Intramuscular injection of DNA vaccines elicits potent humoral and cellular immune responses in mice. However, DNA vaccines are less efficient in larger animal models and humans. To gain a better understanding of the factors limiting the efficacy of DNA vaccines, we used fluorescence-labeled plasmid DNA in mice to 1) define the macroscopic and microscopic distribution of DNA after injection into the tibialis anterior muscle, 2) characterize cellular uptake and expression of DNA in muscle and draining lymph nodes, and 3) determine the effect of modifying DNA distribution and cellular uptake by volume changes or electroporation on the magnitude of the immune response. Injection of a standard 50-μl dose resulted in the rapid dispersion of labeled DNA throughout the muscle. DNA was internalized within 5 min by muscle cells near the injection site and over several hours by cells that were located along muscle fibers and in the draining lymph nodes. Histochemical staining and analysis of mRNA expression in isolated cells by RT-PCR showed that the transgene was detectably expressed only by muscle cells, despite substantial DNA uptake by non-muscle cells. Reduction of the injection volume to 5 μl resulted in substantially less uptake and expression of DNA by muscle cells, and correspondingly lower immune responses against the transgene product. However, expression and immunogenicity were restored when the 5-μl injection was followed by electroporation in vivo. These findings indicate that distribution and cellular uptake significantly affect the immunogenicity of DNA vaccines.
Immunization with DNA vaccines results in humoral and cellular immune responses that protect against disease in preclinical models of infectious diseases, cancer, and autoimmunity (for review, see Ref. 1). In mice, DNA vaccines readily induce strong cell-mediated immunity, although Ab responses are typically weaker than that of corresponding protein-based vaccines administered with adjuvant (2). In larger animals, multiple immunizations at high DNA doses often achieve only modest immune responses (3). A few recent clinical trials have demonstrated the immunogenicity of DNA vaccines in humans, although their potency is limited for reasons that are still unclear (4, 5, 6). These limitations must be understood before DNA vaccines can be developed to their full potential.
The mechanisms of action of DNA vaccines have been mainly investigated in mice using either the intradermal or i.m. routes of immunization. Keratinocytes are the main site of transgene expression after intradermal administration by particle bombardment (7), whereas muscle cells principally express transgene administered i.m. by needle injection. However, bone marrow-derived dendritic cells are central to the induction of immune responses by DNA vaccines utilizing either immunization route (8, 9, 10, 11, 12, 13). Immune responses are promoted by transgene product expressed by transfected dendritic cells (direct priming) or by nonlymphoid cells (cross-priming).
The factors limiting DNA vaccine immunogenicity are not fully identified. In addition, the relative contributions of nonlymphoid cells and dendritic cells to the initiation and magnitude of immune responses remain to be fully determined. Hence, the goals of the present study were to 1) determine the macroscopic and microscopic distribution of DNA after i.m. injection, 2) characterize cellular uptake and expression of DNA in muscle and draining lymph nodes, and 3) determine the effect of modifying DNA distribution on the magnitude of the immune response.
To address these questions, we followed the distribution of functional fluorescence-labeled DNA after i.m. injection into BALB/c mice. The distribution and cellular uptake of labeled DNA were examined by fluorescence and confocal microscopy in the tibialis anterior muscle and draining lymph nodes within 5 min, and at 24 h after injection. Mononuclear cells (MNC)3 that internalized labeled DNA were characterized by flow cytometry after isolation by enzymatic digestion of tissue followed by cell separation on a density gradient. Transgene expression was followed by RT-PCR and by histochemical staining. Finally, we examined the effects of modifying DNA vaccine distribution on Ag expression and immunogenicity by reducing the injection volume and electroporation. We found that distribution and cellular uptake present significant limitations to DNA vaccine potency. Therefore, techniques that enhance DNA delivery should prove to be useful in developing improved DNA vaccines.
Materials and Methods
Pathogen-free BALB/c mice 6–8 wk of age were obtained from Charles River Laboratories (Hollister, CA). All experimental procedures were approved by the Committee on Animal Research of the University of California, San Francisco, and the Animal Care and Use Committee at Chiron Corporation.
The distribution of DNA in vivo was followed using plasmid DNA that was directly and irreversibly labeled with a fluorescence-labeled peptide nucleic acid “clamp” (Gene Therapy Systems, San Diego, CA). The function and conformation of labeled plasmid prepared in this way are not altered (14). Rhodamine- or FITC-labeled DNA contained the CMV immediate early gene promoter and intron A at the 5′ end of a cDNA encoding β-galactosidase. Plasmid pCMVlux contained the CMV promoter and a cDNA encoding the gene for firefly luciferase (15). Plasmid CMVgagmodSF2 contained the CMV promoter and a cDNA encoding the gag gene of the HIV SF2 strain (16).
A vaccine dose contained 10 μg rhodamine-labeled DNA encoding β-galactosidase (Gene Therapy Systems) or 10 μg unlabeled HIV gag DNA (previously demonstrated to be an effective dose (17)) dissolved in PBS, pH 7.4. Anesthetized mice (87 mg/kg ketamine and 13 mg/kg xylazine i.p.) were immunized with a single vaccine dose injected at right angle to the skin into the central portion of the right tibialis anterior muscle using a 0.5-ml tuberculin syringe or a 0.025-ml Hamilton syringe with a 28-gauge needle. To control the depth of needle penetration, the needle was covered with polyethylene tubing (internal diameter, 0.38 mm) to expose only 2 mm of the bevel. The rate of injection was 10 μl/s. The distribution of the label alone (control) was assessed by injecting a 50-μl dose of rhodamine-labeled peptide nucleic acid dissolved at 1 pmol/μl in PBS (Perkin-Elmer, Norwalk, CT).
Electroporation in vivo
Details of electroporation are given elsewhere (17). Briefly, the skin overlying the tibialis anterior muscle of anesthetized mice was shaved, and a single dose of plasmid DNA was injected. For electroporation in vivo, a two-needle array electrode pair (Genetronics, San Diego, CA) was inserted into the muscle immediately after DNA delivery. The distance between electrodes was 5 mm, and the array was inserted parallel to the muscle fibers. Six electric pulses of 100 V were delivered at 1-s intervals using a BTX 820 square wave generator.
Processing and immunohistochemical staining of tissues
Groups of four mice were used for fluorescence and confocal microscopic studies. The distribution of DNA was determined in situ by injecting labeled DNA into the tibialis anterior muscle under a MZ FLIII Leica fluorescence stereomicroscope (Leica Microscopy Systems, Heerbrugg, Switzerland) in anesthetized animals. For tissue sections, anesthetized mice (50 mg/kg sodium pentobarbital i.p., Abbott, North Chicago, IL) were fixed by vascular perfusion with 1% paraformaldehyde in PBS, pH 7.4, at 120 mm Hg immediately or at 24 h after injection of DNA. The right tibialis anterior muscle and the draining popliteal lymph node were then removed and fixed further in paraformaldehyde at 4°C overnight. Tissue samples were cut into 150-μm sections with a Vibratome (Technical Products International, St. Louis, MO) and mounted on Superfrost slides (Fisher Scientific, Pittsburgh, PA).
For immunohistochemical staining, sections were incubated with or without primary Ab overnight at 37°C. The CD11b Ag expressed on cells of myeloid origin was identified with a rat mAb M1/70 (1:100; PharMingen, San Diego, CA). Sections were then incubated with goat anti-mouse IgG conjugated with FITC (PharMingen). Slides were mounted in Vectashield (Vector, Burlingame, CA) and examined within 24 h.
Fluorescence and confocal microscopy
Slides were examined using a Zeiss Axiophot fluorescence microscope with rhodamine and FITC bandpass filters (Chroma, Brattleboro, VT) and Fluar objectives, or a Zeiss LSM 410 confocal microscope equipped with a krypton-argon laser and optimized photomultiplier tubes. Images were recorded on Kodak Ektachrome film (ASA 400, Eastman Kodak, Rochester, NY) or as digital confocal image files. Images were printed on a digital printer (Fujix Pictography 3000, Fuji Film, Tokyo, Japan) from Adobe Photoshop (Adobe Systems, Mountain View, CA).
Mononuclear cells were isolated from the epimycium of the tibialis anterior muscle or the draining popliteal lymph nodes by enzymatic digestion with 1 mg/ml collagenase D (Boehringer Mannheim, Mannheim, Germany) in 1 ml RPMI containing 2% FCS for 30 min at 37°C. The enzymatic digestion was stopped by adding 10 volumes of HBSS without calcium and magnesium containing 10 mM EDTA. After homogenization in a Dounce homogenizer, the tissue digest was spun for 5 min at 1000 rpm, resuspended in 2 ml HBSS, and layered onto a metrizamide gradient (1.077 g/cm3). MNC were isolated by collecting cells in the light density fraction after centrifugation at 900 × g for 20 min.
Cells immunoreactive for CD11b were isolated by magnetic cell sorting with the Vario-MACS (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. Mononuclear cells isolated from the epimycium or draining popliteal lymph nodes were incubated with magnetic beads coupled to antiCD11b, and cells were positively selected after passage through a MS+ ferromagnetic column (Miltenyi Biotec).
The pattern of transgene expression was determined by histochemical staining of β-galactosidase activity 24 h after i.m. injection of a lacZ reporter gene. β-Galactosidase activity was measured in groups of four mice as described previously (17). Briefly, Vibratome sections were incubated at 37°C for 18 h in a reaction mixture containing 100 μg/ml 5-bromo-4-chloro-3-indolyl-6-d-galactose (X-gal, Life Technologies, Gaithersburg, MD), 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, 2 mM magnesium chloride, 0.01% sodium deoxycholate, 0.02% Nonidet P-40 in PBS, pH 7.4. The X-gal was dissolved in dimethylformamide at 10 mg/ml and then diluted into the reaction mixture. After incubation, sections were washed three times for 10 min with 3% DMSO in PBS and mounted on glass slides with Vectashield for microscopic analysis.
The amount of transgene expression was quantified in groups of 10 mice by measuring luciferase activity after i.m. injection of a lux reporter gene. At 14 days after injection, the tibialis anterior muscle (30–50 mg wet weight) was removed, frozen, and stored at −70°C. Muscle was pulverized with a mortar and pestle on dry ice until a fine powder was obtained. Muscle powder was extracted in lysis buffer and assayed for luciferase activity following the manufacturer’s instructions (Promega, Madison, WI). Uninjected control muscle extracts were prepared in the same manner.
cDNA generation, amplification, and analysis
Total RNA was isolated from the tibialis anterior muscle or isolated MNC purified from muscle or draining lymph nodes according to the Ambion RNA isolation kit manual (Ambion, Woodlands, TX). Contaminant DNA was eliminated by enzymatic digestion with 50 U RNase-free DNase I (Roche, Indianapolis, IN) for 15 min at 37°C, after which the enzyme was inactivated for 10 min at 75°C. cDNA was generated from the RNA template using a 15-mer poly(dT) oligonucleotide and avian myeloblastosis virus RT according to the Promega RT-PCR System protocol (Promega). The cDNA was then amplified by PCR using primers specific for the gag region of CMVgagmodSF2 yielding a 294-bp product. As a control, the cDNA was amplified using primers specific for β-actin yielding a 514-bp product. PCR products were analyzed by agarose gel electrophoresis in the presence of 1 μg/ml ethidium bromide. The sensitivity of the method was determined by spiking uninjected muscle extracts with dilutions of total RNA from 293 cells transfected with CMVgagmodifSF2, followed by RNA recovery and PCR amplification. The RT-PCR was able to amplify a 294-bp product specific for HIV gag from total muscle RNA spiked with the equivalent of 100 transfected 293 cells.
Serum Ab titers
Groups of 10 mice were immunized, and 6 wk later blood samples were drawn from the orbital sinus using a heparinized capillary tube (Oxford Labware, St. Louis, MO). Blood was clotted and spun down, and serum was frozen at −20°C until thawed for analysis by ELISA. For Ag-specific total Ig ELISA, Immulon 2 plates (Dynatech Laboratories, Chantilly, VA) were coated with 5 μg/ml recombinant HIV gag Ag at 37°C for 1 h. Plates were washed and blocked with PBS containing 0.3% Tween 20 and 1% goat serum for 1 h at 37°C. Diluted serum samples were then added and incubated at room temperature for 1 h. After being washed with PBS containing 0.3% Tween 20 and 1% goat serum, 100 μl goat anti-mouse Ig IgM + IgG conjugated to HRP (1:6000; Boehringer Mannheim, Indianapolis, IN) antiserum was added, and plates were incubated at 37°C for 1 h. Finally, plates were washed six times with PBS and developed with orthophenylamine diamine at room temperature. After 5 min, the reaction was stopped with H2SO4, and the OD492 nm was determined. The titer was defined as the reciprocal serum dilution giving an OD492 nm of 0.5. All samples were assayed in duplicate on separate plates, and the titers were calculated as the average of the two.
Measurement of T cell responses
Spleens were harvested and stimulated with the H-2d-restricted p7g gag peptide (18) and stained for intracellular IFN-γ, as follows. Erythrocyte-depleted single-cell suspensions were prepared by treatment with Tris-buffered NH4Cl (Sigma). Nucleated spleen cells (1 × 106) were cultured in duplicate at 37°C in the presence or absence of 10 mg/ml p7g peptide. Brefeldin A (PharMingen) was added to block cytokine secretion. After 3–5 h, cells were washed, incubated with anti-CD16/32 (PharMingen) to block Fcg receptors, stained with FITC-conjugated CD8 mAb (PharMingen), and fixed overnight at 4°C in 2% (w/v) paraformaldehyde. The following day, cells were treated with 0.5% (w/v) saponin (Sigma) and incubated with PE-conjugated mouse IFN-g mAb (PharMingen) in the presence of 0.1% (w/v) saponin, washed, and analyzed using a FACScalibur flow cytometer (Becton Dickinson, San Jose, CA).
Statistical evaluations for each group (n = 10) are reported as the mean ± SE, except for Ab titers which were reported as the geometric mean titer (GMT). Differences among groups were performed using the ANOVA test. Experimental group means were considered significantly different from control groups if p was < 0.05.
Several strategies were used to determine the fate of DNA after i.m. injection. First, we used rhodamine-labeled DNA to delimit the macroscopic and microscopic distribution of DNA after i.m. injection. Second, we characterized cellular uptake of labeled DNA in tissue sections and in MNC isolated from muscle and the draining lymph nodes. Third, we analyzed transgene expression by RT-PCR and histochemical staining. Finally, we evaluated the effects of modifying the distribution and cellular uptake by reducing volume or electroporation on the immunogenicity of DNA vaccines by measuring expression of a reporter gene and immune responses. The goal was to determine whether distribution, cellular uptake, and transgene expression were directly related to the immunogenicity of DNA vaccines.
DNA distribution and cellular uptake
The overall localization of a standard dose (10 μg in 50 μl saline solution) of labeled DNA was determined after injection into the tibialis anterior muscle of BALB/c mice (Fig. 1). Using a low magnification fluorescence stereomicroscope, labeled DNA could be observed directly after injection (Fig. 1, A–C). Immediately after injection, the vaccine which clearly exceeded the fluid capacity of the muscle caused swelling of the anterior epimysial sheath. Thereafter, the epimysial sheath rapidly returned to normal position, apparently causing accumulation of labeled DNA along the myotendinous junction areas (arrowhead).
Localization of labeled DNA was further examined in sections cut transversely or longitudinally through the muscle. Labeled DNA was present throughout the entire transverse section 5 min after injection, in the interstitial space between muscle cells (Fig. 1,D, arrowheads). Some labeled DNA was found inside bundles of muscle cells localized near the injection site (arrow). Although injection of saline alone did not result in fluorescence of such intensity, it did cause a small yet detectable increase in low intensity autofluorescence of muscle cells near the injection site (data not shown). In longitudinal sections, most of the labeled DNA was located between muscle cells 5 min after injection (Fig. 1,E). Labeled DNA was internalized by cells that were round or oval 24 h after injection (Fig. 1,F). Confocal microscopy showed that labeled DNA (red) was contained within small cytoplasmic vesicles of these cells, but not in the nucleus (green) as determined using the nuclear stain YO-PRO-1 (Fig. 1 G). Furthermore, the rhodamine signal colocalized with a marker of endolysosomes (lysosome-associated membrane protein 1, LAMP-1), suggesting that the DNA was internalized by phagocytosis into a degradative compartment (data not shown).
Labeled DNA was also found in the draining popliteal lymph nodes after i.m. injection (Fig. 2). Fluorescence was detected there as early as 3 h, and fluorescence intensity was greatest at 24 h after injection (Fig. 2,A). At this time point, fluorescence was mostly located in the subcapsular sinus of the lymph nodes. Cellular uptake of DNA in the lymph nodes was characterized further by isolating MNC on a density gradient after enzymatic digestion with collagenase D. Analysis by confocal microscopy showed that all cells that internalized labeled DNA in the draining lymph nodes were immunoreactive for the myeloid cell marker CD11b (Fig. 2 B).
To determine whether MNC that internalized DNA can potentially function as APCs, we analyzed the expression of costimulatory molecules by flow cytometry (Fig. 3). Low density MNC were isolated from muscle and draining lymph nodes 24 h after i.m. injection of FITC-labeled DNA and stained with PE conjugated mAbs. By flow cytometry analysis, the contour plot of cells that internalized FITC-labeled DNA was clearly distinguishable on the green channel. MNC that internalized labeled DNA were immunoreactive for the myeloid marker CD11b, regardless of whether they were isolated from muscle or the draining lymph nodes (Fig. 3, A and B). In addition, MNC that internalized FITC-labeled DNA isolated from the draining lymph nodes also expressed the costimulatory molecules CD80 and CD86 (Fig. 3, C and D). These costimulatory molecules were not expressed on MNC isolated from muscle (data not shown).
DNA transgene expression
We next sought to determine whether muscle and MNC that internalized DNA were capable of expressing the transgene after i.m. injection. Transgene expression was assessed by RT-PCR and histochemistry at the time when maximum fluorescent DNA was observed in the draining node, 24 h after injection of plasmids encoding HIV gag or β-galactosidase, respectively. Template cDNA was prepared using RNA extracted from the tibialis anterior muscle or from MNC isolated from the draining lymph nodes on a density gradient after enzymatic digestion with collagenase D. Extracts were subsequently amplified by PCR using HIV gag-specific primers to monitor transgene expression and mouse β-actin specific primers as a control for the cDNA synthesis and PCR. An HIV gag-specific PCR product was readily detected in whole muscle extracts, but not in MNC isolated from the draining lymph nodes at 24 h postimmunization (Fig. 4). MNC isolated from draining nodes at 48 h and 7 days and from muscle at 24 h were also negative for transgene expression by RT-PCR (data not shown). As a positive control, we used RNA extracts from 293 cells that were transfected with the same HIV gag plasmid that was injected in vivo. Whereas a PCR product specific for β-actin was detected in all extracts, no PCR product was detected in the absence of the reverse transcriptase step. Thus, transgene expression was readily detected by RT-PCR in muscle, but not in MNC after i.m. injection. Similarly, tissue sections of muscle and lymph node stained for β-galactosidase did not show detectable levels of transfection in non-muscle cells (data not shown).
Effects of distribution and cellular uptake on DNA vaccine immunogenicity
In the following set of experiments, we sought to determine the effect of modifying distribution and cellular uptake on DNA vaccine expression and immunogenicity. Two methods were used toalter the distribution of DNA in vivo. First, the volume of the injected DNA was reduced to limit the overall dispersion of DNA within the tissue. Such a reduction is also likely to decrease the hydrostatic pressure caused by the injection. Second, cellular uptake of DNA was facilitated by electroporation of the tibialis anterior muscle in vivo immediately after injection.
The effects of modifying the injection volume and of electroporation were determined by microscopy immediately after injection of a rhodamine-labeled plasmid encoding β-galactosidase (Fig. 5). Changes in cellular uptake were monitored by tracking the localization of labeled DNA, whereas changes in transgene expression were monitored by histochemical staining for β-galactosidase activity. A 50-μl dose of labeled DNA typically localized in the interstitial space separating muscle cells and inside muscle cells near the injection site (Fig. 5,A; see also Fig. 1,D). Labeled DNA was also detected in the interstitial space after injection of the same DNA dose in a 5-μl volume, although it dispersed to a lesser extent through the muscle (Fig. 5 B). Cellular uptake of labeled DNA by muscle cells was clearly diminished after injection of DNA in a low volume.
Electroporation in vivo caused a sharp increase in the intensity and number of muscle cells that internalized labeled DNA (Fig. 5,C). The effect of electroporation in vivo on DNA uptake by muscle cells was particularly striking, given that labeled DNA could be detected by confocal microscopy within the nuclei of the muscle cells (Fig. 5, D–F). The red channel was used to identify rhodamine-labeled DNA and the green channel to identify nuclei after sections were stained with YO-PRO-1. Superimposition of images demonstrated colocalization of labeled DNA within nuclei of muscle cells (arrowheads). Muscle cell nuclei could be distinguished unambiguously from nuclei of other cell types, because mature muscle cells have multiple elongated nuclei located near the plasma membrane. Under these conditions, electroporation in vivo did not result in a detectable increase of DNA uptake by MNC (data not shown). Thus, electroporation in vivo seemed to facilitate delivery of labeled DNA to the nucleus of muscle cells.
To examine the extent to which the DNA taken up was functional, we examined transgene expression after DNA injection with or without electroporation. Expression was analyzed in transverse sections of the tibialis anterior muscle 24 h after injection. Histochemical staining showed that typically 2–10 muscle cells per field were stained after injection of a 50-μl dose of DNA (Fig. 5,G). However, typically 20–100 muscle cells were stained after injection of the same dose followed by electroporation in vivo (Fig. 5 H). No staining was observed inside MNC located in the interstitial space between muscle cells (data not shown). These results demonstrate that increased DNA uptake by muscle cells results in increased transgene expression.
The effect of modulating the distribution and cellular uptake of DNA was determined by quantifying DNA gene expression and immunogenicity (Fig. 6). For DNA expression, a luciferase reporter plasmid was injected i.m., and luciferase activity was determined 14 days later (Fig. 6 A). Injection of the luciferase plasmid in 50 μl saline solution resulted in an average expression value of 10 ± 3 pg/muscle (n = 10 mice/group). Luciferase activity was increased ∼18-fold when the 50-μl injection was followed by electroporation in vivo (184 ± 37 pg/muscle). Conversely, injection of the reporter plasmid in a small volume (5 μl) resulted in substantially less luciferase activity (2 ± 0.2 pg/muscle). However, electroporation restored the decreased luciferase expression (13 ± 4 pg/muscle).
For DNA vaccine immunogenicity, a plasmid encoding HIV gag was injected into the tibialis anterior muscle, and serum titers were measured by ELISA 6 wk later (Fig. 6,B). Injection of the HIV gag plasmid in a standard volume (50 μl) of saline solution resulted in modest Ab titers (GMT =932; n = 10 mice/group). The magnitude of this response was increased ∼6-fold by electroporation in vivo (GMT = 5513). Conversely, injection of the same amount of HIV gag plasmid in a small volume resulted in substantially lower Ab titers (GMT = 13). As with gene expression, electroporation in vivo was able to enhance the efficacy of the vaccine given in a small volume (GMT = 3287). CD8 T cell responses were similarly affected by electroporation and low volume (Fig. 6 C). Taken together, these results are consistent with the hypothesis that DNA vaccine potency, as measured by both Ab and CD8-restricted T cell responses, can be influenced by cellular uptake and expression by muscle cells.
In the present study, we used fluorescence-labeled plasmid DNA to examine the delivery of DNA vaccines. By visualizing the fate of DNA vaccines, we have identified some of the potential barriers to efficient transfection of cells in situ, leading to a better understanding of some of the key cellular events responsible for priming an immune response after i.m. injection of DNA vaccines. We showed that a standard vaccine volume of 50 μl exceeds the fluid capacity of the mouse tibialis anterior muscle, resulting in vaccine dispersion not only throughout the entire muscle but predominantly into the interstitial space between the muscle body and the epimysial sheath immediately after injection. Fluorescence was detected even in the draining lymph node within 3 h of injection (Data not shown). Under these conditions, labeled DNA was taken up rapidly by muscle cells near the injection site, and gradually by MNC that were located between muscle fibers and in the draining lymph nodes. However, most, if not all, transgene expression was confined to muscle cells, and could be influenced by factors modifying DNA uptake by muscle cells. The immunogenicity of DNA vaccines was affected correspondingly, substantiating the role of muscle cells as the primary source of Ag after i.m. injection.
The mouse represents a widely used experimental model for characterization of immune responses induced by DNA vaccines. Here, the distribution of labeled DNA after i.m. injection was determined by microscopy. The injection of 50 μl caused immediate swelling of the anterior epimysial sheath, spreading of DNA throughout the tibialis anterior muscle, and DNA uptake by muscle cells. Such uptake is consistent with previous reports by other groups, which have shown that factors such as needle type, orientation and speed of injection, volume and type of injection fluid, and preinjection of hypertonic solutions can influence gene expression after i.m. injection (15, 18, 19, 20, 21). In one of these studies, transgene expression was detected as early as 2 min after injection of a luciferase reporter plasmid into mouse muscle (15). Although extracellular DNA is rapidly degraded and cleared from muscle within hours (22), administration of inhibitors of DNA degradation did not increase expression of a luciferase reporter gene (21). These observations suggest that in the mouse model, uptake and expression of foreign DNA by muscle cells occur very rapidly by a process that is still unknown. One possibility is that DNA entry into mature muscle cells is facilitated by T tubules, which are found only in skeletal and cardiac muscle (23). In addition, the multiple nuclei in muscle cells may increase the probability of DNA reaching the nucleus (24). However, several groups have reported uptake and expression of DNA in other tissues such as liver and lung after direct injection (25, 26, 27), and further studies are necessary to fully understand the mechanism of cellular entry of DNA in vivo. Our findings suggest that muscle cells may be particularly susceptible to hydrostatic pressure created by the injection of the vaccine into the relatively small tibialis anterior muscle. Such pressure could be generated as the total vaccine volume physically distends the extracellular space in the muscle. This process may modify the permeability of muscle cells, thereby facilitating the transfer of macromolecules across the plasma membrane. The injection of 5 μl, while resulting in dispersion of vaccine throughout the tibialis anterior muscle, did not result in swelling of the muscle, demonstrable uptake of DNA at the site of injection, or strong immune response. In vivo electroporation, however, increased DNA uptake and expression at the site of injection and immune response for both injection volumes. These results may provide insight into one of the variables leading to lower immunogenicity of DNA vaccines in larger animals, where the relative volume of inoculum to mass of muscle is much smaller than that thus far used in the mouse and the corresponding hydrostatic pressure is lower. These findings provide additional rationale for both routine use of small volume injections in the mouse model and continuation of ongoing investigation into the utility of electroporation in larger animals.
As shown by rhodamine signal, MNCs also internalized substantial amounts of plasmid DNA. However, DNA uptake by MNCs appeared to occur through a mechanism distinct from that of muscle cells. Detectable DNA accumulation in MNCs was gradual, restricted to cytoplasmic vesicles, and independent of injection volume. This pattern suggests that DNA uptake by MNCs is part of a constitutive pathway for sampling extracellular molecules, which is consistent with the capacity of these cells to endocytose a wide variety of molecules. MNCs that internalized DNA failed to express detectable amounts of either gag RNA or β-galactosidase under our conditions. However, MNC that are transfected by plasmid DNA after i.m. injection likely contribute to the magnitude and quality of the immune response. It has been shown that dendritic cells transfected with DNA vaccines in vitro can efficiently prime immune responses after transfer into naive animals (28, 29). In addition, transfection of APCs in vivo by DNA vaccines has been shown indirectly and directly. First, Torres et al. (30) demonstrated that removal of the injection site shortly after DNA immunization with a 50-μl i.m. injection did not abrogate priming of immune responses, indicating that cells distal to the injection site, possibly APCs, were transfected under these conditions. These data are consistent with our observation that fluorescent DNA is found in the draining node at early time points after injection of 50 μl. Second, APCs isolated from DNA vaccine-injected tissue were shown to present Ag in vitro, indicating that Ags were either expressed in or acquired by these cells (12). Third, targeting Ag for rapid degradation and presentation by MHC class I molecules can enhance the magnitude of CTL priming (31), which would be expected to happen if the Ags were expressed within APCs. Finally, dendritic cells and macrophages containing plasmid DNA and the transgene product have been found in the draining lymph nodes and in the spleens of vaccinated mice after administration by scarification of the skin (13), i.m. injection (32), and intradermal injection (33). In addition, plasmid DNA taken up by APCs may contribute to the priming of Ag-specific immune responses through activation of innate immune responses. Plasmid DNA contains unmethylated CpG motifs that induce lymphoid cells to release cytokines such as IFN-γ, IL-12, and IL-18 (34, 35). In turn, these cytokines may direct immune responses toward a T helper type 1 profile, which is consistent with the strong cell-mediated immunity induced by DNA vaccines. Thus, uptake of plasmid DNA by APCs may contribute to Ag presentation independently of or synergistically with Ag expression by these cells.
A similar body of evidence now exists for the involvement of non-APCs, such as myocytes, in the induction of immune responses by DNA vaccines. First, transplantation of stably transfected myoblasts into F1 and bone marrow chimeric mice has shown that transfer of Ag from muscle cells to APCs can occur (10, 36). Second, Doe et al. (8) demonstrated that adoptive transfer of APCs into immunodeficient scid mice was capable of supporting CTL priming up to 3 wk after DNA immunization, suggesting that transfer of Ag from transfected host cells to APCs occurred. Third, CTL responses can be induced in mice even when expression of Ag is limited to muscle cells through the use of a muscle-specific promoter (37). Finally, using a controllable plasmid expression system and adoptive transfer, Corr et al. (38) recently demonstrated that the bulk of the immune response after needle injection of plasmid DNA is dependent on Ag expression by nonlymphoid tissues and subsequent transfer of Ag to APCs. Therefore, cross-priming also plays a role in the induction of immune responses by DNA vaccines. The correlation we present here between DNA distribution, uptake, expression by muscle cells, and the magnitude of immune responses is consistent with this hypothesis.
Implications for DNA vaccine design
It is highly desirable to increase the potency of DNA vaccines and several avenues can be envisaged to attain this goal. As demonstrated in the present study, one approach to improving immunogenicity is to increase DNA uptake by muscle cells, which can be accomplished by electroporation in vivo (17, 39). Alternatively, facilitating transfection of APCs in vivo may increase the potency of DNA vaccines, and several studies have demonstrated the feasibility of this approach (7, 40, 41). For example, a gene gun can be used to bombard DNA inside APCs residing within the skin, which subsequently migrate and prime immune responses in the draining lymph nodes. Because injected DNA requires cellular uptake, it is found inside endolysosomes of APCs after i.m. injection. Hence, formulations that facilitate the egress of plasmid DNA out of this degradative compartment may augment transfection of these cells. It appears there is a window of time for this to occur before DNA degradation, because we were able to isolate intact and functional plasmid from these cells for ∼18 h after injection (our unpublished observations). DNA entry into the cytoplasm may be facilitated through the use of DNA formulations that may destabilize the endosomal membrane, such as chloroquine (42), fusogenic peptides (43), or DNA adsorbed onto cationic microparticles (44).4 In any case, transfection opens the possibility of introducing additional genes inside APCs such as immunoregulatory molecules, making it possible to induce or suppress immune responses in the host. Other means of increasing immune responses induced by DNA vaccines are currently under investigation, including coadministration of DNA-encoding cytokines, chemokines, costimulatory molecules, liposomes, and other experimental adjuvants.
In summary, we have used sensitive techniques to follow the distribution, cellular uptake, and expression of DNA vaccines to better understand the limitations to transfection in situ. We found that the only cells detectably transfected after i.m. injection were muscle cells and established a direct relationship between muscle transfection and DNA vaccine potency. However, much of the injected DNA was phagocytosed into a dead end compartment within MNCs, which could explain their inability to express detectable transgene product. Therefore, strategies to increase DNA uptake by muscle cell or to facilitate DNA entry into the nucleus of APCs are likely to increase the potency of DNA vaccines.
We thank Marie-Laure Denereaz for invaluable help in the preparation of the manuscript, Phil Felgner (Gene Therapy Systems) for discussions pertaining to the rhodamine-labeled plasmid, and Georg Widera (Genetronics) for assistance in setting up the electroporation studies.
This work was supported in part by National Institutes of Health Grant HL 24136 from the National Heart, Lung, and Blood Institute and by Grant S97-25 from the University of California BioSTAR Project.
Abbreviations used in this paper: MNC, mononuclear cells; X-gal, 5-bromo-4-chloro-3-indolyl-6-d-galactose; GMT, geometric mean titer.
K. S. Denis-Mize, M. Dupuis, M. L. MacKichan, M. Singh, D. O’Hagan, J. Donnelly, D. McDonald, and G. Ott. Plasmid DNA adsorbed onto PLG-CTAB Particles mediates target gene expression and antigen presentation by dendritic cells. Submitted for publication.