects on the
pharmacokinetics and non-target-related pharmacodynamics, (5) Ease, con-
trol, and reproducibility of the antibody-liposome conjugation process,
including lot-to-lot consistency, and industrial scalability, and (6) the poten-
tial for multiple antibodies/specificities on the same particle.
One particular strategy seems to have met the above criteria well and was
adopted by the authors. This strategy, outlined in Fig. 7.1A, also known
as “micellar insertion,” involves the following steps: First, producing the
antibody molecule that has a unique conjugation site for attachment of the
linker group that will link the antibody to the surface of the liposome. For
example, one very convenient conjugation site is created by a thiol group of
cysteine residue exposed at the C-terminus of the antibody polypeptide
chain. It may be engineered into the protein sequence by recombinant
methods, or proteolytically generated. The antibody in this context is any
polypeptide carrying antibody-binding domain(s) of an immunoglobulin
molecule, such as Fab
, Fv, or scFv. While we focus on the antibody ligands,
it is clear that any protein ligand can be used. The second step is producing
an amphiphilic linker molecule that includes a hydrophobic domain, such as
a lipid moiety that includes one or more hydrocarbon chains or a sterol, and
a hydrophilic spacer arm attached to the hydrophobic domain, typically a
flexible, hydrated polymer, such as poly (ethylene glycol) (PEG). On the
terminus, contralateral to the hydrophobic domain, the spacer arm bears a
functional group reactive to the conjugation site on the protein molecule.
A third step includes forming a conjugate between the antibody protein
and the linker molecule in an aqueous solution. Due to self-association of the
hydrophobic domains and relatively large size of the hydrophilic portions of
its molecule, the linker–antibody conjugate remains in solution in the form of
micelles and can be purified from the unconjugated protein at this time.
Finally, incubating the liposome with the micellar antibody–linker conjugate,
during which step the conjugate molecules leave the micelles and their
hydrophobic domains merge into the outer leaflet of the liposome bilayer,
forming a link between the liposome and the antibody. Optionally, the
residual micellar material can then be separated from the antibody-conjugated
liposomes. Each of these steps is considered in more detail below.
2.6. Engineering antibody molecules having C-terminal
Nonimmunogenic short flexible C-terminal conjugation sequences can be
engineered into the antibody fragments. The most convenient way of intro-
ducing this feature is expression of a recombinant protein with a single
C-terminal cysteine. While pepsin digestion of the whole IgG molecule
is an established method for generation of Fab
fragments having free thiols
at the hinge region, we have found that the presence of a CPPC hinge
sequence and/or extra free Cys (e.g., in the case of human IgG1) makes
proteolytically generated Fab
prone to forming stable intramolecular disulfide
bridges at the hinge and otherwise to attachment of more than one linker
chain, both leading to low yields of the liposome-linked antibody. In addition,
we have removed Myc-His
sequences typically found in the C-terminus of
many antibody libraries in order to reduce nonspecific immunogenicity and an
observed rapid clearance observed in immunoliposome constructs prepared
from them. For scFv antibody fragments, we have used a preferred sequence
of Gly4Cys (Nellis et al., 2005a), while for Fab’s we have typically employed
the sequence THTCAA in the C-terminus (Carter et al., 1992).
2.7. Thiol-reactive lipopolymer linkers
The linker molecule generally contains a hydrophilic polymer chain, such as
PEG, terminally linked to a lipid domain (phospholipid or sterol), and bears
a thiol-reactive functional group at the terminus contralateral to the lipid
domain (Fig. 7.6A). The linkers comprising PEG spacers of various size,
phosphatidylethanolamine (PE) lipid anchors of various hydrocarbon chain
length, and terminal maleimide or iodoacetate groups are currently com-
mercially available from Avanti Polar Lipids (Alabama, USA) and NOF
Corporation (Japan). The PE-based lipopolymer linkers introduce an addi-
tional negative charge due to the phosphate residue in the lipid head group,
and contain fatty acid ester bonds are modestly susceptible to hydrolysis
leading to single-chain lipid (lysolipid) anchors that make the conjugates
more prone to leaving the liposome bilayer. To avoid these disadvantages,
thiol-reactive lipopolymers with nonionizable and non-ester lipid moieties,
such as distearylamine, have been reported (Drummond et al., 2005a).
DSPE-PEG-based linkers have advantage due to the availability and
152 Dmitri B. Kirpotin et al.
Maleimide PEG 2000
N Gly Gly Gly Gly Gly (ScFv)
0 2.5 5 7.5
10 12.5 15 17.5
0 2.5 5 7.5 10 12.5
0 2.5 5 7.5 10 12.5 15 17.5
% eluted per fraction
Elution volume (mL)
Figure 7.6 (A) Chemical structure of the antibody-lipopolymer conjugate. Anti-
HER2 single-chain Fv antibody fragment (F5) with appended C-terminal GGGC
sequence (F5-SH) is conjugated through the thiol group to maleimide-terminated
PEG-DSPE linker molecule (Nellis et al., 2005a). (B) Schematic representation of the
“micellar insertion” strategy for the antibody-liposome conjugation. A drug-loaded
liposome (I) (in this case, PEGylated unilamellar vesicle) is incubated with the micelles
of the antibody-lipopolymer linker conjugate (II) whereby the hydrophobic domains of
the conjugate merge into the lipid bilayer, producing an immunoliposome (III). (C)
Monitoring of the antibody-lipopolymer conjugation reaction by SEC. A recombinant
human IgG1 Fab fragment with the C-terminal THTCAA sequence (Fab-SH) was
expressed in Escherichia coli, purified from the periplasmic extract by protein A resin
capture, reduced with 15 mM cysteine and incubated with the micellar solution of
positive clinical experience with PEG-DSPE conjugates as liposome con-
stituents. The antibody-lipopolymer conjugate is retained in the liposome
bilayer as a result of the hydrophobic interactions between the lipid domain
and the bilayer membrane. PEG-PE molecules with shorter acyl chain
lengths than stearic acid have been shown to more rapidly disassociate
from the liposomes in vivo and thus should be avoided (Parr et al., 1994).
2.8. Preparing the protein for conjugation
The next step is to prepare the scFv-Cys or Fab-Cys antibody fragment for
conjugation with the lipopolymer linker. The conjugation requires the pres-
ence of free reactive thiol group at the C-terminal Cys-containing sequence.
Using bacterial expression system, the secretion of the terminally cysteinated
antibody fragment into the culture medium frequently lead to the irreversible
blockade of the terminal Cys by an unknown component. Periplasmic expres-
sion is reported to preserve the reactivity of C-terminal Cys more effectively
(Carter et al., 1992). In our experience, scFv-Cys and Fab-Cys isolated from
the bacterial periplasm, as well as scFv-Cys produced in the mammalian system
(CHO cells), after affinity isolation step (e.g., protein A resin) come out as a
mixture of disulfide-linked dimers and monomers with the terminal Cys
partially blocked (Fig. 7.6D, lane 2). Reduction of the protein with a thiol
compound results in the dissociation of the dimer and partial removal of the
-DSPE linker (protein/linker molar ratio 1:4, mol:mol) at pH 6.0 and
ambient temperature. At indicated times, 0.05-ml aliquots of the reaction mixture were
quenched with cysteine (0.5 mM final) and chromatographed on the TSK Gel Super
SW3000 4.6 30 column in 0.45 M NaCl, 0.01M phosphate buffer pH 7.1 at 0.35 ml/
min with UV detection at 280 nm. Incubation times: bottom panel, zero time; middle
panel, 1 h; top panel, 2 h. Notice the appearance of a micellar protein eluted at 6.5 min
and the gradual decrease of the free Fab-SH eluted at 10.7–11.1 min. (D) SDS-PAGE
analysis of the antibody-lipopolymer conjugate preparation. F5-SH scFv was isolated
from the E. coli periplasmic extract using protein A chromatography (lane 1), reduced
by passage through a Reduce-Imm column (Pierce) at pH 6.0 (lane 2), incubated with
maleimido-PEG-DSPE micelles in an aqueous buffer (lane 3), the resulting conjugate
was incubated with commercial PEGylated doxorubicin liposomes (Doxil, Alza Cor-
poration) at 30–40 scFv/liposome, the liposomes were separated from the unconju-
gated proteins and micelles by Sepharose CL-4B chromatography (see E), solubilized in
the SDS sample buffer, and applied on the gel (lane 4). The gel is stained with
Coomassie Blue R. Notice the disappearance of the F5 dimer band after the reduction
step, and the shift of 3 kD between the unconjugated protein (26 kDa, lanes 1,2) and
the F5-PEG-DSPE conjugate (lanes 3,4; Nielsen et al., 2002). (E) Separation of immu-
noliposomes from unbound micellar antibody-lipopolymer conjugate using Sepharose
CL-4B chromatography. Doxil liposomes incubated with F5-PEG-DSPE conjugate
(30 min, 60
C) were applied on a Sepharose CL-4B column (bed volume 15 ml) and
eluted with PBS. The liposomes were quantified by doxorubicin absorbance at 485 nm,
the micellar and/or free protein was quantified by absorbance at 280 nm. The presence
of liposome-associated scFv was confirmed by SDS-PAGE (D; Nielsen et al., 2002).
154 Dmitri B. Kirpotin et al.
blocking group. The choices of the reducing compound and the reduction
conditions are governed by the specificity toward the terminal cysteine while
keeping intramolecular disulfide bridges intact. Hydrophilic thiols, such as
2-mercaptoethylamine, cysteine, or dithioerythritol, are best suited for the task
due to their tendency to avoid hydrophobic pockets that harbor intramolecu-
lar disulfides. Strong reducing agents such as dithiothreitol (DTT) or tris-
carboxymethylphosphine tend to overreduce the protein. The protein is
incubated with 10–20 mM reducing agent at pH < 7, and the excess of the
reducing agent is thoroughly removed, for example, by SEC on Sephadex
G-25 or similar carrier. Alternatively, the reducing agent removal step can be
combined with the protein purification step, for example, by IEX. Fabs and
scFvs often have high isoelectric points (pI 8.0–9.0) and therefore can be
conveniently purified on a cation exchanger, such as SP Sepharose (GE
Healthcare). The reducing agent, especially of the net zero ionic charge
(DTE, cysteine), usually appears in the eluate at lower ionic strength than
does the protein, making is unnecessary to remove it in advance of the IEX
(Nellis et al., 2005a). During the IEX step, it is important to maintain the pH of
elution buffers sufficiently low (pH 5.5–6.5), to remove the dissolved oxygen
(e.g., by bubbling of argon or nitrogen), and to include 1–5 mM chelator
(EDTA) to prevent heavy metal-catalyzed thiol oxidation. The reactive thiol
group content of the reduced protein is usually about 0.7–0.8/ molecule. The
higher thiol content is often associated with partial reduction of the intramo-
lecular disulfides and should be avoided.
2.9. Antibody-lipopolymer conjuga tion
The purified and reduced antibody fragment can be directly conjugated
to the liposomes bearing thiol-reactive groups. In our preferred method,
the antibody fragment is first conjugated to a thiol-reactive, micelle-forming
lipopolymer and then incubated with the liposomes to effect insertion of
the antibody-lipopolymer into the outer leaflet of the liposome bilayer. The
reaction between maleimide-terminated PEG-DSPE linker (mal-PEG-
DSPE; Fig. 7.6A) and scFv-Cys or Fab-Cys is carried out in aqueous solution
at pH about 6.0; it takes only a few hours and proceeds to completion. The
reaction proceeds to completion even at a stoichiometric ratio between
the available protein thiols and linker maleimides. However, we usually add
an excess of mal-PEG-DSPE (protein/linker molar ratio of 1:4) to act as a
“spacer” that prevents instability of the micelle due to the excluded volume of
the conjugated protein chains. At the end of the reaction, the excess mal-
eimide groups are quenched by a small amount of cysteine or mercaptoetha-
nol. Quenching is especially important if the conjugate is later transferred into
amediumatpH> 6, since the excess maleimide groups have the tendency
to react, albeit slowly, with the side chains of lysine, leading to gradual
accumulation of polyconjugated species that have reduced ability to insert
into the liposome bilayer. If excess mal-PEG-DSPE or the quenching agent
have negative effects on the protein or the liposome, PEG-DSPE molecule
without the terminal maleimide can be substituted for excess mal-PEG-
DSPE (i.e., 1 mol part of mal-PEG-DSPE and 3 mol parts of PEG-DSPE
for 1 mol part of the protein) at the expense of slightly lower yield and longer
incubation time. Although it is not necessary to remove unconjugated protein
prior to the subsequent membrane insertion step (Nielsen et al., 2002), it is
useful to purify the conjugate for the purposes of storage and better characteri-
zation. Due to the large size of the lipopolymer micelles (equivalent molecular
weight 850 kDa; Nellis et al., 2005a), SEC is a convenient way to do so
(Fig. 7.6C). SEC is also convenient for monitoring the reaction by the amount
of protein (OD 280) appearing in the first-eluted, micellar fraction (Fig. 7.6E).
The antibody-lipopolymer conjugate can be concentrated and/or purified
using ultrafiltration on the membrane with 300-kDa molecular weight cutoff,
in a stirred cell or, at a larger scale, using a hollow fiber cartridge. In order to
maintain stability of the conjugate against aggregation in solution, it is useful
to maintain at least 1 mM of a polycarboxylic acid buffer, such as citrate. The
purified conjugate can be stored frozen at 70
C for many months without
deterioration. Alternatively, the purified conjugate can be lyophilized from
the solution containing citrate buffer and a stabilizer, such as 10% sucrose,
stored, refrigerated, and reconstituted in distilled water.
The course of protein reduction, conjugation, and purification of the
conjugate can be monitored using nonreducing SDS-PAGE with the regu-
lar protein Coomassie or Sypro Ruby stain (Fig. 7.6D). Compared to the
unconjugated protein, the conjugate has a small but distinct shift due to the
addition of a 2.9 kDa linker. Protein species having more than one conju-
gated species appear as a ladder of yet slower moving bands and usually are
the sign of the low-quality preparation. Excess lipopolymer appears on the
gel as a faint band near the tracking dye front. Despite the micellar nature of
the conjugate in solution, as long as the protein concentration is low
(<1 mg/ml), it is possible to quantify the protein using UV spectropho-
tometry at 280 nm and the same extinction value as for the free protein. The
effect of light scattering by micelles on the UV spectrum manifests as the
reduced OD280/OD260 ratio (1.60–1.85 vs. 1.90–2.00 for the free pro-
tein). As mentioned above, it is important to include sufficient amount of
citrate (1–10 mM) in all dilution buffers to prevent aggregation.
2.10. “Insertion” of the antibody-lipopolymer conjugate into
Simple coincubation of the preformed liposome and the antibody-
lipopolymer conjugate in an aqueous medium at elevated temperatures
results in the transfer of the conjugate molecules into the lipid bilayer via
anchoring of the lipopolymer hydrophobic domain in the hydrophobic
156 Dmitri B. Kirpotin et al.
inner portion of the bilayer. The process usually is remarkably efficient,
resulting in the capture of at least 70% of the conjugate onto the liposome
membrane, and typically over 80%, with little dependence on the antibody/
lipid on the range of 5–100 proteins/liposome. If the liposome size is close to
100 nm (which corresponds to the average of 80,000 phospholipid mole-
cules/liposome), the density of liposome-conjugated antibody can be quite
accurately defined by the antibody/phospholipid molar ratio in the coincu-
bation mixture. If the inner space of the liposome is loaded with the drug,
such as doxorubicin, the drug should be retained during the insertion process.
Addition of a mixture of several antibody-lipopolymer conjugates having the
same linker molecules results in the liposome bearing the combination of
antibodies essentially at the same ratio as in the conjugate mixture. The rate
of antibody-lipopolymer transfer from the micelle onto the bilayer depends
on the aggregate state of the bilayer lipids. Liquid-crystalline bilayers (above
of the liposome lipid) incorporate the conjugate faster than gel-state
bilayers (below T
); however, even in the gel state, the transfer of the
antibody-lipopolymer onto the liposome is efficient and not overly slow.
To initiate conjugate transfer, the liposomes and the conjugate are mixed in
any liposome- and antibody-compatible aqueous buffer containing 1mM
polycarboxylate (citrate) to maintain stability of the conjugate micelles. If the
denaturation temperature of the antibody so permits, the conjugate–liposome
mixture is quickly heated to the temperature above the lipid T
, and the
incubation continues for 20–30 min with slow agitation. Then the mixture is
quickly chilled down to ambient temperature or below. For less thermostable
antibodies, overnight incubation at 37
C is effective.
The amount of liposome-linked protein can be quantified using SDS-
PAGE essentially the same way as for the conjugate itself (Fig.7.6D). Premade
gradient gels (4–15% for Fab, 10–20% for scFv, e.g., Ready Gel Tris–HCl
from Bio-Rad) are suitable for this purpose. Liposome lipids, complexed with
SDS, move close to the tracking dye front. The antibody conjugate band
is quantified by densitometry of the gel stained with colloidal Coomassie
(Bio-Safe(R) Coomassie G-250 protein stain, Bio-Rad) according to the
manufacturer. The protein standards (made from the conjugate solution of
the known protein concentration) are run concurrently. Typically, the band
staining is linear to the amount of protein in the range of 50–500 ng/lane.
3. Physicochemical Characterization of
Thereareanumberofphysicochemical parameters which play a
critical role in determini ng the effectiveness of an immunoliposome.
Theseincludeparticlesize,drugencapsulation rate, lipid stabil ity, drug
degradation, antibody association rate, and immunoreactivity. As
described above, particle size af fects the rate of clearance from the circula-
tion, stability of encapsulation, and rate of tumor uptake of immun olipo-
3.1. Particle size
The particle size for immunoliposomes is measured ( 1) as an in-process
control to ensur e the completion o f the extrusion step, (2) following drug
loading and antibody co nj ugatio n to ensure th at antibody c onjug atio n did
not result in undesirable interliposomal cross-linking and thus th e forma-
tion of aggregates, and (3) during storage. Dynamic light scattering, also
sometimes referred to as photon correlation spectroscopy, is typically used
to monitor the particle size of these submicron immunoliposomes
(Ostrowsky, 1993). The origi nal Coulter or Nicomp instruments primar-
ily measured light scattering at a 90
angle. However, modern instruments
such as the Malvern Zetasizer can measure light scattering at higher angles
), t hus allowing measurement of liposomes in undilu ted samples
and over a greater dynamic range. The z-average particle diameter and
polydispersity index (PI) are calculated using the cumulants method by the
instrument’s built-in software. The desired average diameter can vary
depending on the formulatio n and a pplication, but is generally in the
range of 80–1 20 nm. For most liposoma l drugs a PI of 0 .2 or less
3.2. Drug encapsulation efficiency
thepercentageofencapsulateddrugand the ratio of drug-to-lipid matrix.
We typically calculate both parameters following measurement of a pre-
and post-column-purified l iposome sample for drug and phospholipid
content. The specific chromatograph employed for se parati ng the free
andencapsulateddrugsmayvarydepending on the active drug entrapped.
However, for many small molecule drugs, we utilize 10 ml PD-10 col-
umns packed with Sephadex G- 75 resin and eluted wi th buffered saline.
When phosphorous buffers and trapping solutions are omitted during
liposome prepa rat ion, then liposoma l p hospho lipi d can be determin ed
indirectly by measuring total inorganic phosphorus following acid diges-
tion with sulfuri c acid (Bartl ett, 1959). The drug is generally ext racted
with an acidic methanol solution, and subsequently quantified by HPLC
with UV or fluorescence detection. The entrapment efficiency i s calcu-
lated as follows: e ntrapment efficiency (%) ¼ 100 [drug/lipid]
, where pre and post refer t o t he respective ratios pre and
post purification on the Sephadex G-75 column.
158 Dmitri B. Kirpotin et al.
3.3. Drug stability
Drug degradation products can be observed in liposome formulations due in
part to the fact that the electrochemical gradient employed to load the drug
can result in extremes of pH, and the fact that liposomes concentrate drugs
at exceedingly high concentrations putting them in close proximity to
another and increasing the propensity for dimerization reactions. It is thus,
helpful to do early stress testing of the free drugs at extremes of pH as well as
accelerated stability studies of the actual liposomal formulations to help
identify potential degradation products early in the formulation optimiza-
tion process. Drug impurities and degradation products are typically
detected using HPLC through two runs at an order of magnitude difference
in concentration to accurately detect both the parent drug, as well as most
minor metabolites. The specifics of the HPLC method vary depending on
the specific drug formulated. However, often solvents optimized for disso-
lution of the corresponding free drugs require modification in order to
solubilize the high lipid concentrations typically present in immunolipo-
3.4. Lipid stability
Analysis of individual lipids will reveal degradation products generated
during the liposome formulation process or subsequent storage. In aqueous
environment, double-chain phospholipids are known to degrade slowly.
The major degradation product in phosphatidylcholine is lysophosphatidyl-
choline, resulting from simple hydrolysis at the sn-2 ester position A robust
method involves the use of a normal phase HPLC method to separate
diacylphosphatidylcholine and lysophosphatidylcholine with either evapo-
rative light scattering detection (ELSD) or charged aerosol detection
(CAD). We use a YMC PVA-Sil column with 5 mm particle size, 120 A
pore size, 2.0 250 mm column size (Waters Corporation) and the sepa-
ration is performed in a gradient fashion using hexanes, isopropyl alcohol
(IPA), and water. The portion of each starts as 57% hexanes: 40% IPA: 3%
water and increases C to 52% hexanes: 40% IPA: 8% water over 20 min,
followed by a 20 min equilibration back to the starting mobile phase ratio.
We prefer a CAD detector due to the increased sensitivity and dynamic
range. For the ESA Corona Plus CAD, the detection range is set to 200 pA
and the nebulizer temperature to 26
C. The input gas (N
) pressure is set at
35 psi and the gas flow rate regulated automatically. Typical retention times
are 18.0 min for HSPC and 25.4 min for lysoPC. The 50 ml sample in
hexanes:IPA (3:2, vol:vol) is injected on at both 500 and 5 mM total
phospholipid to capture the PC and lysoPC at concentrations well within
the linear range of the detection method. Typically, lysoPC is seen in the
range of 0.1–5 mol% in well-controlled liposomes.
3.5. Antibody association rate
An accurate determination of the amount of targeting antibodies or frag-
ments attached to each liposome is required to evaluate the formulation
process and assess influence of antibody density on cellular binding and
internalization efficiency. Quantitation of the antibody fragment is compli-
cated by the complex nature of the immunoliposome, which contains,
drug, loading agent, a variety of lipids, and antibody fragments which are
conjugated to lipids. SDS-PAGE has proven to be valuable, as the technique
isolates each analyte by molecular weight. The antibody-lipid conjugate has
the highest molecular weight of all the individual components which make
up immunoliposomes which simplifies resolution from the mixture. Solu-
bilization of the immunoliposome components by heating in detergent
enables proper separation. Treatment of the liposomes with commonly
used SDS-PAGE sample buffers produces good results; however, for immu-
noliposomes with low protein content (that require larger lipid burden/
sample), it is useful to double the amount of the sample buffer. Premanu-
factured gels (10–20% polyacryamide, e.g., Ready Gel, Bio-Rad) are suit-
able for both scFv and Fab quantification. The lipids in the form of SDS
complexes migrate close to the gel front. The conjugate can be quantified
after regular Coomassie staining and densitometry of the protein band by
comparison of the concurrently run antibody standards in the range of
100–500 ng/lane (for 1 mm gels). Fluorescent stains commonly used for
protein detection on gels allows for detection in the 10 ng range (Nellis
et al., 2005b ). If the antibody is added to the liposomes as a micellar lipid
conjugate, care must be taken to ensure complete removal of the uninserted
antibody-lipid conjugate because there is no way to distinguish it from that
which is associated with the liposome.
4. In Vitro and In Vivo Characterization of
4.1. Optimization of antibody density on liposomal carriers
The antibody density can be varied on the liposome surface by simply
mixing liposomes with varying concentrations of antibody-drug conjugates.
We typically vary the density between 5 and 100 copies/liposome using the
micellar insertion method outlined above. The liposomes are then purified
using Sepharose CL-4B gel chromatography followed by measuring the
phospholipid content by standard phosphorous assay (Bartlett, 1959)and
the antibody by densitometric analysis of fluorescently stained proteins in an
SDS gel (see above). This quality control of the immunoliposome formula-
tions is important to correct for differences in conjugate insertion efficiencies
160 Dmitri B. Kirpotin et al.
between preparations. The liposomes used contain 0.3 mol% of DiIC18(5)-
DS and can thus be used to monitor cell uptake using flow cytometry.
Between 10 and 12 cell lines containing various receptor densities (typically
over two orders of magnitude) are incubated with either a nontargeted
liposome control or an immunoliposome containing increasing concentra-
tions of conjugate (5, 10, 15, 30, 60, and 90 Fab
C for 4 h. Cells are then washed using PBS and analyzed by flow
cytometry using the Cy5 channel. The amount of total cell-associated
fluorescence is plotted as a function of antibody density and evaluated
over the entire panel of cell lines to determine the optimum density with
respect to both specificity and overall cell uptake (Fig. 7.7).
4.2. In vivo pharmacokinetic and biodistribution studies
After selection of the appropriate liposome, drug and targeting strategy
in vitro, (discussed earlier) pharmacokinetic and disposition studies in mice
or rats are a useful predictive tool to see how the liposome formulation would
perform in an in vivo setting. In order to benefit from the EPR effect, the
liposome formulation must be sufficiently stable in circulation to be able to
12 25 37 74 148 12 25 37 74 148 12 25 37 74 148
Figure 7.7 Effect of the grafted antibody density in the uptake of immunoliposomes by
the cells with the high or medium surface abundance of the antigen. Fluorescently
labeled anti-EGFR immunoliposomes of the identical size and various numbers of scFv
copies per liposome were prepared from anti-EGFR scFv with low (C10,
¼ 264 nM), medium (P2/4, K
¼ 15 nM), and high (P2224, K
¼ 0.9 nM) affinity.
The immunoliposomes were incubated at identical concentrations with the high EGFR
(A431) or medium EGFR (MDA-MB-231) cells, and the amount of immunoliposome
uptake by the cells was assessed by flow cytometry. MFI—mean cell fluorescence
intensity (in arbitrary units).
accumulate in the tumor. Many factors contribute to the pharmacokinetic
and biodistribution behavior of liposomal formulations. Lipid composition,
surface charge, size, degree of pegylation, dose, dosing schedule, route of
administration, encapsulated drug, and targeting ligand can all affect the
clearance rates of the carrier and ultimately affect the ability of the immuno-
liposome to deliver drug in a site-specific manner (Drummond et al., 2008).
To maximize the information garnered from such studies, it is important
to measure drug and liposome-associated lipid in the collected plasma and
tissue samples, in order to follow the clearance and distribution of both the
liposomal carrier as well as the encapsulated drug. A nonexchangeable lipid
H]cholesteryl hexadecyl ether (Fig. 7.8A) is used to follow the
liposomal carrier. In this way, individual PK parameters of the drug and
carrier can be measured, but also their relative concentrations can give
information pertaining to the leakage of the drug from the carrier. This is
important, for example, in order to discriminate between a fast leaking
formulation and a fast clearing formulation, that is, changes in the initial
[drug]/[lipid] levels indicate drug leakage. Drug release rates from liposomes
can be characterized by their half-life of release times (T
), and are calcu-
lated using the exponential constant (l), from a single exponential Wt to the
plot of drug/phospholipid ratio versus post injection time [N(t) ¼ N
N(t)isthedrug-to-PLratioattimet and N(0) is the same ratio at time
0(Drummond et al., 2009; Noble et al., 2009). We typically screen multiple
formulations in vivo in PK studies to aid in the engineering of a liposomal drug
with optimum drug release and clearance rates. For simplicity, early studies
focused on formulation stability are completed in the absence of targeting
ligand. However, it is also critical once a stable formulation is identified to
repeat the measurement of clearance and in vivo drug release using the
targeted formulation to ensure that the addition of the targeting ligand has
not disrupted the membrane and increased drug leak rates or increased
clearance via either specific or nonspecific mechanisms. Finally, we compared
optimized targeted and nontargeted liposomes in full biodistribution studies,
looking at the effect of targeting on drug deposition in tumor, organs of the
mononuclear phagocyte system (MPS; i.e., liver and spleen), organs of
potential sites of toxicity, or organs where distribution may occur based on
target receptor expression levels. Again, measuring changes in drug-to-lipid
ratios provides an approximation of drug release rates in various tissues.
In the case of hydrophilic drugs, one can indirectly measure the drug
concentration by assuming that the clearance rate of the unencapsulated drug
is sufficiently faster than the clearance rate of the liposomal drug (Drummond
et al., 2006). Direct measurement of the in vivo encapsulated drug conc entration
can also be obt ained by purifying the liposomes from plasma components using
gel chromatography or solid phase extraction methods (Gabizon et al., 1994;
Zamboni et al., 2007). Typically, HPLC (Noble et al., 2009; Ta ggar et al., 2006)
or photometric (Charrois and Allen, 2004) analysis of plasma or tissue extracts
162 Dmitri B. Kirpotin et al.
0 5 10 15 20 25 30 35 40 45 50
Tim e (h)
% of original drug/lipid
Cholesteryl hexadecyl ether
0 5 10 15 20 25 30 35 40 45 50
Tim e (h)
% ID in blood
Figure 7.8 Determinations of pharmacokinetics and in vivo drug releas e of a
liposomal drug. (A) Chemical structure of the nonexchangeable lipid marker,
H]cholesteryl hexadecyl ether. (B) Mon itoring of topotecan (triangles) and lipid
marker (circles) in the blood of an animal following the administration of a drug-
loaded immunol iposome. (C) Kinetic curve of the in vivo drug release derived from
the data of Panel B.
can be used to measure the drug concentrations. For measuring liposome
concentrations in plasma or tissue extracts, the authors typically use a non-
exchangeable radiolabeled lipid marker ([
H]-CHE) followed by scintillation
counting. Techniques such as ELISA (Park et al., 2002) or gamma counting of
I-labeled antibody may be used to determine the biodistribution of the
targeting ligand in the plasma or tissues. Premature dissociation of the ligand
from the liposome surface will lessen the ability of the carrier to target cells.
Allen, T. M., Cheng, W. W., Hare, J. I., and Laginha, K. M. (2006). Pharmacokinetics and
pharmacodynamics of lipidic nano-particles in cancer. Anticancer Agents Med. Chem. 6,
Bartlett, G. R. (1959). Phosphorous assay in column chromatography. J. Biol. Chem. 234,
Becerril, B., Poul, M. A., and Marks, J. D. (1999). Toward selection of internalizing
antibodies from phage libraries. Biochem. Biophys. Res. Commun. 255, 386–393.
Carter, P., Kelley, R. F., Rodrigues, M. L., Snedecor, B., Covarrubias, M., Velligan, M. D.,
Wong, W. L., Rowland, A. M., Kotts, C. E., Carver, M. E., et al. (1992). High level
Escherichia coli expression and production of a bivalent humanized antibody fragment.
Biotechnology (N. Y.) 10, 163–167.
Charrois, G. J., and Allen, T. M. (2004). Drug release rate influences the pharmacokinetics,
biodistribution, therapeutic activity, and toxicity of pegylated liposomal doxorubicin
formulations in murine breast cancer. Biochim. Biophys. Acta. 1663, 167–177.
Drummond, D. C., Meyer, O., Hong, K., Kirpotin, D. B., and Papahadjopoulos, D. (1999).
Optimizing liposomes for delivery of chemotherapeutic agents to solid tumors. Pharma-
col. Rev. 51, 691–743.
Drummond, D. C., Kirpotin, D. B., and Hong, K. (2005a). Liposomes Useful for Drug
Delivery. (United States).
Drummond, D. C., Marx, C., Guo, Z., Scott, G., Noble, C., Wang, D., Pallavicini, M.,
Kirpotin, D. B., and Benz, C. C. (2005b). Enhanced pharmacodynamic and antitumor
properties of a histone deacetylase inhibitor encapsulated in liposomes or ErbB2-targeted
immunoliposomes. Clin. Cancer Res. 11, 3392–3401.
Drummond, D. C., Noble, C. O., Guo, Z., Hong, K., Park, J. W., and Kirpotin, D. B.
(2006). Development of a highly active nanoliposomal irinotecan using a novel intrali-
posomal stabilization strategy. Cancer Res. 66, 3271–3277.
Drummond, D. C., Noble, C. O., Hayes, M. E., Park, J. W., and Kirpotin, D. B. (2008).
Pharmacokinetics and in vivo drug release rates in liposomal nanocarrier development.
J. Pharm. Sci. 97, 4696–4740.
Drummond, D. C., Noble, C. O., Guo, Z., Hayes, M. E., Park, J. W., Ou, C. J.,
Tseng, Y. L., Hong, K., and Kirpotin, D. B. (2009). Improved pharmacokinetics and
efficacy of a highly stable nanoliposomal vinorelbine. J. Pharmacol. Exp. Ther. 328,
Drummond, D. C., Noble, C. O., Guo, Z., Hayes, M. E., Connolly-Ingram, C.,
Gabriel, B. S., Hann, B., Liu, B., Park, J. W., Hong, K., Benz, C. C., Marks, J. D.,
et al. (2010). Development of a highly stable and targetable nanoliposomal formulation
of topotecan. J. Control. Release 141, 13–21.
Fenske, D. B., and Cullis, P. R. (2005). Entrapment of small molecules and nucleic
acid-based drugs in liposomes. Methods Enzymol. 391, 7–40.
164 Dmitri B. Kirpotin et al.
Gabizon, A., Catane, R., Uziely, B., Kaufman, B., and Safra, T. (1994). Prolonged circula-
tion time and enhanced accumulation in malignant exudates of doxorubicin encapsulated
in polyethylene-glycol coated liposomes. Cancer Res. 54, 987–992.
Haran, G., Cohen, R., Bar, L. K., and Barenholz, Y. (1993). Transmembrane ammonium
sulfate gradients in liposomes produce efficient and stable entrapment of amphiphathic
weak bases. Biochim. Biophys. Acta. 1151, 201–215.
Harding, J. A., Engbers, C. M., Newman, M. S., Goldstein, N. I., and Zalipsky, S. (1997).
Immunogenicity and pharmacokinetic attributes of poly(ethylene glycol)-grafted immu-
noliposomes. Biochim. Biophys. Acta. 1327, 181–192.
Hayes, M. E., Drummond, D. C., Kirpotin, D. B., Zheng, W. W., Noble, C. O.,
Park, J. W., Marks, J. D., Benz, C. C., and Hong, K. (2006). Genospheres: Self-
assembling nucleic acid-lipid nanoparticles suitable for targeted gene delivery. Gene
Ther. 13, 646–651.
Huang, Z., Park, J. I., Watson, D. S., Hwang, P., and Szoka, F. C., Jr. (2006). Facile
synthesis of multivalent nitrilotriacetic acid (NTA) and NTA conjugates for analytical and
drug delivery applications. Bioconjug. Chem. 17, 1592–1600.
Huang, Z., Hwang, P., Watson, D. S., Cao, L., and Szoka, F. C., Jr. (2009). Tris–
nitrilotriacetic acids of subnanomolar affinity toward hexahistidine tagged molecules.
Bioconjug. Chem. 20, 1667–1672.
Kirpotin, D., Park, J. W., Hong, K., Zalipsky, S., Li, W.-L., Carter, P., Benz, C. C., and
Papahadjopoulos, D. (1997). Sterically stabilized anti-HER2 immunoliposomes: Design
and targeting to human breast cancer cells in vitro. Biochemistry 36, 66–75.
Kirpotin, D. B., Drummond, D. C., Shao, Y., Shalaby, M. R., Hong, K., Nielsen, U. B.,
Marks, J. D., Benz, C. C., and Park, J. W. (2006). Antibody targeting of long-circulating
lipidic nanoparticles does not increase tumor localization but does increase internalization
in animal models. Cancer Res. 66, 6732–6740.
Mamot, C., Drummond, D. C., Noble, C. O., Kallab, V., Guo, Z., Hong, K.,
Kirpotin, D. B., and Park, J. W. (2005). Epidermal growth factor receptor-targeted
immunoliposomes significantly enhance the efficacy of multiple anticancer drugs
in vivo. Cancer Res. 65, 11631–11638.
Mui, B. L.-S., Cullis, P. R., Evans, E. A., and Madden, T. D. (1993). Osmotic properties of
large unilamellar vesicles prepared by extrusion. Biophys. J. 64, 443–453.
Mui, B. L.-S., Cullis, P. R., Pritchard, P. H., and Madden, T. D. (1994). Influence of plasma
on the osmotic sensitivity of large unilamellar vesicles prepared by extrusion. J. Biol.
Chem. 269, 7364–7370.
Nellis, D. F., Ekstrom, D. L., Kirpotin, D. B., Zhu, J., Andersson, R., Broadt, T. L.,
Ouellette, T. F., Perkins, S. C., Roach, J. M., Drummond, D. C., Hong, K.,
Marks, J. D., et al. (2005a). Preclinical manufacture of an anti-HER2 scFv-PEG-
DSPE, liposome-inserting conjugate. 1. Gram-scale production and purification.
Biotechnol. Prog. 21, 205–220.
Nellis, D. F., Giardina, S. L., Janini, G. M., Shenoy, S. R., Marks, J. D., Tsai, R.,
Drummond, D. C., Hong, K., Park, J. W., Ouellette, T. F., Perkins, S. C., and
Kirpotin, D. B. (2005b). Preclinical manufacture of anti-HER2 liposome-inserting,
scFv-PEG-lipid conjugate. 2. Conjugate micelle identity, purity, stability, and potency
analysis. Biotechnol. Prog. 21, 221–232.
Nielsen, U. B., Kirpotin, D. B., Pickering, E. M., Hong, K., Park, J. W., Refaat
Shalaby, M., Shao, Y., Benz, C. C., and Marks, J. D. (2002). Therapeutic efficacy of
anti-ErbB2 immunoliposomes targeted by a phage antibody selected for cellular endocy-
tosis. Biochim. Biophys. Acta. 1591, 109–118.
Nielsen, U. B., Kirpotin, D. B., Pickering, E. M., Drummond, D. C., and Marks, J. D.
(2006). A novel assay for monitoring internalization of nanocarrier coupled antibodies.
BMC Immunol. 7, 24.
Noble, C. O., Kirpotin, D. B., Hayes, M. E., Mamot, C., Hong, K., Park, J. W.,
Benz, C. C., Marks, J. D., and Drummond, D. C. (2004). Development of ligand-
targeted liposomes for cancer therapy. Expert Opin. Ther. Targets 8, 335–353.
Noble, C. O., Guo, Z., Hayes, M. E., Marks, J. D., Park, J. W., Benz, C. C.,
Kirpotin, D. B., and Drummond, D. C. (2009). Characterization of highly stable
liposomal and immunoliposomal formulations of vincristine and vinblastine. Cancer
Chemother. Pharmacol. 64, 741–751.
Ostrowsky, N. (1993). Liposome size measurements by photon correlation spectroscopy.
Chem. Phys. Lipids 64, 45–56.
Pagnan, G., Stuart, D., Pastorino, F., Raffaghello, L., Montaldo, P., Allen, T., Caabretta, B.,
and Ponzoni, B. (2000). Delivery of c-myb antisense oligodeoxynucleotides to human
neuroblastoma cells vis disialoganglioside GD2-targeted immunoliposomes: Antitumor
effects. J. Natl. Cancer Inst. 92, 253–261.
Park, J. W., Hong, K., Kirpotin, D. B., Colbern, G., Shalaby, R., Baselga, J., Shao, Y.,
Nielsen, U. B., Marks, J. D., Moore, D., Papahadjopoulos, D., and Benz, C. C. (2002).
Anti-HER2 immunoliposomes: Enhanced efficacy attributable to targeted delivery.
Clin. Cancer Res. 8, 1172–1181.
Parr, M. J., Ansell, S. M., Choi, L. S., and Cullis, P. R. (1994). Factors influencing the
retention and chemical stability of poly(ethylene glycol)-lipid conjugates incorporated
into large unilamellar vesicles. Biochim. Biophys. Acta. 1195, 21–30.
Poul, M. A., Becerril, B., Nielsen, U. B., Morisson, P., and Marks, J. D. (2000). Selection
of tumor-specific internalizing human antibodies from phage libraries. J. Mol. Biol. 301,
Raffaghello, L., Pagnan, G., Pastorino, F., Cosimo, E., Brignole, C., Marimpietri, D.,
Montaldo, P. G., Gambini, C., Allen, T. M., Bogenmann, E., and Ponzoni, M.
(2003). In vitro and in vivo antitumor activity of liposomal Fenretinide targeted to
human neuroblastoma. Int. J. Cancer 104, 559–567.
Sapra, P., and Allen, T. M. (2002). Internalizing antibodies are necessary for improved
therapeutic efficacy of antibody-targeted liposomal drugs. Cancer Res. 62, 7190–7194.
Sapra, P., and Allen, T. M. (2003). Ligand-targeted liposomal anticancer drugs. Prog. Lipid
Res. 42, 439–462.
Sapra, P., Tyagi, P., and Allen, T. M. (2005). Ligand-targeted liposomes for cancer treat-
ment. Curr. Drug Deliv. 2, 369–381.
Stahl, P. H., and Wermuth, C. G. (eds.). (2002). Monographs on acids and bases. In
“Handbook of Pharmaceutical Salts: Properties, Selection, and Use.” Wiley-VCH,
Weiheim, 2002, pp. 265–327.
Sugano, M., Egilmez, N. K., Yokota, S. J., Chen, F. A., Harding, J., Huang, S. K., and
Bankert, R. B. (2000). Antibody targeting of doxorubicin-loaded liposomes suppresses
the growth and metastatic spread of established human lung tumor xenografts in severe
combined immunodeficient mice. Cancer Res. 60, 6942–6949.
Taggar, A. S., Alnajim, J., Anantha, M., Thomas, A., Webb, M., Ramsay, E., and
Bally, M. B. (2006). Copper-topotecan complexation mediates drug accumulation into
liposomes. J. Control. Release 114, 78–88.
Zamboni, W. C., Strychor, S., Joseph, E., Walsh, D. R., Zamboni, B. A., Parise, R. A.,
Tonda, M. E., Yu, N. Y., Engbers, C., and Eiseman, J. L. (2007). Plasma, tumor, and tissue
disposition of STEALTH liposomal CKD-602 (S-CKD602) and nonliposomal CKD-602
in mice bearing A375 human melanoma xenografts. Clin. Cancer Res. 13, 7217–7223.
Zhou, Y., and Marks, J. D. (2009). Identification of target and function specific antibodies
for effective drug delivery. Methods Mol. Biol. 525, 145–160, xv.
Zhou, Y., Zou, H., Zhang, S., and Marks, J. D. (2010). Internalizing cancer antibodies from
phage libraries selected on tumor cells and yeast-displayed tumor antigens. J. Mol. Biol.
166 Dmitri B. Kirpotin et al.