Minimal Toxicity of Stabilized Compacted DNA ... - CiteSeerX

ARTICLE

doi:10.1016/j.ymthe.2003.09.002

Minimal Toxicity of Stabilized Compacted DNA Nanoparticles in the Murine Lung Assem-Galal Ziady,1 Christopher R. Gedeon,2 Osman Muhammad,2 Virginia Stillwell,1 Sharon M. Oette,2 Tamara L. Fink,2 Will Quan,1 Tomasz H. Kowalczyk,2 Susannah L. Hyatt,2 Jennifer Payne,2 Angela Peischl,2 J. E. Seng,3 Robert C. Moen,2 Mark J. Cooper,2 and Pamela B. Davis1,† 1

†

Department of Pediatrics, Case Western Reserve University, Cleveland, Ohio 44106-6006 2 Copernicus Therapeutics, Inc., Cleveland, Ohio 44106 3 Redfield Division, Charles River Laboratories, Redfield, Arkansas 72132

To whom correspondence and reprint requests should be addressed. Fax: (216) 368-4223. E-mail: [email protected].

Nanoparticles containing DNA compacted with poly-l-lysine modified on an N-terminal cysteine with polyethylene glycol can effectively transfect cells of the airway epithelium when applied by the luminal route. To evaluate the toxicity of these nanoparticles, we administered 10 and 100 ␮g DNA compacted into nanoparticles suspended in normal saline by the intranasal route to mice and determined the pulmonary and systemic responses to this challenge, compared to administration of saline alone, and in some experiments, compared to administration of naked DNA, Escherichia coli genomic DNA, or lipofectin-complexed naked DNA. There was no systemic response to either dose of nanoparticles in serum chemistries, hematologic parameters, serum complement, IL-6, or MIP-2 levels or in the activity, growth, and grooming of the mice. Nanoparticles containing 10 ␮g DNA induced responses comparable to saline in all measures, including BAL cell counts and differentials and cytokine levels and histology. However, mice dosed with 100 ␮g DNA in nanoparticles had modest increases in BAL neutrophils 48 and 72 h after dosing, modest increases in BAL IL-6 and KC beginning 24 and 48 h, respectively, after dosing, and, on histology of the lung, a trace to 1ⴙ mononuclear cell infiltrates about the pulmonary veins at 48 h, which were markedly reduced by 10 days and gone by 28 days after dosing. BAL neutrophil and cytokine responses were no greater than those entrained by naked DNA for up to 24 h. However, compared to administration of only 10 ␮g E. coli genomic DNA, the response to compacted DNA was much less. A low dose of lipofectin-complexed DNA (5 ␮g DNA) induced the same response as 20-fold higher doses of DNA nanoparticles. These data indicate that DNA nanoparticles have no measurable toxic effect at a dose of 10 ␮g and a very modest effect, which is not limiting, at a dose of 100 ␮g, which gives maximal gene expression. This favorable toxicity profile encourages development of stabilized compacted DNA for airway administration. Key Words: compacted DNA nanoparticles, PEGylated poly-l-lysine, inflammatory cytokines, CpG motifs, in vivo gene delivery to the lung

INTRODUCTION Nonviral gene transfer into airway epithelial cells has been successful, but for lipid-mediated gene transfer, the inflammatory response becomes limiting at doses at or below those required to achieve genetic correction in patients with cystic fibrosis (CF) [1,2]. Within hours of administration of the lipid–DNA complexes, patients develop fever and a flu-like syndrome, accompanied by elevated serum interleukin (IL)-6 and C-reactive protein. In

948

animal studies, neither the DNA alone nor the lipid alone is strongly inflammatory in the airways, but when the gene transfer complex is formed, significant inflammatory response ensues [3–5]. In mice given lipid–DNA complexes, increases in tumor necrosis factor (TNF)-␣, interferon (IFN)-␥, and IL-12 are observed in the bronchoalveolar lavage (BAL) fluid, as well as increases in BAL cell counts and neutrophils. Because, in animal studies, removing CpG sequences from the plasmid DNA either by mutation or by methylation reduces the inflammatory response, it is be-

MOLECULAR THERAPY Vol. 8, No. 6, December 2003 Copyright © The American Society of Gene Therapy 1525-0016/03 $30.00

doi:10.1016/j.ymthe.2003.09.002

lieved that the inflammation occurs, at least in part, in response to unmethylated bacterial CpG sequences delivered to the cytoplasm of the target cells [3,4,6 – 8]. DNA complexed with polyethyleneimine (PEI) administered to the airways successfully delivers reporter genes in vivo, but also provokes an inflammatory response, with increased TNF-␣ and IL-1␤ in the BAL fluid of treated animals [9]. Inflammatory responses have thus been limiting for successful nonviral gene transfer strategies in the airway, and if an effective nonviral means of gene delivery could be developed that was not limited by the host response, this would constitute an advance. A different strategy for nonviral gene transfer, which does not involve lipids, is also successful in increasing delivery of reporter genes to airway epithelial cells, compared to naked DNA [10]. In this method, single molecules of plasmid DNA are compacted using poly-l-lysine (poly K) covalently linked to polyethylene glycol (PEG) to shield any excess charge of the complex and to prevent aggregation in normal saline solution. These complexes are stable indefinitely in saline, have the same transfection efficiency after storage at 4°C for 2 years, and have a half-life in mouse serum of 2– 4 h [10]. One advantage of compacted DNA is its ability to access the nucleus of nondividing cells, such as those in the airway epithelium [11]. Thus, substantial gene transfer, increased about 200fold over naked DNA, has been observed following intranasal or intratracheal administration of these complexes [10]. However, the inflammatory potential of complexes of compacted DNA has not been formally tested. We report here in vivo tests of the inflammatory response to compacted DNA delivered via the airways to mice.

RESULTS Experimental Design We conducted six separate experiments, five with Balb/C mice and the sixth with C3H/HeBFEJ mice. Mice were dosed with normal saline (25 ␮l), compacted DNA at a concentration of 0.4 mg/ml (25 ␮l, for a dose of 10 ␮g per animal, “low dose”), or compacted DNA at a concentration of 4 mg/ml (25 ␮l, for a dose of 100 ␮g per animal, “high dose”). These doses were selected because the high dose was found, in other studies, to give the maximal gene transfer based on luciferase activity in lung 48 h after dosing [10]. We also studied the response to a dose a log below the maximum, which gives significantly increased gene expression compared to naked DNA, but significantly less than the high dose. In some experiments, we applied naked plasmid or Escherichia coli genomic DNA in the same volume of saline. We administered the dose, in 5-␮l aliquots, to the nose of anesthetized mice maintained in the upright posture, as described previously [10]. There was no immediate mortality associated with administration of the complexes or saline. For experiments 2– 6, we sacrificed animals by pentobarbital overdose and imme-

MOLECULAR THERAPY Vol. 8, No. 6, December 2003 Copyright © The American Society of Gene Therapy

ARTICLE

diately exsanguinated them by cardiac puncture. Animals in experiment 1 were sacrificed by carbon dioxide narcosis. The investigators dosing, sacrificing, and performing cell counts or assays on the animals were unaware of the contents of the coded tubes of gene transfer reagent or saline or the treatment group to which the animals were assigned. The single exception was when genomic DNA was included, for to duplicate preparation in the literature, this tube was stored on ice. Therefore, those who dosed the animals were aware of the treatment assignment of that group, but not the others. However, those who performed the assays were not aware of treatment assignment. In experiments with multiple time points, those who performed the assays were also unaware of the time point at which the samples were taken. All protocols were approved by the local Institutional Animal Care and Use Committees. Experiment 1 In experiment 1, 15 female and 15 male Balb/C mice received saline, 30 females and 30 males received the low dose complex, and 30 females and 30 males, the high dose complex. We weighed the animals on study days 1, 3, 8, 11, 15, 22, and 29. We selected animals at random for sacrifice on days 3, 11, and 29 (2, 10, and 28 days after dosing). All animals grew, fed, and groomed normally throughout the course of the experiment. We sacrificed 5 control mice and 10 experimental mice from each group of each gender at each time point. Animals were exsanguinated and full blood chemistries and hematologic studies were performed. We obtained body weights and organ weights, all of which were comparable within gender for all groups. We conducted histologic examination of the following tissues: brain, heart, lung, liver, kidneys, spleen, and gonads. Except for the lung, which is described below, all examinations were normal. We performed the following serum chemistries: total protein, albumin, globulin, glucose, cholesterol, total bilirubin, blood urea nitrogen, creatinine, sodium, potassium, chloride, ALT, AST, alkaline phosphatase, ␥-glutamyl transferase, calcium, phosphorus, and triglycerides. No abnormal values for serum chemistries were demonstrated at any time point in the treated animals. Values that differed from those of the saline control group at one time point were very similar to the saline controls at other time points and probably represented the statistical variations inherent in multiple comparisons. None of the variations of the treated group was a clinically significant abnormality. Values that differed from those of the saline-treated control group at any time point are shown in Table 2. We made the following measurements: white blood cell and differential counts, red blood cell count, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and platelet count. No clinically significant abnormalities in hematologic parameters were

949

ARTICLE

doi:10.1016/j.ymthe.2003.09.002

TABLE 1: Histologic abnormalities on lung sections from mice treated as indicated (number abnormal/number scored) Dosing group

Day 2

Day 10

Day 28

Saline, male

0/5

0/5

0/5

Saline, female

0/5

0/5

0/5

Compacted DNA (10 ␮g), male

0/10

0/10

0/10

Compacted DNA (10 ␮g), female

0/10

0/10

Compacted DNA (100 ␮g), male Compacted DNA (100 ␮g), female

0/10

a

3/10

0/10

9/10a

0/10

0/10

10/10

a

a

Mononuclear infiltrates graded trace to mild were present around the pulmonary veins at day 2, and only trace grades of infiltrates were present at day 10.

demonstrated at any time point in the treated animals compared to the saline controls. None of the variations of the treated group was outside the normal range nor a clinically significant abnormality. Specifically, there was no leukocytosis or change in the differential count at day 2 when mononuclear infiltrates were observed about the pulmonary veins on histologic examination of the lungs. At day 10, female mice who were treated with either 10 or 100 ␮g DNA had statistically significant reductions in white blood cell counts, mostly accounted for by reduced numbers of lymphocytes (data not shown), though all values were in the normal range and there was no effect of DNA dose. Values that differed between experimental and saline-treated control groups are shown in Table 3. Total serum complement (CH50) was also measured and was within the normal range in all groups at all time points. Pathologic Examination of the Lung In the lung, most animals in all groups had areas of pulmonary hemorrhage attributable to the carbon dioxide narcosis that was used to sacrifice the animals. Otherwise, the lungs of all animals treated with saline or the low dose complexes were normal at every time point

examined (2, 10, and 28 days after dosing). The lungs of animals treated with the high dose complexes (100 ␮g DNA), 2 days after dosing, showed mononuclear infiltration about the pulmonary veins, graded 1–2⫹ (trace to mild) by the pathologist, in all 10 males and 9 of 10 females. Ten days after dosing, all females treated with the high dose complex had normal lungs, but 3 of 10 males showed 1⫹ (trace) mononuclear infiltrates about the pulmonary veins. At day 28, no animals showed any infiltrates. Table 1 summarizes these results. None of the animals had peribronchial or alveolar infiltrates, and none had neutrophils in the infiltrates. Since the muscle in the pulmonary veins is comparable to that of the heart in mice, we carefully examined the hearts for mononuclear cell infiltrates, but found none. Fig. 1 shows representative sections from a saline-treated animal, an animal treated with the low dose complex, and a mouse treated with the high dose complex, the latter at two different time points. Experiments 2 and 3 We divided Balb/C mice into groups such that 10 males and 10 females received saline, 10 males and 10 females received the high dose complex, and 7 males and 7 females received the low dose complex. At an “early” time point, 8 –9 h after dosing, we sacrificed 5 males and 5 females given saline, 2 males and 2 females given the low dose complex, and 5 males and 5 females given the high dose complex. At the “1 day” time point, 24 –28 h after dosing, we sacrificed 5 males and 5 females from all dose groups. We obtained blood by cardiac puncture and performed BAL. We selected these time points based on the clinical studies that show an increase in serum IL-6 in patients dosed with lipid–DNA complexes in the lung by 1–3 h that peaks at about 6 h following administration and persists for at least 24 –30 h [1,2] and on animal studies demonstrating peak cytokine levels in lung at 5– 8 h and in BAL fluid at 24 h following aerosolization of DNA–polyethyleneimine complexes [9]. We obtained cell

TABLE 2: Serum chemistries for which either experimental group differed from saline-treated control (⫾SD) Day—sex 2—male

2—female

Test

Saline

Compacted DNA 10 ␮g


2 ⫾ 2.1

0 ⫾ 0.4

Globulin

2.3 ⫾ 0.6

1.9 ⫾ 0.3

1.8 ⫾ 0.2a

Total protein

5.1 ⫾ 0.25

4.6 ⫾ 0.3

4.7 ⫾ 0.2a

Alkaline phosphatase

131 ⫾ 23

101 ⫾ 11a

100 ⫾ 16a

Phosphate

GGT

a

a

0 ⫾ 0a

10.4 ⫾ 1.2

8.5 ⫾ 0.6

9.3 ⫾ 1

Globulin

1.6 ⫾ 0.2

1.3 ⫾ 0.2a

1.4 ⫾ 0.1a

10—male

ALT

337 ⫾ 116

228 ⫾ 106

193 ⫾ 82a

28—male

Cholesterol

118 ⫾ 4.4

117 ⫾ 9

105 ⫾ 8a

28—female

Triglycerides

150 ⫾ 37

139 ⫾ 40

a

a

94 ⫾ 27a

Significantly different from control (saline-treated animals), P ⬍ 0.05

950


ARTICLE

doi:10.1016/j.ymthe.2003.09.002

TABLE 3: Hematologic parameters that differ from control (⫾SD) Compacted DNA 10 ␮g


RBC (million)

10.7 ⫾ .5

9.75 ⫾ 0.5a

9.69 ⫾ 0.5a

Hematocrit (%)

53.5 ⫾ 2.3

48.5 ⫾ 2.55a

MCH

Day—sex 2—male

Test

Control

48 ⫾ 2.3a

16.4 ⫾ 0.2

17.1 ⫾ 0.5

a

17.4 ⫾ 0.4a

MCHC

32.8 ⫾ 0.3

34.3 ⫾ 0.9a

35 ⫾ 0.8a

MCHC

35.6 ⫾ 0.6

34.7 ⫾ 0.5

10—female

White blood ct (thousand)

11.5 ⫾ 1.75

8.7 ⫾ 1.5

9.0 ⫾ 1.8a

10—female

MCH

18.2 ⫾ 0.3

17.6 ⫾ 0.4

17.5 ⫾ 0.6a

28—male

MCH

16.4 ⫾ 0.2

16.8 ⫾ 0.4a

17.1 ⫾ 0.3a

MCHC

32.8 ⫾ 0.5

33.4 ⫾ 0.7

33.9 ⫾ 0.7a

Platelet ct (thousand)

853 ⫾ 138

890 ⫾ 117

1014 ⫾ 120a

2—female

28—female a

a

a

35.8 ⫾ 0.6

Significantly different from control (saline-treated animals), P ⬍ 0.05

counts and differentials on BAL fluid and measured the following cytokines by ELISA: IL-6, macrophage inhibitory protein (MIP)-2, keratinocyte chemoattractant (KC), TNF-␣, and IL-1␤. The limits of detection for these cytokines were, respectively, 15.6, 7.8, 15.6, 24.3, and 7.8 pg/ml. We performed experiment 3 with methods the same as for the second experiment, but we sacrificed 6 male Balb/C mice from each dose group at time points 4 – 6, 24, 48, and 72 h following dosing. We performed BAL and analyzed the fluid for cell count and differential and cytokine content. In addition, we obtained blood at sacrifice. We obtained cell counts and differentials on BAL fluid and measured the following cytokines: IL-12, IFN-␥, TNF-␣, IL-1␤, IL-4, IL-5, IL-2, IL-10, granulocyte–macrophage colony stimulating factor (GM-CSF), IL-6, MIP-2, and KC. The limits of detection for these cytokines were, respectively, 80, 3, 16, 3, 3, 3, 3, 3, 16, 3, 15.6, and 7.8 pg/ml. Because the methodologies for experiments 2 and 3 were similar, the data are presented together. BAL Cytokine Response to Compacted DNA or Saline For experiments 2 and 3, many values for BAL cytokines were below the limits of detection of the assays. For IL-1␤, no values were above the limits of detection of the assay in either experiment. We measured IL-12, interferon-␥, IL-4, IL-5, IL-2, IL-10, and GM-CSF only in experiment 2 and they did not exceed the limits of detection in any sample. For TNF-␣, only three values in experiment 1, and none in experiment 2, were above the limits of detection: one in the saline group, one in the low dose group, and one in the high dose group at the early time point. These values are not shown. For MIP-2, values were above the limits of detection at 4 – 6 h, but were the same in all dosing groups (saline, low dose, and high dose). After this time point, only scattered samples had values that exceeded the limits of detection. Figs. 2A and 2B illustrate the IL-6 and KC responses. For both of these cytokines,


there were highly variable responses at the 4- to 6-h time point in all treatment groups, including saline alone. At 24 h, KC levels were comparable in the three groups, but IL-6 was significantly elevated in the high dose group. At 48 and 72 h, the mice treated with high dose complex had significantly elevated levels of both KC and IL-6. Serum MIP-2 and IL-6 Only the occasional value (6 of 54, distributed in the saline, low dose, and high dose groups) for MIP-2 was above the limits of detection (data not shown). For IL-6, most values were above the limits of detection in all three groups from both experiments, but there were no significant differences among the groups at any time point (Fig. 2C). BAL Cell and Differential Counts Fig. 3A shows the cell counts for animals included in experiments 2 and 3. There were no significant differences across the time course. Fig. 3B shows the percentage of neutrophils in the cells recovered by BAL. Values were comparable for all three groups at the 4- and 24-h time points, but the high dose group had significantly increased percentage neutrophils at 48 and 72 h after dosing. Experiments 4, 5, and 6 In experiment 4, we compared directly instillation of saline with instillation of 10 ␮g naked DNA, 100 ␮g naked DNA, 10 ␮g compacted DNA, and 100 ␮g compacted DNA in male Balb/C mice. We sacrificed six mice per group at 24 h and obtained BAL fluid and blood. In BAL fluid, we measured cell counts and cytokine content. We performed experiment 5 on six male Balb/C mice per group that were given saline, 10 ␮g single-stranded E. coli genomic DNA (Sigma Chemical Co., St. Louis, MO, prepared as described [5]), 100 ␮g compacted DNA, or 100 ␮g naked DNA. We sacrificed the animals 4 h later and obtained BAL cell counts and cytokine levels. For experi-

951

ARTICLE

doi:10.1016/j.ymthe.2003.09.002

FIG. 1. Sample lung histology from Balb/C mice dosed with 100 ␮g compacted DNA 48 h postdosing (A, airway; B, pulmonary blood vessel), 100 ␮g compacted DNA 10 days postdosing (C, airway; D, pulmonary blood vessel), 10 ␮g compacted DNA 48 h postdosing (E, airway; F, pulmonary blood vessel), or saline 48 h postdosing (G, airway; H, pulmonary blood vessel). All images were photographed at 20⫻ magnification.

ment 6, we used male mice of the C3H/HeBFEJ genotype. This genotype is insensitive to endotoxin and so isolates the inflammatory response to CpG sequences, whereas the Balb/C mice could respond to the endotoxin as well as the CpG nucleotides. These mice received saline, 10 ␮g

952

single-stranded E. coli genomic DNA (Sigma Chemical Co., St. Louis, MO, prepared as described [5]), 100 ␮g compacted DNA, 100 ␮g naked DNA, or 5 ␮g naked DNA complexed with the lipofectin reagent. We sacrificed the mice 4 h after dosing and obtained BAL fluid and blood.


ARTICLE

doi:10.1016/j.ymthe.2003.09.002

Comparison of Naked DNA, Compacted DNA, Genomic DNA, and Lipid-Complexed DNA We compared compacted DNA to the same plasmid administered as naked DNA. There were no significant differences between the inflammatory responses to the compacted DNA and to the naked plasmid DNA at either the 4- or the 24-h time points, times at which strong CpG responses have been observed in other systems [3–5,8]. To verify that the Balb/C mice we studied were capable of generating an inflammatory CpG response, we administered 10 ␮g E. coli DNA prepared as described by Schwartz et al. [8], as well as 100 ␮g compacted DNA and 100 ␮g plasmid DNA and 5 ␮g naked plasmid DNA complexed with the lipofectin reagent. Bacterial genomic DNA provoked a significant BAL neutrophilia and elevated MIP-2 and KC at 4 and 24 h after dosing compared to mice given saline, or 100 ␮g naked or compacted DNA, or 5 ␮g lipofectin-complexed DNA, indicating that these mice can mount an inflammatory response to instilled DNA (data not shown). However, we did not detect TNF-␣, IFN-␥, or IL-12 in the BAL fluid of these mice, even those dosed with E. coli genomic DNA. To verify that the modest response to compacted DNA was not unique to Balb/C mice, and to isolate the response to DNA from that induced by the endotoxin that contaminated the E. coli genomic DNA preparations, we studied C3H/HeBFEJ mice, which are endotoxin resistant. The pattern of KC and MIP-2 responses in these animals was similar to that of the Balb/C mice (data not shown), but IFN-␥, IL-1␤, and IL-12 were below the limits of detection in BAL fluid from these C3H/HeBFEJ mice. Since the C3H/HeBFEJ mice have minimal response to endotoxin, these cytokine elevations were probably entrained by the DNA.

DISCUSSION Balb/c mice treated intranasally with 10 ␮g DNA compacted with CK30 coupled to PEG are indistinguishable from mice treated with saline in all the parameters we measured, including lung histology, BAL indicators of inflammation (cell counts and cytokines), and systemic indicators of inflammation (cell counts, IL-6). Yet this dose of DNA produces gene expression more than 200fold greater than a comparable dose of naked DNA and greater than the maximum gene expression from any dose of naked DNA and comparable to expression from PEI-complexed DNA [10], which reportedly, at even lower doses, produces much larger inflammatory responses [9]. When the dose is increased to 100 ␮g compacted DNA, the dose that gives maximal gene expression, which is 200-fold greater than a comparable dose of naked DNA FIG. 2. Cytokine levels in BAL fluid and serum following treatment of Balb/C mice with saline, low dose compacted DNA (10 ␮g), or high dose compacted DNA (100 ␮g). Mice were sacrificed at the indicated times after dosing, and BAL fluid and blood were retrieved and assayed by ELISA. Concentrations in BAL fluid are adjusted for dilution by the urea method. (A) KC levels in BAL. (B) IL-6 levels in BAL. (C) IL-6 levels in serum. *Group differs from the low dose


group, by ANOVA, P ⬍ 0.001. **Group differs from the saline and low dose groups, by ANOVA, P ⬍ 0.001.

953

ARTICLE

doi:10.1016/j.ymthe.2003.09.002

FIG. 3. Cell counts and differential counts of neutrophils in BAL fluid from Balb/C mice dosed with saline, low dose compacted DNA (10 ␮g), or high dose compacted DNA (100 ␮g). Mice were sacrificed at the indicated times after dosing, BAL fluid was retrieved, and cell counts were performed. (A) Cell counts. (B) Percentage of neutrophils. *Significant differences from the low dose group by ANOVA or t test, P ⬍ 0.05. **Significant differences from the saline and low dose groups by ANOVA or t test, P ⬍ 0.05.

[10], treated mice display a mild, reversible, dose-related inflammatory response to these complexes in the lung. Mononuclear cells were observed, in trace to small amounts, around the pulmonary veins at 48 h in 19 of 20 mice that received 100 ␮g compacted DNA; by day 10, only 3 of 20 animals still had collections of mononuclear cells about the veins, and there were no abnormalities at day 28. There were no peribronchial or alveolar infiltrates on histologic examination at 48 h, and there was no accompanying systemic response, i.e., no serum leukocytosis or shift in the differential counts, no increase in serum IL-6, and no change in feeding or grooming behavior. Animals treated with 100 ␮g compacted DNA had increased neutrophils in BAL fluid at 48 and 72 h, as well as elevations in the chemokine KC and the regulatory cytokine IL-6. Since KC is a known neutrophil chemoattractant, elevation of KC is consistent with the influx of neutrophils into the BAL fluid. There was no elevation in BAL fluid of IL-1 ␤, IFN-␥, IL-12, or TNF-␣, cytokines that have been reported to be elevated in association with intrapulmonary administration of DNA complexed with other reagents [3–9]. In addition, GM-CSF, IL-2, IL-4, IL-5, and IL-10 did not increase at any time point up to 72 h. There were none of the systemic effects (such as leukopenia, thrombocytopenia, elevation of IL-6) that can be seen with systemic administration of lipid–DNA complexes [14]. In sum, the inflammatory response to compacted plasmid DNA was mild and transient, only evident for the highest dose of compacted DNA (100 ␮g), and qualitatively and temporally different from the responses reported previously for nonviral complexes that are capable of delivering DNA effectively (more than 100-fold above naked DNA) to airway epithelium [3–9]. The mechanism by which this inflammatory response

954

occurs is not entirely clear. Its pattern, nature, and timing differ substantially from earlier reports of the inflammatory response to unmethylated CpG sequences contained in the plasmid DNA prepared in bacteria, which occurs during the first 24 h, is accompanied by elevations in serum IL-6 and BAL TNF-␣ and IL-12, and is accompanied by marked neutrophilia [4,6 – 8,14,15]. Despite the large dose of DNA (100 ␮g per mouse) relative to those in other reports of CpG responses to lipid–DNA complexes (⬃20 – 60 ␮g DNA) or PEI–DNA complexes (⬃5 ␮g/ mouse), little response characteristic of the CpG response was seen, and none in excess of naked DNA. Our assays were capable of detecting 10 –20% of the cytokine levels reported by others for PEI-complexed DNA [9] and so would have been able to detect such a response had it occurred. The plasmids used have few of the most intensely stimulatory CpG sequences, however— only 21 RRCGYY sequences are found in our CFTR plasmid, and 23 in luciferase—and may simply not provide an intense stimulus. It is also possible that compacted DNA does not enter the critical responding cells for the CpG response well (plasmacytoid dendritic cells, B cells, monocytes, and natural killer cells) [3,15–17,22,24] or that the ability of compacted DNA to access the nucleus rapidly [11] provides less opportunity to interact with toll-like receptor 9, which is located in endosomes [15,18,20]. In experiments aimed at assessing gene transfer rather than toxicity, in which mice were maintained in microisolator cages rather than the hanging cages used in the experiments reported here, only two of eight C57BL/6 mice dosed by intratracheal inoculation with 100 ␮g compacted DNA (containing the lacZ reporter gene) had perivascular mononuclear infiltrates on day 2 [10]. For these two mice, reporter gene expression was also observed in the adjacent endothe-


ARTICLE

doi:10.1016/j.ymthe.2003.09.002

lium, so the inflammatory response may be associated with gene expression in the blood vessels. None of the eight mice in the gene transfer experiment had histologic evidence of bronchial or alveolar inflammation. It seems likely that exposure of the animals to ambient rather than sterile conditions, with the possibility for epithelial irritation and associated breaches of epithelial integrity, accounts for the increased incidence of perivascular infiltrates in the current experiments, though strain differences and dosing routes cannot be excluded as contributors to the differences we observed. However, if clinical applications are to be considered for these reagents, studies under ambient, rather than sterile, conditions are probably more relevant. The reason for the delayed response of neutrophils, KC, and IL-6 to DNA compacted with PEG-CK30 is not clear. One possibility is that the complex itself is benign, but the foreign gene contained in the complex (human CFTR or firefly luciferase), once it is expressed, provokes some response. Another possibility is that the complex degrades, but slowly, and at 48 h sufficient uncomplexed DNA has accumulated to stimulate a small CpG-type response, or free poly K has been released and, despite the protection of the covalent linkage to PEG, is now available to produce local activation of complement, which provokes a modest inflammatory response [12,13]. These possibilities can be sorted out by following the fate of the complex components in vivo by molecular imaging techniques and by experiments designed to separate effects of gene expression from the presence of the compacted DNA complex. Serum complement was not reduced in any of the mice following treatment with any dose of compacted DNA. This point is noteworthy since free poly K causes complement activation as well as other toxicities, including bronchospasm [12,13]. The centrifugal elutriation step during complex preparation ensures that free CK30-PEG has been removed from the DNA complexes [11]. Moreover, the stability of these complexes after incubation in mouse serum at 37°C also suggests that little, if any, poly K is released from the complexes prior to cellular entry [10,11]. The relatively benign response to these complexes, even at very high doses of DNA, suggests that their utility as gene transfer agents, at least to the airway epithelium, may not be limited by the inflammatory response. On the other hand, any adjuvant effect of the inflammatory response will also be absent, so gene transfer for vaccination purposes or cancer killing using CK30-PEG–DNA complexes may be less effective than other means of gene delivery [26]. In addition, although these complexes appear to be quite benign in the pristine lungs of experimental animals, their impact in a lung that is infected and inflamed remains to be determined. This question is important since such a situation occurs in one important target disease for pulmonary gene transfer, cystic fibrosis.


Animal models exist that permit this question to be addressed [27].

METHODS Animals. Balb/C mice 7–10 weeks of age were used in most experiments. This strain was selected because in preliminary experiments, they tolerated the anesthesia regimen better than other strains tested. Since we wanted to assess toxicity, we wished to minimize the complications of interpretation of data that could be introduced by anesthetic deaths. For one experiment, C3H/HeBFEJ mice were studied because they are endotoxin resistant, and we wished to examine the effects of E. coli genomic DNA in our system. Animals were acclimated to the facility for at least 1 week after shipping prior to use in experiments. They were maintained in hanging cages and fed normal mouse chow and had unlimited access to food and water throughout the experiment. Preparation of condensing peptides. CK30-PEG10k was prepared by mixing CK30 (trifluoroacetate salt at 20 –50 mg/ml in 15 ml of 0.1 M phosphate-buffered saline, pH 7.2, with 5 mM EDTA) with 10% molar excess of 10,000 MW methoxy-PEG-maleimide (in 15 ml of dimethyl sulfoxide) at room temperature (RT) overnight. The methoxy-PEG-maleimide was added drop-wise over ⬃5 min to the CK30 solution mixing on a Vibrax shaker. The reaction mixture was then fractionated on a Sephadex G-25 column (HiPrep 26/10 desalting, Amersham Pharmacia Biotech) equilibrated with 50 mM ammonium acetate and fractions containing peptides based on absorbance at 220 nm were then pooled and lyophilized. Formulation of compacted DNA nanoparticles. Nine volumes of DNA at a concentration of 0.222 mg/ml in distilled water was added at a rate of 4 ml/min to 1 volume of CK30-PEG10k (acetate counterion) (6.4 mg/ml) vortexing in water at room temperature, for a final DNA concentration of 0.2 mg/ml and end-point ratio of positive to negative charges (NH3⫹/PO4⫺) of 2:1. The compacted DNA was sterile filtered and concentrated to ⬃4 mg/ml using centrifugal filter devices Ultrafree-4 or Ultrafree-15 with Biomax-100K (or 50K) membrane (Millipore). The concentrated DNA was diluted back to the original volume with 0.9% NaCl and the concentration step was repeated. During this process, excess free CK30-PEG10k was removed and water was exchanged with saline. Concentrated compacted DNA was stored at 4°C. Compacted DNA used in this study met or exceeded a series of qualification parameters, including size and shape characteristics as determined from transmission electron micrographs (EMs) and other attributes determined from light scattering, gel, serum stability, and salt stability analyses. EM analysis demonstrated rod-like DNA nanoparticles with a radius of ⬃12–15 nm and a length of 100 –150 nm (data not shown). Gel analysis showed no free or degraded DNA, and these complexes demonstrated essentially no DNA degradation after incubation in 75% mouse serum at 37°C for 2 h. No aggregation was seen on EM. Colloidal stability of CK30-PEG10k-compacted DNA in physiologic saline was assayed by sedimentation at 3400g for 1 min at RT: no less than 90% of the DNA remained in the supernatant. The plasmid backbone contains the kanamycin resistance gene and the cytomegalovirus immediate early promoter. For experiments 1–3, the transgene was CFTR, for experiments 4 – 6, the plasmid encoded luciferase. Endotoxin assay. DNA endotoxin levels were measured using the KineticQCL Limulus Amebocyte Lysate (LAL) Assay (BioWhittaker, Walkersville, MD, Cat. No. 50-650U). All reagents used in the assay were pyrogen-free. The assay was performed according to the test kit manual with the following modifications: all naked DNA samples and compacted DNA from experiment 1 were diluted in LAL water, whereas the remaining compacted DNA samples were diluted in 50 mM Tris buffer (BioWhittaker Cat. No. S50-642) containing 0.5% (v/v) Pyrosperse reagent (BioWhittaker Cat. No. N188) to minimize false positive results due to the polycationic compacting polymer. Positive controls in each assay included diluted test samples that were spiked with endotoxin (0.5 EU/ml). In these controls, endotoxin activity was not diminished by addition of Pyrosperse, al-

955

ARTICLE

though Pyrosperse blocked false-positive signal generated from CK30-P10k. Plasmid DNA samples were diluted from 1:80 to 1:400 because these dilutions provided the best positive control recovery, while maintaining a low limit of detection of the assay. The bacterial E. coli genomic DNA that was used to provide a positive control in the cytokine study was diluted 1:106 to allow the endotoxin levels to fall within the standard curve. After samples were aliquoted, the plate was preincubated for 10 min at 37°C before the colorimetric substrate was added. The rate of reaction was measured at 405 nm and 37°C. Endotoxin content was calculated from the kinetic data using software provided for this purpose with the instrument. The endotoxin levels of both naked and compacted plasmid DNA used in these studies were 9.3 EU/mg, or 40.9 EU/dose/kg, assuming 20-g mice. The bacterial genomic DNA endotoxin levels were on the order of 103–106 EU/mg or 2 ⫻ 105 EU/dose/kg, assuming 20-g mice. Bronchoalveolar lavage. BAL was performed as previously described [19], using three 1-ml aliquots of warmed Hanks’ buffered salt solution, which were pooled for each animal. Lavage fluid return was about 2.5 ml. The pooled lavage fluid was mixed and cell counts were performed in a hemocytometer. An aliquot was taken for cytospin and stained with Wright’s stain for differential counts. At least 100 cells were counted for the differential. The remainder of the fluid was centrifuged, aliquoted, and immediately stored at ⫺80°C until assayed for cytokines. Cytokine assays. Cytokines were assayed by the CF Core Center Mediator Assay Core by ELISA as previously described [19] or in a Luminex instrument. All assays of MIP-2 and KC were performed separately by ELISA, because we discovered significant cross reactivity of the antibodies to MIP-2 with KC when the Luminex assay was used. Comparison of the same samples in Luminex assay and ELISA showed excellent (r ⬎ 0.95) rank correlation between the assays (that is, the highest samples by Luminex are also the highest by ELISA and the lowest, also the lowest in ELISA) but the absolute values differed slightly from one another. The dilution of the BAL fluid was estimated by the urea dilution method as previously described [19].

ACKNOWLEDGMENTS The authors thank Alma Wilson for her expertise with animal maintenance and Christopher vanHeeckren and James Poleman for their expertise with BAL fluid preparation and analysis. This work was supported by Copernicus Therapeutics, Inc., and by grants from the National Institutes of Health (DK27651, HL07415, and DK58318) and the Cystic Fibrosis Foundation (RDP grant and fellowship to A.G.Z.). Drs. Ziady and Davis hold equity in Copernicus Therapeutics, Inc. RECEIVED FOR PUBLICATION JULY 23, 2003; ACCEPTED SEPTEMBER 6, 2003.

REFERENCES 1. Alton, E. W. F. W., et al. (1999). Cationic lipid-mediated CFTR gene transfer to the lungs and nose of patients with cystic fibrosis: a double-blind placebo-controlled trial. Lancet 353: 947–954. 2. Ruiz, F. E., et al. (2001). A clinical inflammatory syndrome attributable to aerosolized lipid–DNA administration in cystic fibrosis. Hum. Gene Ther. 12: 751–761. 3. Yew, N. S., et al. (1999). Contribution of plasmid DNA to inflammation in the lung after administration of cationic lipid: pDNA complexes. Hum. Gene Ther. 10: 223–234. 4. McLachlan, G., Stevenson, B. J., Davidson, D. J., and Porteous, D. J. (2000). Bacterial DNA is implicated in the inflammatory response to delivery of DNA/DOTAP to mouse lungs. Gene Ther. 7: 384 –392.

956

doi:10.1016/j.ymthe.2003.09.002

5. Freimark, B. D., et al. (1998). Cationic lipids enhance cytokine and cell influx levels in the lung following administration of plasmid: cationic lipid complexes. J. Immunol. 160: 4580 – 4586. 6. Yew, N. S., et al. (2000). Reduced inflammatory response to plasmid DNA vectors by elimination and inhibition of immunostimulatory CpG motifs. Mol. Ther. 1: 255–262. 7. Yew, N. S., et al. (2002). CpG-depleted plasmid DNA vectors with enhanced safety and long-term gene expression in vivo. Mol. Ther. 5: 731–738. 8. Schwartz, D. A., Quinn, T. J., Thorne, P. S., Sayeed, S., Yi, A. K., and Krieg, A. M. (1997). CpG motifs in bacterial DNA cause inflammation in the lower respiratory tract. J. Clin. Invest. 100: 68 –73. 9. Gautam, A., Densmore, C. L., and Waldrep, J. C. (2001). Pulmonary cytokine responses associated with PEI–DNA aerosol gene therapy. Gene Ther. 8: 254 –257. 10. Ziady, A. G. et al. (2003). Transfection of airway epithelium by stable PEGylated poly-l-lysine DNA nanoparticles in vivo. Mol. Ther. 8: 936 –947. 11. Liu, G., et al. (2003). Nanoparticles of compacted DNA transfect post-mitotic cells. J. Biol. Chem. 278: 32578 –32586. 12. Plank, C., Mechtler, K., Szoka, F., and Wagner, E. (1996). Activation of the complement system by synthetic DNA complexes: a potential barrier for intravenous gene delivery. Hum. Gene Ther. 7: 1437–1446. 13. Coyle, A. J., Ackerman, S. J., and Irvin, C. G. (1993). Cationic proteins induce airway hyperresponsiveness dependent on charge interactions. Am. Rev. Respir. Dis. 147: 896 –900. 14. Tousignant, J. D., Zhao, H., Yew, N. S., Cheng, S. H., Eastman, S. J., and Scheule, R. K. (2003). DNA sequences in cationic lipid: pDNA-mediated systemic toxicities. Hum. Gene Ther. 14: 203–214. 15. Krieg, A. M. (2002). CpG motifs in bacterial DNA and their immune effects. Annu. Rev.. Immunol.. 20: 709 –760. 16. Gursel, M., Verthelyi, D., Gursel, I., Ishii, K. J., and Klinman, D. M. (2002). Differential and competitive activation of human immune cells by distinct classes of CpG oligodeoxynucleotide. J. Leukocyte Biol. 71: 813– 820. 17. Hartmann, G., and Krieg, A. M. (2000). Mechanism and function of a newly identified CpG DNA motif in human primary B cells. J. Immunol. 164: 944 –953. 18. Kandimalla, E. R., Zhu, F. G., Bhagat, L., Yu, D., and Agrawal, S. (2003). Toll-like receptor 9: modulation of recognition and cytokine induction by novel synthetic CpG DNAs. Biochem. Soc. Trans. 31(Pt): 3 ): 654 – 658.. 19. Chu, R. S., Targoni, O. S., Krieg, A. M., Lehmann, P. V., and Harding, C. V. (1997). CpG oligodeoxynucleotides act as adjuvants that switch on T helper 1 (Th1) immunity. J. Exp. Med. 186: 1623–1631. 20. Ahmad-Nejad, P., Hacker, H., Rutz, M., Bauer, S., Vabulas, R. M., and Wagner, H. (2002). Bacterial CpG–DNA and lipopolysaccharides activate Toll-like receptors at distinct cellular compartments. Eur. J. Immunol. 32: 1958 –1968. 21. Zelenay, S., Elias, F., and Flo, J. (2003). Immunostimulatory effects of plasmid DNA and synthetic oligodeoxynucleotides. Eur. J. Immunol. 33: 1382–1392. 22. Krug, A., et al. (2001). Toll-like receptor expression reveals CpG DNA as a unique microbial stimulus for plasmacytoid dendritic cells which synergizes with CD40 ligand to induce high amounts of IL-12. Eur. J. Immunol. 31: 3026 –3037. 23. Huang, L., Krieg, A. M., Eller, N., and Scott, D. E. (1999). Induction and regulation of Th1-inducing cytokines by bacteria, DNA, lipopolysaccharide, and heat-inactivated bacteria. Infect. Immun. 67: 6257– 6263. 24. Hartmann, G., and Krieg, A. M. (1999). CpG DNA and LPS induce distinct patterns of activation in human monocytes. Gene Ther. 6: 893–903. 25. Yasuda, K., Ogawa, Y., Kishimoto, M., Takagi, T., Hashida, M., and Takakura, Y. (2002). Plasmid DNA activates murine macrophages to induce inflammatory cytokines in a CpG motif-independent manner by complex formation with cationic liposomes. Biochem. Biophys. Res. Commun. 293: 344 –348. 26. Liu, Y., et al. (2002). , Adenovirus-mediated CD40 ligand gene-engineered dendritic cells elicit enhanced CD8(⫹) cytotoxic T-cell activation and antitumor immunity. Cancer Gene Ther. 9: 202–208. 27. van Heeckeren, A., Walenga, R., Konstan, M. W., Bonfield, T., Davis, P. B., and Ferkol, T. (1997). Excessive inflammatory response of cystic fibrosis mice to bronchopulmonary infection withPseudomonas aeruginosa. J. Clin. Invest. 100: 2810 –2815.
