| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Microbiology and Immunology, University of Leicester, United Kingdom;
Tumor Targeting Group, Section of Oncology & Cellular Pathology, Division of Genomic Medicine, University of Sheffield Medical School, United Kingdom; and
YCR Cancer Research Unit, Department of Biology, University of York, United Kingdom
Correspondence: Dr. B. Burke, Department of Microbiology and Immunology, Maurice Shock Building, University of Leicester, University Road, Leicester LE1 9HN, UK. E-mail: bb14{at}leicester.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGENE TRANSFER TO MACROPHAGESMACROPHAGES AS GENE DELIVERY...MACROPHAGES AS IN VIVO...CONCLUDING REMARKSREFERENCES |
|---|
Key Words: vector • adoptive immunotherapy • transfection • homing • transcriptional targeting
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGENE TRANSFER TO MACROPHAGESMACROPHAGES AS GENE DELIVERY...MACROPHAGES AS IN VIVO...CONCLUDING REMARKSREFERENCES |
|---|
Many of the events regulating the infiltration and functions of macrophages in such diseased tissues have now been identified. This has prompted attempts to transfer therapeutic gene constructs into macrophages to use these cells in adoptive immunotherapy protocols or to alter their deleterious or defective activities in vivo. Most studies have used gene transfer methods to manipulate gene expression in macrophages in vitro. However, as is often the case, attempts to replicate these studies in the more complex in vivo environment have proved difficult with high-level, macrophage-specific transfection proving problematic.
|
GENE TRANSFER TO MACROPHAGES |
|---|
|
TOPABSTRACTINTRODUCTIONGENE TRANSFER TO MACROPHAGESMACROPHAGES AS GENE DELIVERY...MACROPHAGES AS IN VIVO...CONCLUDING REMARKSREFERENCES |
|---|
Viral methods
Adenoviruses
Adenoviruses have become a popular method of gene transfer to
primary macrophages due to their ability to infect nondividing cells
with high efficiency and reasonable longevity (up to several weeks) of
transgene expression [3
]. Various studies have reported
transfection efficiencies of 10–80% with human macrophages, depending
on cell culture conditions and the multiplicity of infection (MOI;
number of infectious virus particles per cell) used
[4
5
6
7
]. It is noteworthy that adenoviruses appear to be
relatively ineffective in infecting monocytes, presumably due to the
paucity of appropriate integrins (the receptors for adenoviruses) on
these cells [5
]. A recent study using an adenoviral
vector to transfer interleukin (IL)-18 antisense into the murine
monocytic cell line RAW 264.7 only achieved a transfection efficiency
of 20% [8
]. One study solved the integrin problem by
incubating primary human monocytes in 100 ng/ml macrophage-colony
stimulating factor (M-CSF) for 72 h to stimulate expression of the
integrins 
vß3 and
vß5. With this method, using MOIs of 100
or 50, infection rates of greater than 90% were achieved
[9
].
Retroviruses
Retroviruses (with the exception of the lentivirus subclass) are
generally considered ineffective in the transfection of
nonproliferating or poorly proliferating cells such as primary
macrophages. One explanation for this may be that such viruses are
incapable of entering the nuclei of nonproliferating cells, and thus
are unable to integrate their genome into the chromosomes of the
infected cell, which is essential for significant levels of retroviral
gene expression. Parveen and co-workers [10
] have
addressed this problem by introducing a nuclear localization signal
(NLS) sequence into the matrix protein of the C type retrovirus spleen
necrosis virus (SNV). This enabled the SNV vector to penetrate the
nucleus of primary human monocyte-derived macrophages, allowing
infection of 90% of cells compared with 0% for the control virus
without the NLS.
There is some evidence that gene expression from integrated retroviruses often lasts only a few weeks, especially in vivo, which may be due to "transcriptional silencing" caused by methylation of DNA near the promoter and by integration of the virus genome into condensed chromatin regions in which the transgene would be inaccessible to the transcription machinery (reviewed in ref [3 ]). The recent discovery of two types of DNA element, locus control regions and ubiquitously acting chromatin opening elements, which alter the conformation of the surrounding chromatin in such a way as to enhance expression of nearby genes, offers a potential solution to the problem of insertion into transcriptionally inactive regions of chromatin [3 ].
CD34+ bone marrow stem cells, precursors of macrophages
(and other leukocytes), are capable of cell division, rendering them
susceptible to retroviral infection. Murine leukemia viruses have most
often been used for this, but other retroviral vectors have also
recently been developed, for example based on human foamy virus, which
can give a transfection efficiency of up to 80% [11
].
Other studies have shown that retrovirally transduced CD34+
cells continue to stably express transgenes after being induced to
differentiate into macrophages. In one study, stable expression of a
reporter gene in 7–29% of CD14+ macrophages (as well as
4–23% of CD1a+/CD14- dendritic cells) was
achieved by retroviral transduction of granulocyte (G)-CSF-mobilized
peripheral blood CD34+ cells. These had been grown in the
presence of IL-3, IL-6, GM-CSF, and stem-cell factor prior to being
induced to differentiate into macrophages (and dendritic cells) with
GM-CSF and tumor necrosis factor (TNF-) [12
]. The
ability to stably transfect these precursor cells opens up the
possibility of reinfusing such genetically modified stem cells into a
patient, thereby providing a circulating (and possibly bone marrow
populating) source of therapeutically modified macrophages (and other
leukocytes that would develop from these precursors in vivo).
Other retroviruses of the lentivirus group including HIV are able to infect and integrate into the chromosomes of nonproliferating cells including monocyte/macrophages [13 , 14 ], and therefore have the potential to effect stable, long-term gene transfer to such cells. In the case of HIV, sequences in the viral proteins Matrix, Vpr, Integrase, and Pol have been implicated in enabling nuclear import of the viral genome in nondividing cells [15 ]. However, it should be noted that productive HIV infection (i.e., producing progeny virus particles) of monocytes is reported to require cell proliferation [16 , 17 ]. There is also evidence that lentiviral vectors offer the advantage of being less prone to transcriptional silencing than retroviral vectors (reviewed in ref [3 ]). Although HIV seems to offer the most effective, current method for the stable transfection of DNA into macrophages, there are safety concerns over its use (even when using replication-deficient variants) due to the danger of generating replication-competent viruses by recombination. However, considerable effort is being focused on the development of novel packaging systems that could significantly reduce this risk [18 ].
Adeno-associated viruses (AAV)
AAV have potential for gene therapy as they exhibit a number of
attractive features including their ability to integrate into the
genome of host cells and mediate long-term expression of transgenes (up
to 2 years in mice), and to target proliferating and nonproliferating
cells [3
]. AAV have recently been used to efficiently
transduce monocytes and dendritic cells [19
,
20
].
Poxviruses
A highly attenuated poxvirus, modified vaccinia Ankara, has been
used to transfer genes expressing tumor-specific antibodies into
GM-CSF-activated macrophages in vitro [21
,
22
]. In addition, a herpes simplex virus-1-derived vector
has been used to transfer transgenes to macrophage cell lines
[23
].
Limitations of current viral vectors
A potentially important limitation in the use of adenoviral,
retroviral/lentiviral, and AAV vectors is that they impose relatively
low size limits (approximately 7.5 Kb, 8 Kb, and 4.5 Kb, respectively)
on the amount of foreign DNA that can be incorporated. In contrast, the
use of poxviruses such as vac-cinia is not hampered by this
problem because these viruses have much larger genomes. It should be
noted that the clinical use of certain viral vectors, notably
adenoviruses and AAV, may prove to be limited due to the immunogenicity
of the virus-infected cells. In the case of adenoviruses, "gutless"
vectors, which lack parts of the viral genome and therefore express
fewer potentially immunogenic viral proteins, are under development.
Such vectors have the additional advantage of a greater capacity for
heterologous DNA, although these vectors present a number of problems
in terms of manufacture (reviewed in ref [3
]).
A problem limiting the use of viral vectors in transfecting macrophages in vivo is their lack of specificity for macrophages—most infect a broad range of cell types. However, promising recent studies have shown progress in altering the cell tropism of adenoviruses to enhance their ability to transduce target cells in specific diseased sites. One of these studies produced a bispecific single chain antibody or "diabody," which binds a viral protein (the knob domain of the adenovirus fiber proteins on the virus surface, which are responsible for adenovirus-cell interactions) at one end and a cell type-specific surface protein (CD105) at the other [24 ]. This antibody, effectively an artificial virus cell-binding receptor, acts as an adapter between the virus and the target cell type. Other workers, such as Krasnykh et al. [25 , 26 ] have taken the more direct route of "genetic targeting" adenoviruses by genetically manipulating the genes coding for the adenoviral fiber protein. Such work could potentially lead to the development of adenoviruses targeted to specific cell types such as macrophages.
Nonviral methods
A number of nonviral methods have been tested for the transfer of
exogenous nucleic acids (plasmid DNA, oligonucleotides, and ribozymes)
into macrophages, including electroporation and nonimmunogenic,
synthetic DNA carriers such as liposomes, lipoplexes, and
diethylaminoethyl (DEAE)-dextran, and they will be discussed in the
following sections. However, these methods usually give low
transfection efficiencies and short durations of transgene expression
in target cells in vitro. They also have only limited use for systemic
application. Cationic liposome/plasmid DNA or oligonucleotide
complexes, for example, are rapidly cleared from the circulation, with
the highest levels of activity usually observed in organs such as the
lungs, spleen, and liver—that is, tissues with a known role in the
removal of particulate and foreign matter from the bloodstream.
Interactions between macrophages and naked DNA
Macrophages have been shown to take up exogenous DNA, possibly via
a specific transport mechanism. Most of this is thought to be degraded
in endosomes [27
]. However, exogenous DNA has been
reported to be transported to the nucleus and expressed by a macrophage
cell line (RAW 264.7) in vitro when used at very high concentrations.
Maximal reporter gene expression was observed at a DNA concentration of
500 µg/ml [28
]. Gene transfer to macrophages using
naked DNA has also been reported in vivo in a murine wound model
[29
], although in this study there was no cell-type
specificity, with adipocytes being the predominant cell type expressing
the transgene, followed by macrophages and fibroblasts.
However, scenarios exist in which constraints such as the lack of target cell specificity and low levels of transgene expression may not be important, as in genetic immunization. For example, Condon and co-workers [30 ] showed that cutaneous immunization with naked DNA, attached to microscopic particles and projected forcefully into the skin by a burst of gas from a "gene gun," results in transfection of dendritic cells in the skin, giving potent, antigen-specific, cytotoxic T lymphocyte (CTL)-mediated protective tumor immunity.
Another finding of significance in the transfection of macrophages is that these cells have specific mechanisms that recognize bacterial DNA because of the presence in this of unmethylated CpG dinucleotides (i.e., a cytosine followed by a guanine), which are rare in eukaryotic DNA (reviewed in ref [31 ]). Macrophages respond to such DNA by undergoing activation similar to that caused by lipopolysaccharide (LPS). This could be a complicating factor in gene therapy protocols using plasmid DNA or even oligonucleotides containing CpG dinucleotides, as unmethylated CpG DNA is possibly the most potent adjuvant known [32 ]. The problem could possibly be avoided by treating the DNA to be transfected with a methylase before transfection, which has been shown to be effective in the case of plasmid DNA [31 ].
Electroporation
Electroporation involves passing a brief pulse of electricity
through a suspension of cells and exogenous DNA. The electric pulse is
thought to cause the formation of pores in the cell membrane, allowing
the cells to take up the exogenous DNA, which is believed to move
directly into the nucleus, bypassing the endosomal degradation pathway.
It has been noted that the conditions applied, such as the voltage and
capacitance, ionic strength and temperature of the medium, and quantity
of DNA used have to be optimized for each cell type
[33
]. It has also been reported that the proportion of
cells in S phase is important in transfection of granulocyte-macrophage
progenitors by electroporation [34
].
Electroporation has achieved moderate levels of transfection efficiency for human monocytic cell lines such as MonoMac 6 (40–50%; B. Burke and C. E. Lewis, unpublished observations) and monocyte-derived macrophages (>30%; ref. [33 ]) in vitro, but appears less effective with promonocytic/leukemic cell lines (<3%; refs. [35 , 36 ]). The main disadvantage of electroporation is the high level of cell death. This ranges from 5 to 60% with cell lines within 72 h after electroporation (depending on the electropulsing method used [35 , 36 ]), to 30 to 75% within 24 h using primary cells [33 ]. However, few studies have been able to confirm that electroporation is an effective way of transfecting primary macrophages in vitro.
Li et al. [37 ] used electroporation to transfect CD34+ precursor cells in vitro and used a postpulse-pelleting method and caspase inhibitors to reduce osmotic swelling and cell death, achieving a transfection efficiency of approximately 20%. It remains to be seen whether this method will prove effective with macrophages.
Liposomes, lipoplexes, and cationic compounds
These transfection reagents offer a number of advantages. They
have low toxicity, deliver DNA of essentially unlimited size, and do
not evoke an immunogenic or inflammatory response when used in vivo.
However, attempts at using lipid-based reagents to transfect primary
macrophages or myeloid cell lines in vitro have generally yielded poor
transfection efficiencies (<5%), with transgene expression lasting no
more than 24 h [36
]. However, when Mack et al.
[38
] optimized their DEAE-dextran DNA transfection
method for adherent primary human macrophages, they found transgene
expression could be maintained for up to 56 h.
In one study, which compared the efficiency of various liposomal and nonliposomal agents for transfecting the murine macrophage cell line RAW 264.7, LipofectAMINE (in combination with the DNA-condensing agent protamine sulfate) was found to be the most effective, followed by Lipofectin (fourfold less), DOTAP (tenfold less), and DEAE-dextran (20-fold less) [39 ]. The percentage of cells transfected was not given, but rather the total luciferase activity expressed by transfected cells was measured. This is a common approach in transfection experiments, but makes it impossible to assess the proportion of cells transfected, as "high" levels of reporter gene activity could arise from a relatively small proportion of the target cells. The use of a reporter gene such as green fluorescent protein (GFP) would allow transfection efficiency to be calculated using flow cytometry or fluorescence microscopy.
Cationic liposome/DNA complexes have also been shown to be capable of transfecting monocytes/macrophages in vivo in the blood, liver, and spleen. However, the transfection was nonspecific, and endothelial cells throughout the body were the main recipient of the transgene, with other leukocytes also being transfected [40 , 41 ].
The use of liposomes is not restricted to the transfer of plasmid DNA.
They have also been used to transfect primary macrophages with
antisense oligonucleotides and ribozymes in vitro [42
].
Cationic lipids have also been used to deliver ribozymes to murine
peritoneal macrophages in vitro and in vivo with the aim of blocking
TNF- production, and achieved a 70% decrease in release of this
cytokine from macrophages in treated versus untreated mice
[43
]. However, administration of the lipid/RNA complexes
produced a fourfold increase in the numbers of peritoneal exudate
inflammatory cells, which could counteract the desired effect of
reducing inflammation by reducing TNF- expression. Also, the
duration of the inhibitory effect on TNF- was not examined, and only
6% of the ribozyme was taken up by macrophages, the rest being taken
up by organs such as spleen, lung, liver, pancreas, and intestines.
One recent study has used a novel DNA-condensing cationic peptide, CL22, to transfect monocyte-derived dendritic cells in vitro with genes encoding tumor-associated antigens, and showed that these cells then protected mice from a normally lethal challenge with melanoma cells [44 ].
Receptor-mediated gene transfer
Following the detailed characterization of receptor expression by
macrophages over the last two decades, some transfection methods have
been adapted to target specific endocytotic pathways in these cells.
Ligands such as mannose and transferrin have been incorporated into
gene transfer vehicles and have been shown to markedly increase the
efficacy of transfection for primary macrophages in vitro
[45
46
47
]. Erbacher and co-workers [45
]
observed much higher reporter gene expression with human macrophages in
vitro using mannosylated polylysine/DNA complexes than with
DEAE-dextran or Lipofectin. Simoes et al. [46
]
determined that the transfection efficiency for primary human
monocyte-derived macrophages in vitro increased from essentially zero
to 2% of cells when using transferrin.
The main advantage afforded by the use of receptor-targeted DNA complexes is that they can be designed to target specific cell types based on the presence of cell type-specific receptors on the cell surface. This has important implications for the clinic as this method could potentially allow direct injection of these complexes into the bloodstream and negates the need for ex vivo gene transfer. Indeed, in vivo gene transfer to macrophages in the liver, spleen, and lungs [48 ] was achieved using mannosylated polylysine-conjugated plasmids. Kawakami and co-workers [49 ] showed that this was also possible using mannosylated cationic liposomes.
Microorganisms as vehicles for transfection of macrophages
Various bacterial and protozoan microorganisms have evolved the
ability to infect macrophages, evade their antipathogen-defense
mechanisms, and establish chronic infections. Thus, these organisms
represent a potentially powerful method of transferring therapeutic DNA
constructs to macrophages in vivo. Several of these intracellular
microorganisms have been exploited for this purpose.
Macrophages are one of the main cellular targets in
Salmonella infection in humans, and as Salmonella
can easily be manipulated in vitro to carry plasmid DNA, various
workers have attempted to use them to transfect macrophages. For
example, Paglia and co-workers [50
] developed an
attenuated (nonpathogenic) derivative of Salmonella
typhimurium bearing a plasmid encoding murine interferon-
(IFN-) under the control of the cytomegalovirus (CMV) immediate
early promoter. They showed that exposure to such bacteria in vitro or
in vivo resulted in the augmented expression of IFN- by macrophages.
In another report, S. typhimurium were successfully used to
transfect human monocyte-derived macrophages with transfection
efficiencies of 85–95% [51
].
A similar approach may be possible using Leishmania, a protozoan parasite. Although the wild-type strains of this microorganism can be highly pathogenic, attenuated mutants have been developed that can safely be used as live vaccines [52 ]. Vaccaro [53 ] suggested that an attenuated Leishmania strain, engineered to express a therapeutic gene, could be used to infect macrophages in individuals suffering from genetic disorders primarily affecting macrophages. As Leishmania specifically target the lysosomal compartment of macrophages, it could prove useful in the treatment of lysosomal storage disease (LSD), which involve defects in lysosomal enzymes.
The Gram-positive bacterium, Listeria monocytogenes, proved relatively ineffective in transferring reporter genes to the macrophage-like cell lines, J774/A1 and M97, or primary murine macrophages in vitro [54 ]. This accords well with earlier reports of low transfection rates using Listeria with bone marrow-derived macrophages (BMDM) in vitro and of macrophages in vivo following intraperitoneal (i.p.) injection [55 ]. The reason(s) for the low transfectability of macrophages using Listeria remain to be elucidated.
|
MACROPHAGES AS GENE DELIVERY VEHICLES |
|---|
|
TOPABSTRACTINTRODUCTIONGENE TRANSFER TO MACROPHAGESMACROPHAGES AS GENE DELIVERY...MACROPHAGES AS IN VIVO...CONCLUDING REMARKSREFERENCES |
|---|
Ability of macrophages to home to diseased sites after ex vivo
manipulation
Adoptive immunotherapy with macrophages has been attempted in
several studies with varying degrees of success. In a number of animal
studies, infusion of macrophages (following ex vivo activation by such
treatments as LPS, IFN, and exposure to medium conditioned by tumor
cells) into tumor-bearing mice was shown to induce tumor regression
[56
57
58
]. Macrophage adoptive transfer is a relatively
risk-free procedure. Cancer patients have been reinfused with up to
3 x 109 ex vivo-activated macrophages with only mild
side effects, but unfortunately these studies have shown only minimal
therapeutic benefit [reviewed in ref (59)].
Attention began to focus on ways of enhancing the ability of adoptively
transferred macrophages to kill tumor cells or ameliorate other disease
states by making use of their ability to migrate into diseased tissues
to get them to carry therapeutic DNA constructs into such sites.
Various studies have looked at macrophage trafficking after injection
into a host organism, often after ex vivo manipulation, genetically
(e.g., using genes coding for activating cytokines such as IFN-), or
by treatment with activating agents such as cytokines or LPS. The aim
of these studies was to assess their ability to "home" to diseased
sites after local or systemic reimplantation into mice, rats, or humans
[57
58
59
60
]. Homing is a fundamental requirement for their
use as vehicles to target locally acting gene therapy to specific
diseased sites. It is less important for therapeutic strategies
involving ex vivo loading of macrophages with, for example, tumor
antigens, which utilize the antigen-presenting capabilities of
macrophages but do not require them to migrate to specific disease
sites to perform this function.
Routes of reimplantation have included intravenous (i.v.), intra-arterial, intraperitoneal (i.p.), and intrapleural [59 ]. Macrophage homing studies have involved radioactive labeling of macrophages in vitro with isotopes such as Indium 111 or Iodine 125 or expression of a reporter gene [61 62 63 64 ]. Typical findings are that in the short term, up to 2 h after reimplantation, macrophages accumulate primarily in the lungs and, to a lesser extent, in the liver and spleen rather than in the target diseased tissue. Subsequently, a proportion of labeled macrophages in the lungs returns to the circulation and is carried to the liver and spleen. A small proportion of manipulated macrophages, which appears to vary markedly between studies (ranging from 0.2% [63 ] to 28.8% [64 ]) following systemic administration, homes to the diseased site of interest, whether a tumor or a wound. The presence of labeled macrophages in these sites was found to persist for at least 6–7 days in both of these mouse studies. Greater success has been achieved using local administration of macrophages. For example, Chokri and co-workers [64 ] reported that more than 70% of locally injected macrophages homed to the tumor site in mice carrying a subcutaneous melanoma.
In an intermediate means of administration between local and systemic (sometimes referred to as "directed systemic"), rat BMDM and the rat alveolar macrophage cell line, NR8383, were adenovirally infected to overexpress IL-4 and then infused into the renal arteries of rats suffering from experimentally induced glomerulonephritis. Infused macrophages localized to inflamed glomeruli, and the infusion resulted in a remarkable 75% reduction in albuminuria, a sensitive indicator of glomerular damage [65 , 66 ]. Interestingly, the efficiency with which injected macrophages localized to inflamed kidneys was increased markedly by incubation of cells with LPS prior to injection. This means of infusion is unlikely to be routinely possible in humans, but demonstrates the potential usefulness of macrophages as a gene delivery vehicle if the problem of efficient homing to target sites can be resolved.
Macrophages manipulated ex vivo have also been shown to be capable of homing to experimentally damaged muscle, and bone marrow cells from a mouse transgenic for the lacZ reporter gene were capable of engrafting mice and providing a long-term (up to 2 months) supply of genetically modified macrophages capable of migrating into damaged muscle sites when administered i.v. [67 ]. One of the potential advantages of macrophage adoptive transfer is the ability to circulate around the body after i.v. injection and target multiple diseased sites, for example, tumor metastases.
Unfortunately, as described above, several studies have shown that the majority of systemically administered macrophages become trapped in organs such as the lungs, liver, and spleen [60 , 68 , 69 ]. One of these, an early study using murine peritoneal macrophages, found that the organs to which the macrophages homed were determined by the method originally used to elicit them in the peritoneal cavity. Resident macrophages and those elicited by proteose peptone or thioglycollate broth localized initially to the lungs after i.v. reinfusion and then rapidly disseminated to the liver and spleen, but macrophages elicited by Brewer’s thioglycollate medium localized to the lungs and remained there for at least 72 h with little or no migration to the spleen [60 ]. The distribution of macrophages was found to be altered by activation of these cells in vivo through the i.p. injection of the pyran copolymer, MVE-2, prior to adoptive transfer. Activated cells contained a highly differentiated, low-density population of macrophages, which became trapped in the lungs for longer periods of time than nonactivated cells. These cells also exhibited a reduced ability to migrate into the spleen. These different distributions are possibly due to differences in macrophage cell surface proteins induced by the respective treatments. This implies that it might be possible to target macrophages to certain organs or diseased sites by specific pretreatments or even by transfecting them to express particular cell surface-expressed proteins. For example, overexpression of the receptor for macrophage chemotactic protein-1, a cytokine released by many tumors, could possibly be used to maximize migration into such tumors.
Another possible way to enhance macrophage targeting of diseased sites is by using bispecific antibodies. Chokri and co-workers [68 ] showed that a bispecific antibody,which bound both to Fc receptors on the surface of macrophages and also to an adenocarcinoma antigen, increased the tumor cytotoxicity of the macrophages. They proposed that this approach might also be of value in directing the migration of macrophages to particular diseased sites.
It would seem logical to assume that since monocytes, not macrophages, are taken up across the endothelium into normal and pathological tissues, monocytes would prove more effective in "homing" to such tissues. However, work in a murine model has indicated that relatively mature ex vivo-manipulated murine monocytes are capable of migrating into tissues, where they undergo further maturation and cell division [69 ].
Use of transcriptional targeting to overcome the lack of
specificity in macrophage homing to diseased sites
The finding that only a small proportion of reimplanted
macrophages are likely to home to the target diseased site is perhaps
not surprising given the ubiquitous distribution of macrophages around
the body [1
]. However, it emphasizes the need for a
second level of gene targeting, for example, at the transcriptional
level, in order to ensure that genetically manipulated macrophages that
locate to non-target tissues do not express therapeutic genes in these
tissues. This could be achieved using transcriptional control elements
responsive to physiological states or to secreted proteins (e.g.,
cytokines) associated with the diseased tissue to be targeted. The
best-explored type of transcriptional targeting is the use of
hypoxia-responsive elements (HREs) to target expression to solid
tumors, which are known to often contain regions of severe hypoxia (low
oxygen) [70
], and they may also be suitable for
targeting other pathological conditions that are associated with
hypoxia, such as wounds and sites of chronic infection. HREs have been
used to mediate hypoxia-inducible gene expression in macrophages in
vitro [71
, 72
]. The work of Griffiths and
co-workers [71
] in using the macrophage/HRE system to
target tumor spheroids in vitro is discussed later in this review.
Carta et al. [72
] engineered a construct containing the
IFN- gene under the control of three copies of the HRE from the
inducible nitric oxide synthase gene promoter. They used
electroporation to transfect the ANA-1 murine macrophage cell line with
this and found that although these cells secreted basal levels of
IFN- in normoxia (normal levels of oxygen), secretion increased more
than fivefold when cells were exposed to hypoxia (1% O2)
in vitro.
It is not yet known, however, whether hypoxia-responsive enhancers would be sufficiently selective to restrict transgene expression in macrophages strictly to hypoxic areas in vivo, especially as transfected macrophages would be exposed to many other stimuli (including hormones such as insulin, which has been shown to up-regulate hypoxia-inducible factor-1 and thus potentially HRE activity) [73 ] as they pass through body fluids and tissues.
Use of macrophages to deliver therapeutic viruses
An interesting, recent study by Pastorino and co-workers
[74
] transduced the murine macrophage cell line, WGL5,
with a replication-defective retrovirus (bearing the reporter gene
eGFP) and a helper virus. They achieved stable integration, retrovirus
production, and reporter gene expression. A high-expressing clone was
then selected. When the transduced macrophage WGL5 cell line was
administered subcutaneously to allogeneic mice, these cells formed
solid tumors, as might be expected of an immortalized cell line.
CD4+ and CD8+ T cells within the tumor were
shown to be positive for the reporter gene, indicating that recombinant
virus had been released from the transduced cells and was capable of
infecting other cell types. Transduced WGL5 cells were also
administered i.v. and were found not to form solid tumors but rather to
accumulate initially in organs such as the lungs, spleen, and liver.
However, after 24 h, some reporter gene expression was also
observed in the brain. Whether this represented trafficking of
transduced WGL5 cells or free virus to the brain was not clear.
The advantage of this novel approach of using macrophages as "virus factories" is that relatively few macrophages would need to reach the target organ to have a therapeutic effect, due to the amplification effect of each cell producing large numbers of viruses. However, there are important obstacles to be surmounted before this approach is likely to be viable clinically. For example, to transfer the system to primary macrophages rather than a cell line may be difficult because productive retroviral infections require cell division [16 , 17 ]. Secondly, the DNA sequences regulating virus production would need to be modified to incorporate a second level of targeting (e.g., at the transcriptional level), specific to the disease or tissue being treated, to prevent side effects caused by macrophages localizing to non-target tissues and from viruses released into the circulation. Using hypoxia response elements as an "on/off" switch in such a system could possibly enable virus production to be limited to severely hypoxic tumor sites.
Malignant tumors
Macrophages accumulate in large numbers in avascular, hypoxic (low
oxygen) sites in breast [75
, 76
] and
prostate [77
] carcinomas. As mentioned previously,
hypoxia is widespread in malignant human tumors [70
] due
to their poorly organized vasculature and the faster growth of tumor
cells than blood vessels. Macrophages are thought to be attracted by
cytokines released by tumor cells in response to hypoxia and other such
physiological stresses imposed by ischemia (reviewed in ref
[78
]). Exploitation of the tendency of macrophages to
accumulate in such tumor sites, by adoptive transfer of macrophages to
cancer patients for therapeutic gene delivery or antigen presentation,
has long been proposed [79
, 80
].
Recently, Griffiths et al. used an adenoviral vector to transduce primary human macrophages with a gene encoding the prodrug activating enzyme cytochrome P450 2B6 under the transcriptional control of a trimer of an HRE [71 ]. Infiltration of virus-transduced macrophages into tumor spheroids (which contain hypoxic centers) in vitro resulted in a sevenfold decrease in viable tumor cells in the presence of the prodrug cyclophosphamide compared with infiltrated spheroids not treated with the drug. However, the question of whether the basal level of cytochrome produced by the transduced macrophages in normoxic tissues would be low enough to also avoid killing cells in these sites was not examined.
LSD
Eto and Ohashi [81
] infected murine macrophages
derived from the ex vivo expansion of bone marrow cells using a
recombinant adenoviral vector containing the human B-glucuronidase gene
(HBG; a therapeutic gene for one of the LSD) and showed that their
glycosaminoglycan accumulation was markedly reduced in vitro. They also
took macrophages from normal mice (that expressed HBG) and injected
them i.v. into Sly mice (a mouse model for glucuronidase deficiency).
They showed that these cells populate the liver and spleen and, in
doing so, raise HBG enzyme levels in these tissues. However,
adenovirally infected macrophages have yet to be injected into
HBG-negative mice to see whether this is sufficient to correct HBG
deficiency.
.
Alveolar immunodeficiency
Gene therapy for immunodeficiency in the lung is limited, in part,
by the difficulty of transfecting lung cells in vivo. Many options
exist for transfecting cells in vitro, but they are not easily adapted
for use in vivo. To overcome this limitation, macrophages
(specifically, the murine macrophage cell line J774A) were transduced
in vitro with the murine IFN- gene and delivered intratracheally
into immunocompromised scid (severe combined immunodeficiency) mice.
IFN- was detected in bronchoalveolar lavage fluid by 48 h, and
immune function was partially restored in the lungs, with evidence of
enhanced major histocompatibility complex (MHC) class II antigen
expression and increased phagocytosis. However, i.p. administration of
the engineered macrophages did not enhance IFN- levels in the lung.
This study suggests that airway delivery of genetically engineered
macrophages expressing the mIFN- gene may partially restore
significant immune activity in the lungs of immunodeficient mice
[82
].
|
MACROPHAGES AS IN VIVO TARGETS FOR GENE THERAPY |
|---|
|
TOPABSTRACTINTRODUCTIONGENE TRANSFER TO MACROPHAGESMACROPHAGES AS GENE DELIVERY...MACROPHAGES AS IN VIVO...CONCLUDING REMARKSREFERENCES |
|---|
and LPS. They are also capable of phagocytosis of apoptotic
tumor cells and presenting tumor-associated antigens to T cells.
Antigen-specific recognition and subsequent destruction of tumor cells
are the goals of vaccine-based anticancer immunotherapy. However,
macrophage cytotoxicity has been shown to be down-regulated by factors
released by tumor cells, and tumor antigen-specific CTLs are often not
available or are present in an inactive state. In addition, MHC-I
expression by tumor cells is often down-regulated. These situations
would allow tumors to evade tumoricidal mechanisms [83
]. Genes coding for tumor antigen-specific monoclonal antibodies or fragments of them have been cloned into attenuated poxviruses. GM-CSF-activated human macrophages and cytotoxic T cells infected with such viruses acquire the ability to specifically kill tumor cells expressing those antigens in vitro [21 , 22 ]. Whether these novel methods will be applicable in vivo awaits further study, but if so, this method could potentially have a role in anticancer gene therapy, although the high mutation rate of cancer cells makes it likely that variants lacking the epitope would arise.
Although IFN- is one of the most powerful stimulants for macrophage
tumoricidal activity, when this cytokine was evaluated in clinical
trials using ex vivo adoptive cellular immunotherapy protocols, a major
problem encountered was the short duration of ex vivo activation of
macrophages—repeated injections of ex vivo-activated cells were
required to obtain a clinical response. Various studies have tried to
circumvent this problem by transfecting macrophages to overexpress
activating cytokines. Nishihara and co-workers [84
] used
retroviral vectors to engineer a macrophage cell line to express
IFN-, IL-4, IL-6, or TNF- and showed increased in vitro and in
vivo tumoricidal activity by these cells. Ringenbach et al.
[85
] have used polyethylenimine-mediated transfection of
the IFN- gene to enhance the tumoricidal activity of human monocytes
in vitro. Moreover, the cytotoxicity and MHC-II expression of
macrophages were augmented following i.p. injection of
liposome-encapsulated IL-2 and IL-6 DNA into lymphoma-bearing mice
[86
]. However, little attempt was made in these studies
to assess the longevity of gene expression or activation status of
transfected cells in vitro or in vivo.
Evidence for an indirect approach to altering macrophage function in tumors has been provided by Richter et al. [87 ]. Chinese hamster ovary cells were stably transfected with the gene for the immunosuppressive cytokine IL-10. When grown i.p. in mice, these cells showed reduced angiogenesis and tumorigenicity compared with untransfected cells in nude and in SCID mice. The authors suggested that this phenomenon might be linked to the marked reduction in the numbers of macrophages seen in these tumors. Although unproven, this is a plausible hypothesis, as tumor-associated macrophages have been shown to secrete a variety of proangiogenic and prometastatic cytokines and enzymes (reviewed in ref [78 ]), so their elimination from tumors could result in retarded growth.
Arthritic joints
The abundance and activation of macrophages in the inflamed
synovial membrane/pannus correlate closely with the severity of
rheumatoid arthritis [88
]. These cells exhibit
widespread proinflammatory, destructive, and remodeling capabilities
and contribute to the progression of acute and chronic disease.
Furthermore, activation of the monocytic lineage is not locally
restricted, but extends to systemic parts of the mononuclear phagocyte
system. Thus, the selective suppression of macrophage activation is a
possible approach to diminishing local and systemic inflammation, as
well as the prevention of irreversible joint damage.
Macrophage production of the potent proinflammatory molecule TNF-
has been implicated in the pathogenesis of inflammation in arthritic
joints [88
]. For this reason, Kisich et al.
[43
] used cationic, lipid-mediated delivery of ribozymes
to selectively inhibit TNF- production by murine peritoneal
macrophages in vitro (by 80%). They went on to show that following
i.p. injection of cationic lipid/ribozyme complexes, elicited
peritoneal macrophages accumulated the ribozyme, and their
TNF- release in response to LPS was reduced. It remains to be seen
whether such an injection into arthritic joints would suppress TNF-
production by synovial macrophages. Fellowes et al. [89
]
showed that an IL-10 expression plasmid/liposome complex injected i.p.
in mice was taken up and expressed by macrophages in a collagen-induced
model of arthritis. This led to marked and prolonged (up to 30 days
postinjection) amelioration of inflammation in the arthritic joints.
Skin wounds
Macrophages are an important source of mitogenic growth factors
and proangiogenic cytokines and enzymes in healing wounds (reviewed in
ref [90
]). A recent study has shown that it may be
possible to accelerate wound healing (or correct defective
wound-healing processes) by gene transfer to macrophages in wounds.
Meuli et al. [29
] were able to transfect macrophages, as
well as fibroblasts and adipocytes, with the LacZ reporter gene in
surgically wounded mouse skin following local injection of DNA alone or
at a much lower level by i.v. injection of cationic liposomes/DNA
complexes.
LSD
LSD are a group of about 50 monogenic, metabolic disorders caused
by a deficiency in the intralysosomal enzymes involved in macromolecule
catabolism. These defects are most prominently displayed in macrophages
in afflicted patients. CGD is a rare, inherited immunodeficiency
syndrome caused by the inability of macrophages to produce sufficient
reactive oxygen metabolites. This dysfunction is a result of a defect
in reduced nicotinamide adenine dinucleotide phosphate oxidase, the
enzyme responsible for the production of superoxide. It is composed of
several subunits, two of which, gp91phox and p22phox, form the
membrane-bound cytochrome b558, and its three cytosolic components,
p47phox, p67phox, and p40phox, have to translocate to the membrane upon
activation [91
].
gp91phox is encoded on the X-chromosome and p22phox, p47phox, and p67phox, on different autosomal chromosomes, and a defect in any one of these components leads to CGD. Schneider et al. [5 ] used adenovirus-mediated gene transfer to insert a gp91phox gene expressed from a powerful constitutive promoter (CMV) in gp91phox-deficient macrophages. This caused >70% of transfected cells to show respiratory burst activity in vitro and in vivo. These data indicate that autologous macrophages transfected ex vivo or in vivo to express the gp91phox gene may have use in overcoming the life-threatening infections seen in X-CGD patients.
As most of the genes encoding the normal lysosomal enzymes have now been cloned, and the size of the corresponding cDNAs is found to be generally compatible with their transfer by recombinant vectors, macrophages in various forms of LSD may prove to be a viable and clinically useful target in gene therapy protocols [92 ].
Silicotic fibrosis
The finding that TNF- release by alveolar macrophages plays a
central role in the development of inflammation in silicotic fibrosis
prompted Rojanasakul and co-workers [93
] to attempt to
inhibit silica-induced TNF- release by these cells using an
antisense oligonucleotide for TNF- complexed to mannosylated
polylysine (which exploits the endocytotic pathway regulated by the
mannose receptor on macrophages). This inhibited TNF- production by
alveolar macrophages in vitro in the presence of silica. Further
studies are now needed to determine whether systemic or local
application of such complexes would be effective in vivo.
Atherosclerosis
This is an inflammatory disease involving recruitment and
activation of macrophages, smooth muscle cells, and T cells.
Macrophages accumulate low-density lipoprotein (LDL), which is
atherogenic when it undergoes cell-mediated oxidation within the
arterial wall [94
]. Oxidized LDL promotes vascular
dysfunction by exerting direct cytotoxicity toward endothelial cells,
increasing the chemotactic properties of monocytes, transforming
macrophages into foam cells via scavenger receptors (postulated to
enhance their survival), and by enhancing the proliferation of various
local cells such as macrophages and smooth muscle cells. These events
are recognized as important contributing factors in the development of
atherosclerosis. Laukkanen and co-workers [95
] used
adenoviral gene transfer to make murine macrophages express a secreted
form of the human scavenger receptor. This inhibited their ability to
degrade acetylated or oxidized LDL by up to 90% and inhibited their
ability to form foam cells in vitro. Moreover, an adenovirus has been
used to introduce a reporter gene into intimal macrophages (as well as
intimal cells and smooth muscle cells) in atherosclerotic vessels in
organ culture, thereby identifying them as targets for gene transfer in
vivo [96
].
Lung diseases
Ferkol and co-workers [47
] recently attempted
transfer of the -1 antitrypsin gene, an inherited defect in
which is a cause of the chronic lung disease emphysema, to mouse
alveolar macrophages in vivo using mannosylated polylysine-conjugated
DNA. Significant increases in the level of -1 antitrypsin were
achieved in the lungs of treated mice. In another, more recent study,
when an adenovirus encoding heme oxygenase 1 (HO-1) was administered by
direct, intratracheal inoculation into a murine model of acute lung
injury induced by inhaled pathogen, HO-1 was not only expressed
by surface cells in the respiratory epithelium but also by alveolar
macrophages [97
].
HIV infection
Monocytes and macrophages are readily infected by HIV and can
support viral replication, so there have been attempts to transfect
macrophages to render them resistant to this virus. Macrophages derived
from the ex vivo differentiation of CD34+ cells have been
retrovirally transduced with the IFN-ß gene in vitro. This enhanced
their resistance to HIV infection and inhibited the replication of the
virus in these cells, as well as increasing their release of IL-12 and
IFN- (which in turn is likely to stimulate other cellular immune
responses to HIV) [98
].
In another study, macrophages were transfected in vitro using liposomes containing antisense oligonucleotide to the Rev response element (responsible for nuclear export of viral mRNAs) or a ribozyme complementary to the HIV-1 5' long-terminal repeat, which contains an important viral promoter. Both resulted in the inhibition of HIV replication by up to 90% in macrophages in vitro [42 ].
Liver disease
Kupffer cells play a significant role in the pathogenesis of many
inflammatory liver diseases including early alcohol-induced liver
injury. Therefore, a potential therapeutic strategy would be to
modulate the activities of macrophages in the liver via a gene delivery
system. A recent report has described how adenoviral gene transfer can
be used to infect Kupffer cells in the liver in vitro and in vivo
[99
]. However, other cell types are likely to be
infected using this approach, so some form of macrophage-specific,
transcriptional targeting may prove useful to ensure expression of the
transgene in Kupffer cells alone. This could involve placing transgenes
under promoters for general, macrophage-specific genes such as c-fms
(the gene encoding the receptor for M-CSF-1) [100
], the
LPS receptor CD14, CD68 [101
], or genes up-regulated by
macrophages only in diseased sites. Examples of the latter are the
genes for TNF-, transforming growth factor-ß [102
],
or platelet-activating factor [103
], all of which have
been shown to be up-regulated at the mRNA level in Kupffer cells in
liver disease.
|
CONCLUDING REMARKS |
|---|
|
TOPABSTRACTINTRODUCTIONGENE TRANSFER TO MACROPHAGESMACROPHAGES AS GENE DELIVERY...MACROPHAGES AS IN VIVO...CONCLUDING REMARKSREFERENCES |
|---|
Despite interest in the use of macrophages as a gene delivery system, relatively few in vivo studies have been carried out to look at the ability of ex vivo-transfected macrophages to migrate into (and express transgenes in) target sites. Nor have the effects of different differentiation/activation states of the cells used, cell numbers injected, or routes of administration for reinfusion been fully examined. Such studies are now warranted if the value of such novel, macrophage-based, adoptive gene therapies is to be accurately assessed.
The targeting of macrophages in vivo, using various viral and nonviral methods, looks promising, as this utilizes the natural ability of macrophages to take up cell surface ligands (via general phagocytotic or specific receptor-ligand internalization pathways) to ensure uptake of such gene vectors in vivo. However, such vectors cannot be relied on to deliver genes exclusively to macrophages, so transcriptional targeting of transgene expression (for example, using macrophage-specific promoters) may prove essential if gene expression in nontarget cells and thus unwanted side effects are to be avoided. Once again, however, the paucity of clinical studies showing the efficacy of in vivo targeting of transgenes to macrophages means that possible therapeutic applications remain conjecture rather than fact at this stage.
Another salient fact to emerge from this review is that promonocytic and leukemic cell lines are rarely good models to use when assessing gene transfer to macrophages, as methods used for their transfection may prove relatively ineffective in primary macrophages. Moreover, primary cells may express transgenes in a different manner (if at all) compared with these transformed cell lines.
|
ACKNOWLEDGEMENTS |
|---|
Received February 7, 2002; revised April 4, 2002; accepted April 4, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGENE TRANSFER TO MACROPHAGESMACROPHAGES AS GENE DELIVERY...MACROPHAGES AS IN VIVO...CONCLUDING REMARKSREFERENCES |
|---|
