Attenuated Shigella as a DNA Delivery Vehicle for DNA-Mediated Immunization

Direct inoculation of DNA, in the form of purified bacterial plasmids that are unable to replicate in mammalian cells but are able to direct cell synthesis of foreign proteins, is being explored as an approach to vaccine development. Here, a highly attenuated Shigella vector invaded mammalian cells and delivered such plasmids into the cytoplasm of cells, and subsequent production of functional foreign protein was measured. Because this Shigella vector was designed to deliver DNA to colonic mucosa, the method is a potential basis for oral and other mucosal DNA immunization and gene therapy strategies.

[H202] with a correlation existing between stimulated catalase-loaded cells suggests that the 22. Confluent VSMCs with and without catalase loading the magnitude and duration of an increase fluorescent signal after the addition of growth factor were stimulated for 30 min with 10 mM SNP. Cells in [H202] and the level of tyrosine phos-1.is predominantly derived from H202 (see Fig. 2E).
were subsequently lysed with 6% trichloroacetic acid 1.VSMCs deprived of serum for 3 days were stimulated (TCA). Lysates were neutralized, and equal amounts phorylation (10). These latter observations with the indicated concentration of H202 or PDGF-AB (2 x 1 05 cells) were used to determine amounts of also strengthen the case that in many ways (5 ng/ml) for 20 min, after which cells were harvested cGMP by radioimmunoassay (Amersham) according H 02 fulfills the definition of an intracel-and lysed in RIPA buffer 150 mM tris-CI (pH 7.5), 150 to the manufacturer's recommendation.
fluenced by extracellular catalase in vivo.
ferred to a nitrocellulose filter. When indicated, filters 24. Confluent VSMCs were maintained in media without were probed with either an antibody to phosphotyserum for3 days and then, as indicated, stimulated with Recent epidemiological studies suggest a car-rosine (RC20), a broadly reactive antibody to MAP PDGF-AB (5 ng/ml). After 18 hours, cells in quadruplidioprotective effect of antioxidants (15 (3,4). The that in the absence of DAP, 15D and assay point, a loss of 1 to 1.5 log uInits of resultant 15D construIct, a Aasd isolate of 15D(pCMVP) entered 13% and 10% of the viable bacteria occUrred with no notable Shigella flcxneri 2a strain 2457T, was able to cuIltuired BHK cells, respectively. By con--difference between strains 15FL) and inaintain the eUkaryotic expression vector trast, 33%Xo (15SD) and 29% [151D(pCMVP)J 151)(pCMVP). However, at both the 24-pCMV3 (5) in thc absence of antibioticof the BHK cells contained bacteria when iandl 48-hour assay points, increasing Units of selective presstLire. The plasmnid pCMV3 cx-DAP was present dLring the invaision step.
3-galactosidase activity were readily detectpresses Escherichia coli P-galactosidase uinder AlthoUgh both constrtucts were able to ined in extracts of BHK cells infected with the control of the itm-mediate early promloter vade BHK cells, the addition of DAP dUring, 1 S)D(pCMV) The detected P-galactosiand enhancer fromn the humliian cytomegallothe invasion step increased thie nlluimber of dase activity did not resuLlt fromn expression virus (CMV) in mammalian cells; thi's per-BHK cells infected and the numiber of via-witlhin the bacteria because, although no mitted easy analysis of mnammalian cell-hle bacteria recovered (9).
activittv Xas metsured at the first two assay meiediated gene expression after delivery.
To test the ability of 15D to deliVer points, large nuLmbers of viable bacteria Strain 15D was screenlied tO eIstire that plasmid DNA, we followed intracellular were precsent. In addition, aln isolate of the large plasmiid that is essential for bacte-bacterial viability and P-galactosidaise ac-I 5D(pCMVP) that did not express IpaB rial invasion of marmmaliln cells had not tivity (Fig. 1)  been lost duLring the genetic mnanipUtlatiols. (8, 10). Initially, 1 x 107 to 3 X 10' viable WiaS Unable to bring iboUt P-galactosidase Immu1noblots verified that the strain con-bacteria of eaclch strain were recovered fromii activity at the 24-hoLIr assay point.
tinuled to express tlhe invasion-associated mionolayers of BHK cells with no detectable Infected moionolayers of BHK cells were lpaB and IpaC polypeptides (6) and thus P-galactosidase activity in cell extracts. No immuALInLOstained to examiine P3-galactosidase showed no loss of tlhe invasion plasmild. To P3-galactosidase activity coultd be detected expression within individUal cells (Fig. 2) confirmiearlier observations, we tested 1 5) in bacterial extracts that were eqUivatlent to (8, 1 1 ). No intracellular imlulllnostainiing ainid 15D(pCMVP) for the ability to invadce tlhc total numiber of bacteria added. At eaclh w.as observed in monolayers infected with cuLltured baby hamster kidney (BHK) cells with and without DAD suIpplemientation Fig. 2. Intracellular imdrn thei 90 mmii allowed for inigv2aItaclllaoin dluring thc '0 m i (li ufetltor llx cl ion munostaining to detect (7)(8)(9). Exarmination by light microscopy of expression of P-galac-  (8). In this assay, indicated times, washed DAP (50 pLg/ml) was added to concentrated bac-monolayers were fixed in terial suspensions before these suspensions were phosphate-buffered 4% added to flasks of semiconfluent BHK cells (-1 x paraformaldehyde for 5 107 cells). At the indicated times, BHK cells were min and then blocked removed by trypsinization and washed in phoswith 3%o goat serum phate-buffered saline. A portion of the cell sus-(Gibco-BRL) in HBSS for pension was lysed with a 0.2% Triton X-100 solu-30 min. BHK cells were tion, diluted, and plated on TSA Congo red DAP then permeahilized for 1 min with HBSS containing 0.1r saponin (Sigma). A monoclonal antihody to plates to determine the numher of viahle hacteria.
3-galactosidase (Sigma) was diluted 1: 2000 in HBSS containing 0.1 % saponin and applied tot 30 min at 3-Galactosidase activity was measured in the re-37 C in a humidified chamher. Fluorescein isothiocyanate-conjugated secondary antihody to mouse IgO maining cell extract hy a standard hiochemical (Fc-specific, Sigma) was diluted 1::32 and applied for 30 min at room temperature. Between each step, assay (10) [units of 3-galactosidase =380 x chamher slides were washed extensively with HBSS containing 0.100 saponin. A final wash step of HBSS 0D400/time (in minutes)]. 3-Galactosidase activi-alone was used to close permeabilized cells. Fluorescent images were visualized with a Nikon Microphot ties were standardized to 1 mg of total protein, as with epifluorescence attachment or with an Olympus VANO4-S with fluorescence attachment. Original determined with a BCA* protein assay kit (Pierce).

300
-(CIEN(FE * V(L . 270 * 1 3 OCTOBE)RR 19)5) either strain at the 30-min assay point (Fig. foreign gene expression. Because pCMV, is staining were observed in the outer region of 2B). Only slight intracellular immunostain-a 7164-base pair plasmid that occurs in the cornea near the.sclera of the right eyes ing was detected at the 4-hour assay point -500 copies per bacterial cell, each bacteri-that received 15D(pCMV,B), except those in monolayers infected with 15D(pCMVP) um is estimnated to contain -3.93 x 10-9 ,ug from day 8, in which staining was detected ( Fig. 2, C and D). By the 24-and 48-hour of DNA. Thus, intracytoplasmic delivery of in only one of three comeas. Several areas assay points, positive immunostaining of no more than 4 x 10-9 to 20 x 10-9 ,g of typical of the staining observed in comeas several cells per field was observed in mono-DNA by ShigeUa was sufficient for expression that received 15D(pCMV,) are shown in layers infected with 15D(pCMVP) (Fig. 2, of 3-galactosidase. Fig. 3B. No apparent endogenous ,-galacto-E and F). Staining throughout the cyto-To demonstrate that gene delivery was sidase activity was detected in eyes inoculatplasm suggested that the plasmid DNA had not restricted to BHK cells, we infected ed with 15D. Histology experiments will be been released from the bacterium into the murine P815 cells that express H-2d class I needed to examine in greater detail the percell cytoplasm, leading to transcription and major histocompatibility complex (MHC) centage of cells and cell type(s) invaded by translation by the mammalian cell. Immu-molecules with 15D(pCMVP). As shown in 15D(pCMV,B) and those staining positive nostained cells also appeared to be rounded, Table 2, 56.25 units of 13-galactosidase ac-for 3-galactosidase. In an initial experiment, possibly because of the presence of a large tivity were detected in lysates from P815 spleen cells from intranasally inoculated quantity of ,B-galactosidase protein. As cells infected with 15D(pCMV3). Further BALB/c mice showed a moderate proliferameasured by fluorescence-activated cell experiments will be necessary to determine tive response to 0-galactosidase protein (2.5 sorter (FACS) analysis, 1 to 2% of 5000 whether these cells can present Shigella-jig/ml) (13,14). The stimulation index (14) 15D(pCMV,3)-infected BHK cells ex-delivered DNA-encoded foreign antigens in was 3.6 when the inoculum was supplementpressed 3-galactosidase at the 24-hour assay the context of class I. ed with DAP compared with 2.1 in the point (8, 10).
Studies of the ability of 15D to deliver absence of DAP. Although preliminary, Visual examination of Leukostat-stained plasmid DNA in vivo have begun in two these experiments indicate that bacteria can chamber slides of 15D(pCMVP)-infected small animal models, the guinea pig kerato-be used to deliver plasmid DNA in vivo.

BHK cells indicated that 28% of the cells conjunctival and murine intranasal models,
Our method for delivering functional contained one to five visually intact bactewhich are used to study Shigella pathogenic-DNA inside cells need not be restricted to rial cells, with 1.7% containing five bacteria ity and immunobiology (12,13). To deter-Shigella because the invasion genes used by (Table 1). Four hours after gentamicin treat-mine whether 15D could deliver pCMV,B to Shigella can be inserted into other bacteria ment, 26% of the cells contained visually the ocular surface of the guinea pig eye, we such as E. coli ( 15). Likewise, other bacteria intact bacteria, with <1% of the cells con-stained comeas for ,B-galactosidase activity such as Listeria are able to invade cells and taining four bacteria. Therefore, invasion and visually examined them at various times break out of the phagocytic vacuole into with one to five bacteria was required for after inoculation (12). Varying amounts of the cytoplasm (16). Although we have no formal proof that such a release of bacteria from the phagocytic vacuole into the cell