Novel biodegradable cationic core shell nanoparticles for codelivery of drug and DNA 2

6


Chapter 2
Literature Review

The previous chapter gave a brief introduction of gene delivery and its current
challenges as well as my research objective. The purpose of this chapter is to review and
explore topics pertinent to the main thrust of my research. The general focus in my study
is on polymer micelles and non-viral vectors, in particular, cationic polymers, for
delivery of drugs or genes and codelivery of drugs and genes. I will also examine some of
the difficulties involved with current gene delivery systems using non-viral vectors.

2.1 Polymeric micelles
Micellization phenomena produced by amphiphilic materials have been observed long
before and micellization has been an important area of research in colloid science
[Wihelm M., Gao Z., et. al. 1991]. Initially the study began from low molecular
amphiphiles such as ionic surfactants [Wihelm M., Gao Z., et. al. 1991]. In the last few
decades, micelles fabricated by amphiphilic block copolymers began to attract much
attention from both academia and industry. A great number of papers have reported
mechanism of self-assembling, properties and characterization methods of polymeric
micelles since 1970s’. For example, Tuzar et. al. found that block copolymers of
polystyrene (PS) and poly(ethylene oxide) (PEO) could form spherical micelles in water
when the length of soluble PEO is significantly longer than that of the insoluble PS
portion of the molecule [Tuzar Z. et. al. 1976]. Wihelm et. al. developed a fluorescence

7

probe technique to study the critical micelle concentration [Wihelm M. et. al. 1991]. Gao
and Eisenberg established a model of micellization for block copolymers in aqueous

solution [Gao Z. et. al., 1993]. Astafieva et. al. investigated the critical micellization
phenomena in block polyelectrolyte solutions [Astafieva I. et. al., 1993]. Chen et. al.
studied the effect of block size and sequence on the micellization of ABC triblock
methacrylic polyampholytes [Chen W-Y. et. al., 1995]. This section will give a review on
mechanism of micelle formation, micelle structure, methods for measuring the critical
micelle concentration (CMC) of polymeric micelles, and biomedical applications of
micelles.

2.1.1 Mechanism of polymeric micelle formation
Formation mechanism and properties of polymeric micelles have been well studied.
Micelle formation is driven by two opposing forces including an attractive force between
the amphiphiles leading to aggregation and a repulsive force that prevents unlimited
growth of the micelles into a distinct macroscopic phase [Astafieva I., 1991].
Thermodynamically, micelle formation is mainly due to positive standard entropy of
micellization. The formation of polymeric micelles in aqueous solution is influenced by
the chain length of hydrophilic and hydrophobic blocks, temperature and ions presented
in the aqueous solution.

2.1.2 Structure of polymeric micelles
A simple but important finding obtained by the studies in the past few decades is that
the polymeric micelles usually have core-shell structure (see Figure 2.1) in aqueous

8

solution [Jones M-C., 1999]. The core consists of hydrophobic segments of copolymer
while the shell is made of hydrophilic segments including ionic species [Astafieva I.,
1993; Luisi P. L. et. al., 1984; Pileni M.P., 1989]. This kind of core-shell model is not the
only structure that amphiphilic copolymers may have. Under some conditions they can
also form long and rod like micelles, although this structure is more common for small
molecular weight amphiphiles [Astafieva I., 1993; Price C. 1983; Price C. et. al., 1986;

Canham P.A. et. al., 1980]. In non-aqueous solution, the core-shell model can even
inverse completely, i.e. a hydrophilic core and hydrophobic corona [Astafieva I., 1993;
Price C. 1983; Price C. et. al., 1986; Canham P.A. et. al., 1980].
In fact, the core-shell concept of polymeric micelles was theoretically borrowed from
those formed by small molecular weight amphiphiles. However, this core-shell structure
can be evidenced by some techniques such as fluorescent probes and
1
H-NMR. For
example, fluorescence of bis(1-pyrenyl-methyl)ether (dipyme) [Winnik F. M., et. al.
1992) and pyrene [Wihelm M. et. al., 1991] is sensitive to the polarity change in their
local environments. Therefore, by studying the fluorescent change of these compounds,
one can know the polarity and hence the hydrophobicity of the core.
1
H-NMR can also
provide some information about the core-shell structure [Jones M-C. et. al., 1999]. The
1
H-NMR spectrum of a copolymer in a solvent (e.g. CDCl
3
) where micelle formation is
not expected should exhibit the characteristic peaks corresponding to the hydrophilic and
hydrophobic segments of the polymer. However, in D
2
O, the presence of micelles with a
highly inner viscous hydrophobic core results in a restricted motion of the protons as
demonstrated by the weak signals associated with the hydrophobic segments of the
copolymer [Jones M-C. et. al., 1999; Nakamura K., 1977; Bahadur P., et. al., 1988].

9

2.1.3 Critical micelle concentration (CMC) and its measurements
CMC is one of the most important parameters of the polymeric micelles. It is well
known that the micelles exist only above a certain concentration, i.e., the critical micelle
concentration. However, the CMC characterization of amphiphilic polymer is slightly
different from that of small molecular amphiphiles. The characterization of CMC is not
only essential as the evidence of the formation of micelles but also very important for
polymeric micelles as a drug delivery system, this will be discussed later.
A number of techniques on the determination of CMC were reported. In principle, one
can use any physical property that shows sudden changes at or near CMC. Most
frequently, breaks or discontinuities in plots of such properties as the surface tension,
electrical conductivity, osmotic pressure, interfacial tension, or light scattering as a
function of polymer concentration have been used for this purpose [Wihelm M., et. al.
1991; Jones M-C., et .al. 1999]. However, since the CMC of block copolymers is
normally much lower than that of low molecular weight surfactants, many techniques,
which are suitable to low molecular weight amphiphiles, cannot be simply extended to
amphiphilic copolymers [Wilhelm M. et. al., 1991]. For example, light scattering is
known as one of the most powerful techniques for the determination of the size, shape,
and aggregation numbers of micelles as well as values of the diffusion coefficient of the
micelles. However, the scattering techniques are usually not sensitive enough to detect
the particles at the CMC as low as that of block copolymers [Wilhem M. et. al., 1991].
Gel permeation chromatography (GPC) under aqueous conditions has been employed to
study the CMC of amphiphilic copolymers since single chains and micellar chain

10

fractions of copolymers exhibit different elution volumes [Weissig V. et. al., 1998;
Yokoyama M. et. al., 1993]. It is also possible to determine molecular weight and
aggregation number of micelles by GPC. It is important to note that the integrity of
polymeric micelles should be maintained during their elution through the size exclusion
column. However, adsorption of the polymer on the column may be a problem

[Yokoyama M. et. al., 1993], especially at concentrations close to the CMC, where
micelles consist of large loose aggregates.
Because of the shortages of these techniques, another method was developed based on
the fluorescent probe techniques [Kalyanasundaram K. et. al., 1977; Wilhelm M., et. al.
1991]. The fluorescent probe used is usually polarity-sensitive compound such as pyrene
(see Figure 2.2 for its chemical structure). Pyrene is a condensed aromatic hydrocarbon,
which is highly hydrophobic and sensitive to the polarity of the surrounding environment.
Below the CMC, pyrene is solubilized in water, a medium of high polarity. When
micelles form at a polymer concentration above the CMC, pyrene partitions preferentially
into the hydrophobic domain afforded by the micellar core and thus, experiences a non-
polar environment [Kalyanasundaram K., et. al., 1977; Jone M-C., 1999]. Consequently,
numerous changes such as an increase in the fluorescence intensity, a change in the
vibrational fine structure of the emission spectra and a red shift of the (0,0) band in the
excitation spectra, can be observed. The apparent CMC can be obtained from the plot of
the fluorescence intensity ratio of pyrene such as the I
1
(the first peak)/I
3
(the third peak)
ratio from emission spectra or the I
333
/I
338
(peaks at 333 and 338nm) ratio from excitation
spectra, against polymer concentration. A major change in the slope indicates the onset of
micellization [Figure 2.3] [Jone M-C., 1999; Kalyanasundaram K. et. al., 1977]. The I
1
/I
3

11

ratio is measured at a constant excitation wavelength (339nm) and variable emission
wavelengths corresponding to I
1
and I
3
. Many researchers have applied the fluorescent
probe technique to various polymeric amphiphiles [Kalyanasundaram K. et. al., 1977;
Wihelm M., Gao Z., et. al. 1991; Astafieva I. et. al., 1993]. Some claimed that CMC
might be better ascertained by the I
333
/I
338
ratio since the I
1
/I
3
ratio is affected by the
wavelength of excitation and may result in an erroneous CMC [Shin I. L., et. al., 1998;
Astafieva I., 1993]. The CMC determined with fluorescence techniques needs to be
carefully interpreted for two reasons. Firstly, the concentration of pyrene should be kept
extremely low (10
-7
M) so that a change in slope can be precisely detected as
micellization occurs. Secondly, a gradual change in the fluorescence spectrum can
sometimes be attributed to the presence of hydrophobic impurities or association of the
probe with individual polymer chains or permicellar aggregates [Chen W. Y. et. al.,
1995]. Changes in anisotropy of fluorescent probes have also been associated with the

onset of micellization [Zhang X., et. al., 1996].

Figure 2.1 Schematic representations of block and random copolymer micelles. [Jones
M-C. et. al., 1999]

12



Figure 2.2 Molecular structure of pyrene.
360 370 380 390 400 410
0
50
100
150
200
250
300
350
400
450
I
3
I
1
Fluorescent Intensity
Wavenumber (nm)
300 310 320 330 340 350 360
0
100

200
300
400
500
600
700
Fluorescent Intensity
Wavenumber (nm)

Figure 2.3 Emission spectra (left, excited at λex=339nm) and excitation spectra (right,
monitored at λem=390nm) of pyrene.
The reason of specially reviewing the two characteristics of polymeric micelles, the low
critical micelle concentration and core-shell structure in aqueous phase, is that the two
characteristics of polymeric micelles are especially significant for the micelles to be
applied as drug carrying system. The following section will review in detail on how
researchers have applied polymeric micelles’ characteristics for drug delivery.

2.1.4 Polymeric micelles as drug delivery carrier
Delivery systems for drugs used in the treatment of human diseases have had an impact
on nearly every branch of medicine including cardiology, endocrinology, oncology,
ophthalmology and immunology. Generally speaking, successful drug therapy may be

13

achieved in three ways: delivering the drug efficiently to the target, modifying the drug
for increased efficiency, or finding a novel drug of inherently high efficacy. Of the three,
devising an efficient means of delivery is the most cost-effective. In the search for
delivery systems, Yolls and his co-worker reported the use of lactide-based copolymers
for drug delivery in 1970 [Yolls S., 1975]. Such devices have become much more varied
since then. Sustained-release tablets [Ng S.Y., et. al., 2000; Ng S.Y , et. al., 1997;

Sintzel M.B., et. al., 1998], polymeric matrices (e.g. rods, discs and cylinders) [Hilton
A.K., et. al., 1993], microparticles [Yang Y.Y., et. al., 2000; Chia H.H., et. al., 2001]
hydrogels [Jeong B. et. al., 2000], lipsomes [Zalipsky S., 1995], nanoparticles [Lee J. H.,
2003] and drug-polymer conjugates [Lu Z-R. et. al., 2000] are the commonly used
formulations for drug delivery. The polymeric micelles were first proposed as a drug
carrier by Bader et. al. in 1984 [Bader H., et. al. 1984]. Polymeric micelles exhibit a
number of advantages over other forms of drug carriers because of their core-shell
structure, low CMC and targeting ability.

2.1.4.1 Core of polymeric micelles as a reservoir for hydrophobic therapeutics
The core of the polymeric micelles can serve as a reservoir for an insoluble drug.
Incorporation of insoluble drugs into the core of micelles can be achieved by chemical
conjugation or by physical entrapment through dialysis or emulsification techniques.
Simple equilibration of a drug and micelles in water may not result in high levels of
incorporated drug [Kwon G. S. et. al., 1995; 1997]. Chemical conjugation implies the
formation of a covalent bond, such as amide bond, between specific groups on the drug
and the hydrophobic core of the polymer. Such bonds are normally resistant to enzymatic

14

cleavage mainly because of steric hindrance and thus difficult to be hydrolyzed unless a
spacer group is introduced [Ulbrich K. et. al., 1987]. On the other hand, the chemical
conjugation of a drug to another compound may change its pharmacokinetics and
pharmacodynamics. Therefore, the incorporation of a drug by a physical procedure is
preferred. Physical entrapment of the drug is generally done by a membrane dialysis
method or an oil-in-water emulsion procedure [Jones M-C. et. al., 1999]. For the dialysis
method, a drug and a copolymer are dissolved in a solvent (e.g. ethanol or N, N-
dimethylformamide) in which they are both soluble. The mixture is then dialyzed against
water by using a membrane. As the solvent is replaced by water, the hydrophobic
segments of the polymer and the drug molecules interact and associate to form the core of

micelles while the hydrophilic segments arrange towards the aqueous phase to form the
shell. In the case of the oil-in-water emulsion method, a drug and a copolymer are
dissolved in a water-insoluble volatile solvent (e.g. dichloromethane) and the solution is
then added to an aqueous phase with stirring to form an oil-in-water emulsion. The drug-
loaded micelles are formed as the solvent evaporates. The main advantage of the dialysis
process over the emulsion method is that the use of toxic solvents such as chlorinated
solvent can be avoided. The drug loading level of micelles could reach up to 8%-25%,
depending on the fabrication techniques [Kwon G. S. et. al., 1995;1997], chemical
structure of the drug and polymer, temperature and pH. For example, an increase in the
length of a hydrophobic segment of polymer facilitates the entrapment of a hydrophobic
drug in the core [Torchilin V. P., 2001]. Encapsulation efficiency of the drug also
depends on initial drug loading [Kwon G. S. et. al., 1995; 1997] and aggregation number
of copolymer. G. S. Kwon and T. Okano reported that by using oil-in-water emulsion

15

method, the encapsulation efficiency of Doxorubicin reached 65% [Kwon G. S., 1996].
In my study, the encapsulation efficiency of indomethacin reached 80% (see Chapter 4).
Theoretically, the loading capacity and encapsulation efficiency of polymeric micelles
depend on partition coefficiency of drug between the core phase and aqueous phase
[Torchilin V. P., 2001].
It should be noted that although polymeric micelles have mostly been studied as
delivery systems for drugs, they could also be used to carry plasmid DNA, anti-sense
oligonucleotides or diagnostic agents [Torchilin V. P., 2001; Kataoka K. et. al., 2001].

2.1.4.2 Nanoscale size and hydrophilic shell enabling passive targeting
Another characteristic of polymeric micelles is their nanoscale size and hydrophilic
nature of the shell. The size of drug-loaded polymeric micelles made from amphiphilic
copolymers was reported to range from 10 to 200 nm [Kataoka K. et. al., 2001; Jones M.
et. al., 1999; Trubetskoy V. S. et. al., 1996; Shin I. L. et. al., 1998; Yokoyama M. et. al.,

1992]. The size and size distribution of polymeric micelles can be measured using
dynamic light scattering (DLS) or observed directly under transmission electron
microscopy (TEM) [Yu B. G. et. al., 1998], scanning electron microscopy (SEM) [Kim S.
Y. et. al., 1998] or atomic force microscopy (AFM) [Cammas S. et. al., 1997; Kohori F.
et. al., 1998]. Kwon G. S. and Kataoka K. reported the micelles made from PEO-b-BM
[block copolymer micelles containing poly(ethylene oxide)] having a hydrodynamic
diameter of 10-30nm [Kwon G. S. et. al., 1995]. The reversible, secondary association of
a fraction of PEO-BM occurred giving an aggregate diameter of about 100nm. Another
group of researchers studied a variety of drugs and tracers loaded micelles fabricated

16

using PEO-b-BM block copolymer [Jones M. C. et. al., 1999]. The size of these
polymeric micelles ranged from 15 to 165nm. Polydispersity of polymeric micelles is
usually very low compared with the copolymers used to form the micelles. For example,
the polydispersity index of acetal-PEG-b-PLA [poly (ethylene glycol-b-poly(lactic acid)]
micelles was reported to be 0.03 [Nagasaki Y., et. al., 1998], which is even narrower than
that of the polymer synthesized by anionic polymerization.
The nanoscale size of the polymer micelles together with the hydrophilic nature of the
shell not only allows easy sterilization by filtration and minimizes the risks of embolism
in capillaries, contrary to larger drug carriers [Kwon G. S. et. al., 1996] but also makes
polymeric micelles especially advantageous to avoid fast renal clearance and other
barriers such as the reticuloendothelial system (RES) and mononuclear phagocyte system
(MPS), prolonging blood circulation time. Another important advantage that polymeric
micelles can provide is their passive targeting ability provided via enhanced permeability
and retention effect at tumor and inflammatory sites (i.e. EPR effect) [Matsumura Y.,
1986]. This is because there are increased vascular permeability and impaired lymphatic
drainage at the pathological site and this cause accumulation of the drug at the
pathological site. Drug targeting will be discussed in detail in Section 2.1.3.

2.1.4.3 Low CMC of polymeric micelles enabling long time circulation in blood
stream
Amphiphilic copolymers usually possess a much lower critical micelle concentration
(CMC) compared to low molecular weight surfactants. The low CMC of the polymers is
a great advantage as the stability of their micelles would be unaffected under conditions

17

of extreme dilution, as would be the case in the physiological environment upon
administration. For instance, the CMC of PEO-b-PBLA and PNIPA-b-PSt was reported
to be between 0.0005%-0.002% [La S. B. et. al., 1996; Cammas S. et. al., 1997].
Therefore, stable delivery systems mean long circulation time in blood. In section 2.1.5, it
will be presented that long circulation time in blood is one of the key prerequisites for
efficient passive targeting. Torchilin V. P. pointed out that it is important to differentiate
between thermodynamic and kinetic stability of polymer micelles [Torchilin V. P., 2001].
The CMC provides the information about thermodynamic stability. Below the CMC,
equilibrium between unimers and micelles is shifted towards unimer formation, rendering
the dissociation of core-shell structure. On the other hand, the kinetic stability is related
to the actual time that micelles dissociation takes. Even upon dilution to a concentration
below CMC, the preformed micelles can still exist long enough to perform their carrier
function. The kinetic stability depends on many factors including the physical state of the
micelle core, the contents of a solvent inside the core, the size of the hydrophobic
segment and the ratio of the hydrophobic segment to hydrophilic segment [Tian M., et.
al., 1993; Wang Y. et. al., 1995; Creutz S. et. al., 1998; Yokoyama M. et al., 1996].

2.1.5 Targeting ability of polymeric micelles as drug carrier
Two important issues that concern researchers for an effective drug therapy include
temporal and distribution control of drug. The temporal control refers to the ability to
adjust the duration of drug release or to the possibility to trigger the drug release at a
specific time. Temporal control can be implemented by introducing temperature-sensitive

function to the hydrophilic block of the copolymer [Cammas-Marion S. et. al., 1999;

18

Chung J. E. et. al., 2000; Topp M. D. C., 1997] or by incorporating auxiliary agents such
as channel proteins [Meier V. et. al., 2000; Gelder P. V. et. al., 2000] and magnet
nanoparticles [Sershen S. R. et. al., 2000; Kato N. et. al., 1998; Simpson C. R. et. al.,
1998], which are sensitive to external stimuli (electrolytes, IR light or magnetic field).
The distribution control is also called drug targeting, which is aimed to direct the drug to
the desired site where pharmacological activities are required.
Drug targeting may be classified into two categories: active and passive targeting.
Active targeting aims to increase delivery of drugs to the target by utilizing biologically
specific interactions such as antigen-antibody and ligand-receptor binding. Biological
signals used in this method include antibodies and ligands. On the other hand, passive
targeting is defined as a way to increase the amount of drugs delivered to targeted tissues
by minimizing non-specific interactions with non-target organs, tissues, and cells.
Carriers included in passive targeting utilize physicochemical interactions such as
hydrophobic and electrostatic interactions and physical factors of carriers such as size and
mass.
With advances in the chemistry of amphiphilic polymers, polymeric micelles can be
designed to have reactive functional groups on their surfaces for conjugation of pilot
molecules. Thus, the stage is set for promising research in active drug targeting. However,
over the past decade, passive targeting has shown greater success than active targeting.
Passive targeting achieved much higher in vivo selectivity due to the EPR effect
[Torchlin V. P., 2001]. The evidence showed that drug-loaded polymeric micelles
preferably accumulated to a greater extent than free drug into tumors and displayed a
reduced distribution in non-targeted sites [Weissig V. et. al., 1998]. That is because

19

malignant or inflamed tissues have impaired lymphatic drainage, increasing vascular
permeability. The tumor vessels are also leakier and less permeation selective than
normal vessels. This phenomenon is called the EPR effect. However, implementation of
in vivo selective drug delivery with an active targeting system is not as easy as a simple
extension of in vitro specificity to selective cells. Several difficulties obstruct this simple
extension, including loss of antibody specificity caused by chemical conjugation and/or
enzyme degradation in the blood, non-specific uptake of drug-antibody conjugates, and
increased antigenicity of antibody or antibody-drug conjugates stemming from multiple
doses.
Moreover, the passive component of drug targeting is important in active targeting
systems for the following three reasons: (1) the majority of a living body comprises non-
target sites. Non-specific target sites may capture the major dose fraction. Therefore,
minimization of non-specific capture at non-target sites may be important to maximize
amounts of drugs delivered by active targeting systems; (2) passive transfer phenomena
precede biologically specific interactions even in most active targeting systems.
Exceptions are cases for intravascular targets such as lymphocytes and vascular
endothelial cells. Most targets are located in the extravascular space. To reach these
targets via the bloodstream, the first step must be translocation through the vascular
endothelia, followed by permeation through the interstitial space to the extravascular
targets. Even for active targeting systems based on biologically specific receptors on cells
such as tumor-specific antigens, the passive transendothelium step is a necessary pre-
requisite; (3) the passive aspects may become very important as drugs are introduced to
active targeting carriers. For antibody-drug conjugates, the adverse physicochemical

20

influence of conjugated drugs on the specificity of the antibody is expected to increase
with an increase in the amount of conjugated drugs. Therefore, minimization of non-
specific interactions is important in both active and passive targeting. More precisely,
inhibition of non-specific distribution/excretion at non-target sites is a critical issue in

active and passive targeting. In living bodies, two major routes for this non-specific
distribution/excretion are present: renal excretion and uptake by RES. Evasion of the
renal excretion can be achieved relatively easily by conjugation of drugs to polymeric
carriers or encapsulation of drugs in polymeric carriers or liposomes. However, evasion
of the RES uptake has not been attained as easily as that of renal excretion. One way to
resolve this problem is to use polymeric micelle structures [Yokoyama M. et. al., 1996;
Karaoka K. et. al., 1993; Okano T. et. al., 1994]. Polymeric micelles possess discretely
separated two phases. The functions generally required by drug carriers can be distinctly
shared by structural separated dual phases of the micelles. The inner core plays the role of
a drug loading/release depot for pharmacological activities, while the outer shell is
responsible for interactions with biocomponents such as cells and proteins. These
interactions determine biodistribution and pharmacokinetic behavior of this drug carrier
system. In vivo delivery of drugs may therefore be controlled by the outer shell segment
independently of the micelle inner core that expresses pharmacological activities.
Polymeric micelles can suppress unfavorable (mainly hydrophobic) interactions between
drugs and the RES systems to evade RES uptake because of the presence of the
hydrophilic shell. Some researchers have also reported that the size is a key factor to
evade the uptake by RES. Mononuclear phagocyte system (MPS) namely, fixed
macrophages of liver and spleen, is another barrier that may shorten the blood circulation

21

time of the drug. Nanoscopic drug carriers can evade recognition and uptake by MPS,
and circulate for a more prolonged period of time.
In summary, the size and relatively high stability of polymeric micelles enable them to
evade the renal clearance and uptake by the RES and MPS, prolonging their blood
circulation and achieving passive targeting. In addition, the shell of polymeric micelles
can be designed to possess functional groups for conjugation of biological signals to
achieve active targeting.

2.1.6 Amphiphilic copolymers used for fabrication of micelles
Both random and block amphiphilic copolymers can self-assemble into spherical
micelles in aqueous solution. A-B diblock copolymers have been the main focus for the
fabrication of micelles. Multiblock copolymers such as poly(ethylene oxide)-
poly(propylene oxide)-poly(ethylene oxide) (PEO-PPO-PEO) were also able to self
organize into micelles [Malmsten M. et. al., 1992; Prasad K. N., 1979].
The hydrophobic core generally consisted of a biodegradable block such as poly(β-
benzyl-L-aspartate) (PBLA), poly(DL-lactic acid) (PDLLA) or poly(ε-caprolactone)
(PCL). The core was also made from a water-soluble polymer [e.g. poly(aspartic acid)]
P(Asp)), which was rendered hydrophobic by the chemical conjugation of a hydrophobic
drug [Yokoyama M. et. al., 1990; 1992; 1996], or formed through the association of two
oppositely charged polyions (polyion complex micelles) [Harada A. et. al., 1995; 1998;
Kataoka K., 1996]. Some non or poorly biodegradable polymers such as polystyrene (PS)
[Zhao C.L. et. al., 1990; Zhang L. et. al., 1995] or poly(methyl methacrylate) (PMMA)
[Inone T. et. al., 1998) were also used as an core-forming block. In addition, a highly

22

hydrophobic small chain such as alkyl chain [Ringsdorf H. et. al., 1992; 1991; Yoshioka
H. et. al., 1995] or a diacyllipid [e.g. distearoyl phosphatidyl ethanolamine (DSPE)]
[Trubetskoy U. S. et. al., 1995] was employed as the core-forming segment. The
hydrophobic chain can be either attached to one end of a polymer [Winnik F. M. et. al.,
1995] or randomly distributed within the polymeric structure [Ringsdorf H. et. al., 1992;
Schild H. G. et. al., 1991].
The shell is generally made from hydrophilic, non-biodegradable but biocompatible
blocks such as PEG. PEG is the most commonly used hydrophilic block for polymeric
micelles. Some other hydrophilic materials with special functions were also employed as
the shell-forming block such as poly (N-isopropylacrylamide) (PNIPAAm) [Cammas S.
et. al., 1997; Chung J. E. et. al., 1997; 1998] and poly(alkylacrylic acid) [Chen W. Y. et.
al., 1995]. PNIPAAm is thermally sensitive while poly(alkylacrylic acid) is pH-sensitive

and could eventually be used to confer bioadhesive properties [Inoue T. et. al., 1998].
Micelles presenting functional groups at their surface for conjugation with a targeting
moiety have also been described in the literatures [Scholz C. et. al., 1995; Kabanov A.V.
et. al., 1992; Nagasaki Y. et. al., 1998; Cammas S. et. al., 1995]. These functional
micelles as discussed in the previous section can help to realize active targeting.

2.2 Gene delivery
Gene therapy has recently received increasing attention especially for combating
cancers. In clinic, the cancers are normally treated by chemo and/or radiation therapy.
However, these strategies exhibit general limitations, especially at the latter stage of
cancer development when metastases have already spread to remote sites of the body. For

23

example, poor specificity of most chemotherapeutic drugs may result in suppression of
the bone marrow and other fast dividing tissues as well as potential genesis of secondary
cancer. Another drawback of chemotherapy is the development of resistant phenotypes,
which no longer respond to chemotherapy. It is also well known that chemotherapy and
radiation therapies can cause serious side effects to healthy tissues or organs. With the
development of the human genome project, the molecular mechanism of cancer has been
substantially studied. This opens the possibility of using gene therapy for treating cancers.
Several strategies have been proposed for manipulation of gene expression either on the
transcriptional or translational level. A deficient gene can be replaced, and the effect of
an unwanted gene can be blocked by the introduction of a counteracting gene. For
instance suicide gene therapy offers the perspective to kill cancer cells selectively by
using prodrug-converting enzymes and tumor specific promoters. Furthermore, antisense
and ribozyme strategies offer the potential to selectively downregulate the expression of
specific genes mainly on the translational level predominantly by sequencing specific
interaction with messenger RNAs.
In fact, most diseases have a genetic component. Hence, gene therapy holds the hope of

curing, not merely treating, a broad range of ailments, including inherited diseases like
cystic fibrosis and even chronic conditions like cancer and infectious diseases like AIDS.
Since the first genetic treatment on September 14, 1990 [Thompson l., 2000], the hope of
gene therapy has inspired a great burst of enthusiasm. However, the main challenge of
gene therapy lies in the development of safe and efficient gene delivery system. This
section will give a review of various types of gene delivery vectors.


24

2.2.1 Gene delivery vectors
Gene delivery vector is the vector that deliver gene into the cell. The simplest and safest
way to deliver gene is to transport the transgene (i.e. naked DNA) to the target site
without the use of vectors. So far, skeletal muscle, liver, thyroid, heart muscle, urological
organs, skin and tumor have been explored as a portal for direct gene injection
[Nishikawa M. et. al., 2001; 2002]. Various physical manipulations such as
electroporation, bioballistic (gene gun), ultrasound and hydrodynamics (high pressure)
have been proposed to improve the efficiency of gene delivery. However, due to the rapid
degradation of DNA by nuclease in the serum and clearance by the MPS, the gene
expression level and the area induced by direct injection of naked DNA were generally
limited. It is known that delivery of the naked gene can only be rarely applied with
reasonable efficiency, as in intramuscular gene therapy [Mumper R. J. et. al., 1998].
Therefore, it is important to develop an efficient system for gene delivery and gene
expression, which is applicable to both basic research and clinical settings.
Basically, gene delivery systems can be classified into two categories: viral and non-
viral vectors [Luo D. et. al., 2000; Merdan T. et. al., 2002; T. Niidome et. al., 2002;
Thomas M. et. al., 2003]. Gene delivery by viral vectors includes employing various
viruses as the vehicles to carry gene to the target cells or organs and express in the host
cell or organ. This process is called infection. Viral vector appeared before non-viral
vector as gene delivery system. The first major gene therapy success was the retrovirus-

based treatment of infants suffering from the X chromosome linked severe combined
immune deficiency (SCID-X1), displaying the potential of long-term or even permanent
cure of hereditary disease [Cavazzana-Calvo M. et. al., 2000]. Because of their highly

25

evolved and specialized components, viral systems are by far the most effective means as
gene delivery vehicles, which achieve high efficiencies (usually >90%) for both delivery
and expression. For example, some retroviruses can integrate their genetic information
into the cell nucleus very effectively. By replacing the natural genetic information of the
virus with the genes desired to express in the defective cells, a genetic disease, at least in
principle, can be cured [Lundstrom K., 2003]. However, viral-mediated delivery has its
own limitations, including toxicity, restricted targeting of specific cell types, limited gene
carrying capacity, difficulties in production and packaging, recombination and high cost
[Luo D. et. al., 2000]. Furthermore, the toxicity and immunogenicity of viral systems also
hamper their routine use in basic research laboratories. The random insertions of
transgene may also result in mutations in the host, leading to carcinogenesis [Thomas M.
et. al., 2003]. The long-term effect of the integrated transgene and the virus in the host is
another serious concern [Thomas M. et. al., 2003]. Therefore, the safety issue makes the
great prospects of viral vector uncertain [Lundstrom K., 2003].
Because of these limitations of viral vectors, non-viral vectors with lower cytotoxcity
and immune response have been purposed. Compared with viral vectors, non-viral
vectors usually do not integrate the transgene into a chromosome of the target cell and
only maintain as an extra-chromosomal genetic element (episome). Therefore, the non-
viral vectors, unlike their viral counterparts, are safe and also amenable to large-scale
production.

2.2.2 Non-viral gene vectors

26

Non-viral gene delivery vectors include cationic lipids, cationic polymers and inorganic
materials. Inorganic materials such as gold nanoparticles [Yang N. S. et. al., 1995] and
multisegment bimetallic nanorods [Salem A. K. et. al., 2003] have been reported to
deliver DNA. However, inorganic materials have attracted very limited attention in the
gene therapy community, possibly due to their poor versatility and biocompatilibility
problem. Cationic lipids and cationic polymers are by far the most widely used vectors in
non-viral gene delivery. Although polycations have been used for the insertion of foreign
DNA into cells long before liposome [Henner W. D. et. al. 1973; Ehrlich M. et. al. 1976;
Fraley R. et. al. 1980], lipids used as a gene vector were more mature than polycations
[Thomas M. et. al., 2003]. Lipsome-based gene delivery formulations were first reported
by Fraley R. et. al. in 1980. In the 1990s, a large number of cationic lipids, such as
quaternary ammonium detergents, cationic derivatives of cholesterol and diacylglycerol
and lipid derivatives of polyamines, were reported [Niidome T. et. al., 2002]. Among
these lipids, lipofectAMINE
TM
exhibits high transfection efficiency. However, it is quite
toxic. The development of new types of lipid molecules appears to be saturated. Most of
the efforts have shifted to improve efficiency of gene delivery and expression by
modification of the existing lipids or pursuing novel polymer-based carriers [Pedroso de
Lima M. C. et. al., 2001; Niidome T. et. al., 2002]. It seems that cationic polymers are the
most promising non-viral vector and are attracting great interest. Recent studies have
shown that polycations are potentially superior to liposomal formulations in many
respects. For example, the formation of the lipoplex (lipsome-DNA complexes) involves
interaction among lipid molecules, in addition to that with DNA itself. A major driving
force for the complex formation is the release of low-molecular weight counter-ions that

27

make a large entropic contribution to the free energy of binding [Matulis D. et. al. 2002].

The hydrophobic segments of lipids are the key determinant in the macroscopic
characteristics of the ensuing liposomes, particularly their size, shape, and stability in the
dispersed state, as well as interactions with other lipids, cell membranes, and DNA. This
in turn affects the transfection efficiency of the resulting lipoplexes [Smisterova J. et. al.,
2001; Zuhorn I.S. et. al., 2002]. Approaches for controlling these parameters are limited,
leading to instability of their macroscopic properties over time [Simberg D. et. al., 2001]
and thus restricting their pharmaceutical potential. Furthermore, liposomal formulations
often require an adjuvant, such as di-oleylphosphatidylethanolamine, for efficient gene
delivery [Hui S. W. et. al., 1996].
Compared with lipids, self-assembly of polyplexes ( polycation-DNA complexes) does
not entail interaction of the polycation molecules with each other, resulting in greater
control of their macroscopic properties, and is quite efficient even without any adjuvants.
Furthermore, polycations can be easily tailored by chemical methods to achieve high
efficiency or cell targeting. Many of the polyplexes have superior transfection efficiency
and serum sensitivity compared to lipoplexes [Gebhart C. L. et. al., 2001]. These
advantages make polycations a compelling target for future exploration in non-viral gene
delivery.

2.2.3 Polycations
Compared with lipids, the greatest advantage of cationic polymers as gene vectors are
that cationic polymers can be easily tailored and synthesized to suit the special
requirements encountered by gene delivery. Some natural polymers have also been

28

employed to deliver genes. In general, the most commonly used cationic polymers as
gene vectors include branched and linear polyethyleneimine (PEI), copolymers of PEI,
poly(L-lysine) and its copolymers, imidazole-modified poly(L-lysine) and chitosan
(Figure 2.4).

Figure 2.4 Cationic polymers most frequently used for nucleic acid delivery. [Merdan T.
et. al., 2002]
Polyethyleneimine (PEI)
PEI polymers have become the gold standard of non-viral vectors. PEI with different
molecular weights and degrees of branching have been synthesized and evaluated in vitro
as well as in vivo. Branched PEI with a molecular weight of 25 kDa is often used as a
control to evaluate gene expression efficiency of other non-viral vectors because of its
high efficiency in gene transfection [Marschall P. et. al., 1999; Campeau P. et. al., 2001].
PEI polymers are able to effectively condense DNA molecules, leading to homogeneous
spherical particles with a size of ~100nm. This enables efficient in vitro gene transfection

29

[Marschall P. et. al., 1999; Campeau P. et. al., 2001]. PEI offers a significantly more
efficient protection against nuclease degradation than other polycations such as poly(L-
lysine), possibly due to its high charge density and efficient complexation with DNA.
The huge amount of positive charges, however, results in a rather high toxicity, which is
one of the major limiting factors especially for its in vivo use. The high density of
primary, secondary and tertiary amino groups exhibiting protonation only on every third
or fourth carbons at pH 7 confers significant buffering capacity to the polymer over a
wide range of pH [Godbey W. T. et. al., 1999]. This property, known as ‘proton sponge’
is the most likely contributed to the high transfection efficiency obtained using PEI-based
polymers [Boussif O. et. al., 1995]. Support for this assumption is that endosomal
acidification is required for efficient gene transfection [Midoux P. et. al., 1999]. Despite
this recognized association, knowledge about relationships between polymer structure
and important biological properties such as toxicity or gene transfection efficiency is
rather limited. An increased molecular weight of PEI led to higher gene transfection as
well as greater cytotoxicity [Fischer D. et. al., 1999].
The linear PEIs have also been reported [Ferrari S. et. al., 1997; Coll J. L , et. al.,
1999]. It has been demonstrated that the linear PEI (Mw 21 kDa) yielded excellent gene

transfection efficiency with a rather low toxicity. Linear PEI has recently been reported to
mediate a cell cycle independent nuclear entry of plasmid DNA [Brunner S. et. al., 2002].
This finding is of particular importance in the therapy of slow dividing tissues.

Poly(L-lysine)

30

Poly(L-lysine) was one of the first polymers used in non-viral gene delivery. Due to
the peptide structure of poly(L-lysine) it is biodegradable, making it suitable for in vivo
use. However, the polymer exhibits moderate to high toxicity. With poly(L-lysine) of
suitable molecular weights and optimized N/P ratios, complexes with plasmid DNA
displayed a size of ~100nm [Wolfert M. A. et. al., 1999] and were taken up into cells as
efficiently as PEI complexes [Merdan T. et. al., 2002]. However, the transfection
efficiency was several orders of magnitude lower. A possible reason is the lack of amino
groups with pKa between 5 and 7, resulting in no endosomolysis and low levels of
transgene expression [Merdan T. et. al., 2002].

Imidazole-containing polymers
The heterocycle imidazole displays a pKa of about 6, possessing a buffering capacity in
the endolysosomal pH range, and thus possibly mediating vesicular escape by a ‘proton
sponge’ mechanism. Therefore, polymers containing the heterocycle imidazole have
shown promising transfection capability. In several approaches, modification of ε-amino
groups of poly(L-lysine) using histidine or other imidazole-containing structures showed
a significant enhancement of reporter gene expression compared to poly(L-lysine) [Fajac
I., Allo J.C. et. al., 2000; Benns J. M. et. al., 2000; Midoux P. et. al., 1999].

Chitosans
Chitosan is a biodegradable and linear aminopolysaccharide with randomly distributed
beta linked N-acetyl-D-glucosamine and D-glucosamine, derived from the common

biopolymer chitin. Chitosan displays a significantly better biocompatibility than PEI.