Sv

JSa

b

c

d

a

ARRAA

KHDSBT

1

eaPcrbrdciTv

v

(

aC

0h

Vaccine 30 (2012) 5733– 5739

Contents lists available at SciVerse ScienceDirect

Vaccine

jou rn al h om epa ge: www.elsev ier .com/ locate /vacc ine

almonella typhi Ty21a bacterial ghost vector augments HIV-1 gp140 DNAaccine-induced peripheral and mucosal antibody responses via TLR4 pathway

ing Wena,1,2, Yi Yanga,1, Guangyu Zhaoa, Shuang Tonga, Hong Yua, Xia Jinb, Lanying Duc,hibo Jiangc,d, Zhihua Koua,∗, Yusen Zhoua,∗

State Key Laboratory of Pathogen and Biosecurity, Beijing Institute of Microbiology and Epidemiology, Beijing 100071, ChinaDepartment of Medicine, University of Rochester, Rochester, NY 14642, USALindsley F. Kimball Research Institute, New York Blood Center, New York, NY 10065, USAMOE/MOH Key Laboratory of Medical Molecular Virology and Institute of Biomedical Sciences, Shanghai Medical College, Fudan University, Shanghai 200032, China

r t i c l e i n f o

rticle history:eceived 31 March 2012eceived in revised form 2 July 2012ccepted 5 July 2012vailable online 20 July 2012

a b s t r a c t

Because of their stability and ease of manipulation, DNA vaccines have considerable potential for elicitingimmune responses. However, they are limited by their weak immunogenicity, especially in humans. Toaddress this challenge, we explored a new strategy of HIV vaccine delivery using Salmonella typhi Ty21abacterial ghosts (BGs). We found that Ty21a BGs loaded with an HIV gp140 DNA vaccine (Ty21a BG-DNA)were readily taken up by murine macrophage RAW264.7 cells, and gp140 was efficiently expressed in

eywords:IV-1NA vaccinealmonella typhi Ty21aacterial ghosts (BGs)LR4/5

these cells. Peripheral and intestinal mucosal anti-gp120 antibody responses in mice vaccinated withBGs-DNA vaccine were significantly higher than those in mice immunized with naked DNA vaccine. Theenhancement of antibody responses was associated with BG-induced production of IL-10 through TLR4pathway. These results demonstrate that Ty21a BGs is a novel and effective delivery vehicle for DNAvaccines, which could therefore be used as a new strategy for development of HIV vaccines.

© 2012 Elsevier Ltd. All rights reserved.

. Introduction

Despite nearly three decades of research and developmentfforts to generate various prototypic HIV vaccines, none is avail-ble for human immunization. Disappointing results of large-scalehase II clinical trials of Merck’s V520 HIV vaccine (a trivalent vac-ine of gag, pol and nef for maximizing HIV-specific CD8+ T-cellesponses) [1,2] and Phase III clinical trials of VaxGen’s gp120-ased AIDS vaccine (AIDSVAX for inducing HIV-specific antibodyesponses) [3,4] have pushed HIV vaccine researchers back to therawing board. However, the encouraging results from a phase IIIlinical trial (RV 144) of the vaccine (ALVAC/AIDSVAX for elicit-

ng both HIV-specific humoral and cellular immune responses) inhailand in 2009 [5] has given new hope that a preventative HIVaccine may be possible. Current consensus holds that a protective

Abbreviations: BG, bacterial ghost; Ty21a BG-DNA, Ty21a BG-based HIV-1 DNAaccine; TSLP, thymic stromal lymphopoitin; PRR, pattern recognition receptor.∗ Corresponding authors. Tel.: +86 10 6385 8045; fax: +86 10 6385 8045.

E-mail addresses: zh 69kou@yahoo.com (Z. Kou), yszhou@nic.bmi.ac.cnY. Zhou).

1 These authors contributed equally to this work.2 Jing Wen is currently working at Shandong Provincial Institute of Dermatology

nd Venereology, Shandong Academy of Medical Science, Ji’nan, Shandong 250022,hina.

264-410X/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.vaccine.2012.07.008

vaccine against HIV infection should elicit T cell immunity, protec-tive antibody responses, and local mucosal immunity [6–9].

In animal models, DNA vaccines have been demonstrated toelicit cellular, humoral, and mucosal immune responses withthe help of a proper adjuvant or a particular delivery system[1,5,10–18]. Novel strategies are needed to improve the immuno-genicity of DNA vaccines. In recent years, a novel vaccine deliverysystem using bacterial ghosts (BGs) has been developed and shownto be an efficient and nontoxic delivery system for DNA vaccinesin vitro and in vivo [19–24]. Current research on the use of BGs asa DNA vaccine delivery system has been limited to two areas: theloading and expression of pEGFP-N1 in vitro, using either E. coli BGsor M. haemolytica BGs, and the evaluation of BG-mediated plasmidDNA delivery in vivo using a pCMV� construct that encodes for�-galactosidase [20–22].

Here, we introduce a novel DNA vaccine delivery system basedon Salmonella typhi Ty21a. Ty21a is a nonpersistent and safe oralvaccine against typhoid fever, and it has been approved by the U.S.Food and Drug Administration for human use. Additionally, it hasbeen used as a delivery vehicle for heterologous antigens [25,26].Importantly, as an intracellular bacterium, Ty21a can target the M

cells of mucosa-associated lymphoid tissue (MALT) [27,28]; thus,Ty21a has the potential to induce effective mucosal and systemicimmune responses. In this study, we prepared Ty21 BGs and opti-mized the conditions for the loading of DNA vaccine into the Ty21a

5 ne 30

BaiebTcts

2

2

tP(mwJ

2

meLBI

2

IDSaSp

2

RvsfmSB

2

bi1paamwttp

734 J. Wen et al. / Vacci

Gs. We found that the Ty21a BGs not only enhanced the peripheralntibody responses of the pSV140 DNA vaccine, but also inducedntestinal mucosal immune responses. To the best of our knowl-dge, this is the first report to show that a bacterial ghost system cane used to augment the delivery of a plasmid DNA vaccine for HIV.he immune regulatory functions of Ty21a BGs appear to be asso-iated with a biased stimulation of Th2-type cytokines and linkedo the activation of TLR4 pathways. These studies suggest a newtrategy for the development of antiviral vaccines.

. Materials and methods

.1. Cell line

The murine macrophage cell line, RAW264.7, was obtained fromhe Cell Culture Center of the Institute of Basic Medical Sciences,eking Union Medical College and maintained in DMEM mediumGIBCO) containing 1% of l-glutamine, 1% of penicillin and strepto-

ycin, and 10% or 2% of fetal bovine serum (FBS) (GIBCO) at 37 ◦Cith 5% CO2 in a CO2 incubator (SANYO Electric Co. Ltd, Osaka,

apan).

.2. Mice

Female 6–8-week-old BALB/c mice (Beijing Experimental Ani-al Center, Beijing, China) were used for all of the mouse

xperiments. The experimental protocols were approved by Theaboratory Animal Center, State Key Laboratory of Pathogen andiosecurity, Beijing Institute of Microbiology and Epidemiology

ACUC’s (The permitted number is BIME 2009-10).

.3. Preparation and identification of Ty21a bacterial ghosts (BGs)

The S. typhi Ty21a strain (kindly provided by Prof. J. Gao, Beijingnstitute of Microbiology and Epidemiology, China) was grown inifcoTM Nutrient Broth or Agar (Becton, Dickinson and Company,parks, MD) at the specified temperature. Ty21a BGs was prepareds previously described [20], and identified using scanning (Hitachi-2400) and transmission electron microscopy (PHILIPS). The pre-ared BGs were lyophilized and stored at −80 ◦C until use.

.4. Preparation of the Ty21a BG-based HIV-1 DNA vaccine

Loading of BGs with plasmid and BG-mediated transfection intoAW264.7 cells were optimized using pEGFP-N1 as described pre-iously [20,22]. Then, a HIV-1 DNA vaccine pSV140 containing aynthetic HIV-1 strain ThaiB-derived gp140 coding sequence (theull-length of gp120 and the ectodomain of gp41with codon opti-

ized for eukaryotic cell expression, gifted by Prof. Jianqing Xu,hanghai Fudan University) was loaded into the lyophilized Ty21aGs (hereafter termed Ty21a BG-DNA) using the same protocol.

.5. Inoculation of Ty21a BG-DNA into RAW264.7 cells

The Ty21a BG-DNA was added either to a 6-well (for Westernlot assay) or 24-well (for cytokine measurements) plates contain-

ng optimally grown RAW264.7 cells at a cell:BG ratio of 1:10 to:100. For cytokine measurements, supernatants were collectedrior to and at 12, 24, 36 and 48 h after Ty21a BG-DNA inoculationnd then stored at −20 ◦C until use. The Ty21a BG, DNA vaccinend medium were used as the experimental controls. For experi-ents detecting the mRNA levels of TLR4 and TLR5, RAW264.7 cells

ere treated with Ty21a BGs for varying times, harvested, washed

wice with PBS, and stored at −80 ◦C prior to RNA extraction. Forhe analysis of HIV-1 gp140 expression in RAW264.7 cells, the cellellets were collected for further western blot analysis, a 1/10,000

(2012) 5733– 5739

dilution of an HRP-labeled goat polyclonal antibody to HIV-1 gp120(AbCam, Cambridge, MA, USA) was used for WB.

2.6. Immunization with the Ty21a BG-DNA vaccine in mice

Mice were subcutaneously inoculated with Ty21a BG-DNA con-taining 100 �g of BGs, 20 �g of naked DNA or endotoxin-free PBS(GIBCO) as a control, and then vaccinated twice with the respec-tive vaccines at a 2-week interval. The mouse serum was collectedby tail-vein bleeding at the baseline time and 10 days after thefinal immunization and then stored at −20 ◦C for the measure-ment of anti-gp120 antibodies (IgG, IgG1 and IgG2a) by ELISA. Thegastrointestinal lavages to detect anti-gp120 IgA antibody werecollected when the mice were sacrificed 14 days after the lastimmunization.

2.7. RT-PCR

Total RNA from RAW264.7 cells was extracted using an RNeasymini kit (Qiagen, Valencia, CA) according to the manufacturer’sinstructions, and RT-PCR was performed as the manufacturer’sinstructions. In brief, first-strand cDNA was synthesized using theSuperscript III first-strand synthesis system (Invitrogen), then Plat-inum PCR Supermix (Invitrogen) was used for PCR at 30 cycles(95 ◦C for 30 s, 55 ◦C for 30 s, and 72 ◦C for 60 s) in a 2720 ThermalCycler (AB Applied Biosystems, Foster City, CA). Primers specific forthe mouse TLR4 and TLR5 genes (Invivogen, San Diego, CA) wereused for detecting the mRNA, and �-actin mRNA was used as acontrol. The PCR products were analyzed by electrophoresis on a1% agarose gel; the density of each band was quantified by Gel-ProAnalyzer 4 and expressed as the ratio of TLR4 or TLR5 signal to the�-actin signal.

2.8. Blocking assay

RAW264.7 cells were cultured overnight. MAb mTLR4/MD2and/or anti-mTLR5-IgG (Invivogen) or isotype control Rat IgG2a (BDBioscience, San Diego, CA) was used to pretreat the RAW264.7 cellsat the optimal concentrations and incubated on ice [29] for 10 min(for TLR4) or 60 min (for TLR5). IL-10 in the supernatants at 12 hpost-inoculation was quantification by ELISA. The concentration ofIL-10 from the BG-only treated cells was denoted as 100%, and thelevels of IL-10 from the other groups were reported as a ratio of thelevels from the BG-only treated cells.

2.9. Measurement of anti-HIV-1 gp120 antibody by ELISA

Vaccine-induced anti-HIV-1 gp120 total IgG, IgG1 and IgG2aantibodies in the serum and IgA antibodies in the gastrointestinallavages were detected by ELISA. The N-terminus immunodom-inant segment of HIV-1 gp120 (amino acid 30–110), which isimmunoreactive with all sera of HIV-1 infected patients, wasexpressed in E. coli (Abcam), was used as the capture antigenfor ELISA. HRP-conjugated goat anti-mouse IgG (BD Bioscience),IgG1, IgG2a, or IgA (Santa Cruz Biotechnology, Santa Cruz, CA)was used as the secondary Abs. The antibody titer was deter-mined as the reciprocal of the highest dilution when the absorbancevalue at 450 nm was greater than the average value of the nega-tive control (before immunization) plus three times the standarddeviation.

2.10. Measurement of cytokines by ELISA

Cytokine levels from cell culture supernatants were detectedusing IL-10, IFN� or IL-12 ELISA kits (Neobioscience, Shen-zhen, China) according to the manufacturer’s instructions, and

J. Wen et al. / Vaccine 30 (2012) 5733– 5739 5735

F ts. Sam ificati

wc

2

socw

3

3

sohf2eot1w

3m

aieittf(up

ig. 1. Scanning and transmission electron micrographs of Ty21a bacterial ghosicroscopy, as described in Section 2. (A) Scanning electron microscopy (SEM, magn

ere expressed as pg/ml, as calculated from the standardurves.

.11. Statistical analysis

Statistical analyses were performed using GraphPad Prism ver-ion 4.0 (GraphPad Software, San Diego, CA, USA). One-way analysisf variance (AVOVA) was used for the mean or geometric meanomparison from the multiple groups. A P-value of less than 0.05as considered statistically significant.

. Results

.1. Characterization of the Ty21a bacterial ghost (BG)

We produced the Ty21a BGs by temperature-induced expres-ion of the gene E from bacteriophage phi X174. The morphologyf the Ty21a BGs was examined by SEM (Fig. 1A) and TEM (Fig. 1B). Itas several notable features: (a) an intact surface structure except

or the lysogenic tunnels with diameter ranging from 100 nm to00 nm; (b) empty bacterial cells; (c) relatively collapsed cellularnvelope, compared to non-induced bacterial cells due to the lossf most cytoplasmic materials (Fig. 1A and B); (d) the lysis rate ofhe fresh Ty21a BGs was 96.77%, and the lyophilized BGs reached00%, suggesting that the lyophilized Ty21a BG lost its infectivityith correspondingly increased safety profile.

.2. The Ty21a ghosts can be efficiently loaded with plasmid andediate efficient expression of HIV-1 gp140

To determine whether the prepared Ty21a BGs could be useds a delivery vector for a DNA vaccine, we first optimized the load-ng and BG-mediated transfection conditions using the eukaryoticxpression plasmid, pEGFP-N1. As shown by a typical experimentn Fig. 2, nearly 100% of the cells were green under these condi-ions as examined by fluorescence microscopy (Fig. 2C), whereashe untreated RAW264.7 cells had no signal (Fig. 2D). This clear dif-

erence between signal and noise was also verified by FACS analysisFig. 2E). We next examined whether such a vector system could besed to deliver HIV-1 antigens. Using the same protocol, the gp140rotein (Fig. 2F) was efficiently expressed, indicating that Ty21a

lmonella typhi Ty21a bacterial ghosts were prepared and harvested for electronon: 20,000×); (B) transmission electron microscopy (TEM, magnification: 24,000×).

BGs can function as an efficient in vitro delivery vehicle for DNAvaccine.

3.3. The Ty21a BG-DNA elicits efficient peripheral and mucosalantibody responses against HIV-1

Having verified that the Ty21a BG-DNA could deliver antigensof interest in vitro, we then investigated whether it could be usedto deliver immunogens in vivo. No weight loss or side effects wereobserved, either during or after immunization with the Ty21a BG-DNA vaccine (data not shown). High levels of the anti-gp120 IgGantibody were detected in the serum of both the BG-DNA and DNA-alone groups, but the BG-DNA elicited a titer of IgG Ab that was 10fold higher than that elicited by the naked DNA (Fig. 3A). This sug-gests that the Ty21a BG is an efficient delivery vehicle for the nakedDNA vaccine and may also have adjuvant activities to augmentantibody responses.

Because Ab subclasses may be associated with vaccine efficacyand reflect the underlying T cell helper activation, we next ana-lyzed the anti-gp120 antibody subclasses. Fig. 3B shows that theBG-DNA elicits higher levels of IgG1 Ab (Th2-associated) than thenaked DNA, whereas, both BG-DNA and the naked DNA inducessimilar levels of IgG2a Ab (Th1-associated); these results indicatethat the BG-DNA augments the Th2-biased response. As gastroin-testinal mucosal immunity is an important armor against HIV-1transmission, we also examined gastrointestinal mucosal IgA anti-body responses. The results showed that both the naked DNA andthe BG-DNA induced detectable levels of IgA that specific for HIV-1gp120, but the BG-DNA elicits higher levels of IgA than the nakedDNA (P < 0.05) (Fig. 3C).

3.4. Cytokine responses induced by Ty21a BG-DNA vaccine

To determine why the BG-DNA appeared to induce a strongerantibody response than the naked DNA, we assessed cytokineinduction in vitro using RAW264.7 cells line as a model APC andstimulated cells with HIV-gp140 DNA vaccine, either formulatedwith the Ty21a BGs or not. As shown in Fig. 4, cells that were stim-

ulated with the BG-DNA or BG-alone produced similarly high levelsof IL-10 (a Th2-type cytokine), but the DNA-alone induced a muchlower level of IL-10. In comparison, the production of IFN� and IL-12(Th1-type cytokines) was not markedly different among the various

5736 J. Wen et al. / Vaccine 30 (2012) 5733– 5739

Fig. 2. Expression of the target gene encoded by the DNA vaccine via a ghost vector in the murine macrophage cell line, RAW264.7. Expression of GFP in RAW264.7 cellstreated by Ty21a BGs loaded with (A, C) or without (B, D) an EGPF expression plasmid pEGFP-N1 was detected by fluorescence microscopy (C and D) and FACS (E), respectively;A and B are corresponding bright field pictures of C and D (magnification fold 400×). Expression of HIV-1 gp140 in RAW264.7 cells treated by Ty21a BGs loaded with theHIV-1 DNA vaccine pSV140 was detected by Western blot using a HRP-anti-gp120 antibody (F).

An

ti- g

p1

20

Ig

G a

nti

bo

dy

tite

r(lo

g10)

A B

An

ti-g

p120 I

gG

su

bty

pe

an

tib

od

y

tite

r(lo

g1

0)

An

ti-g

p1

20

Ig

A a

nti

bo

dy

tite

r

BG+DNA DN A BG1

2

3

4

C

P < 0.01

P < 0.01

P < 0.05 P < 0.05

BG+DNA DNA BG1

2

3

4IgG1

IgG2a

BG+D NA DN A BG0

20

40

60

80

P < 0.05

P < 0.05

P<0.05 P< 0.05

Fig. 3. Induction of peripheral and intestinal mucosal anti-gp120 antibodies in vaccinated mice. BALB/c mice were inoculated with Ty21a BGs loaded with HIV-1 DNA vaccine,pSV140. Anti-gp120 total IgG (A), IgG1 and IgG2a antibodies in serum (B), and IgA antibodies in the intestinal mucosal wash (C) were detected by ELISA. Mean and SD ofresults from 6 mice are presented.

J. Wen et al. / Vaccine 30 (2012) 5733– 5739 5737

Time post inocul ation (h)

Cyto

kin

e l

eve

ls (

pg

/ml)

BG+

DNA

BG

DNA

IL-10

12h 24h 36h 48h0

500

1000

1500

2000

2500

12h 24h 36h 48h0

500

1000

1500

2000

2500

12h 24h 36h 48h0

500

1000

1500

2000

2500

12h 24h 36h 48h0

5

10

15

20

25

12h 24h 36h 48h0

5

10

15

20

25

12h 24h 36h 48h0

5

10

15

20

25

IFNγ IL-12

12h 24h 36h 48h0

10

20

30

40

50

12h 24h 36h 48h0

10

20

30

40

50

12h 24h 36h 48h0

10

20

30

40

50

Fig. 4. Stimulation of cytokine secretion in RAW264.7 cells pulsed with vaccines in vitro. Levels of cytokines in the supernatants of RAW264.7 cells treated with Ty21a BG-DNA,T d by Et aselinr

taa

3

tTae1boc(Bc

ttmtmTi

4

n

y21a BG, or DNA vaccine, respectively, at the different time points were measurehe baseline levels at the different time points. The dashed lines (- - -) indicate the besults are representative of three separate experiments.

reatment groups. These results suggest a correspondence between higher level of IL-10 production and a stronger BG-DNA-mediatedntibody response.

.5. TLR pathways are activated by Ty21a BGs

LPS and flagellin, which are agonists of TLR4 and TLR5, respec-ively, are the major structures on the cell wall of Ty21a [30,31].hus, Ty21a BGs may activate macrophages partially via the TLR4nd TLR5 pathways. We found that BG-treated cells had lower lev-ls of TLR4 mRNA compared with the untreated cells at both 6 and2 h after treatment. In contrast, TLR5 mRNA levels were high toegin with in both the BG-treated and control PBS-treated cells, butnly the BG-treated cells sustained their TLR5 mRNA levels; in theontrol cells, the TLR5 mRNA was remarkably decreased with timeFig. 5A–C). These results imply a potential interaction betweenGs and TLR4/5 pathways that could lead to cellular activation andytokine secretion.

To further test this scenario, we used specific neutralizing Abso TLR4 or TLR5 to block the possible interactions between BGs andhese TLRs. RAW264.7 cells pretreated with the anti-mTLR4/MD2

Ab produced significantly lower levels of cytokines than cellsreated with the isotype control, and treatment with the anti-

TLR5-IgG showed no statistical significant difference (Fig. 5D).hus, interactions between BGs and TLRs may exist, and a BG-TLR4nteraction may play a particularly important role.

. Discussion

One of the key issues of HIV-1 vaccine development is to findew adjuvants and/or delivery systems for various modalities of

LISA. The levels of cytokines in the supernatant from untreated cells representede cytokine levels in cultured RAW264.7 cells without the addition of vaccines. The

HIV-1 vaccines. In this study, we prepared an S. typhi Ty21a BGs andinvestigated its capacity to deliver a plasmid DNA HIV-1 vaccinein vitro and in vivo. We found that this novel experimental systemwas capable of inducing HIV-specific IgG antibodies in the bloodand IgA antibodies in intestinal mucosa. Additionally, the BG-DNAis superior to naked DNA in eliciting these immune responses.

The efficiency of this new delivery system is excellent. We foundthat one lyophilized Ty21a BG could load 103–104 copies of plasmidDNA, a level that is similar to that in the M. haemolytica ghost system[20]. This finding is very different from the results using the live S.typhi Ty21a, which has a very low capacity to carry plasmid DNAvaccines [32].

Another advantage of this new delivery system is its ability toelicit both stronger systematic and mucosal antibody responses.We observed that Ty21a BGs increased the levels of specific anti-gp120 antibody IgG in the serum more than ten-fold relative tothe naked DNA vaccine, and the levels of the Ab subclasses IgG1and IgG2a remained at similar levels in the BG-DNA-immunizedmice, suggesting that Ty21a BG-based HIV-1 DNA vaccine stim-ulated a Th2-dominant Th1/Th2 mixed immune response, whilethe naked DNA vaccine elicited a Th1-dominant Th1/Th2 mixedimmune response. Such findings are in agreement with previousfindings in the M. haemolytica BG delivery of a DNA vaccine [20].

Although the underlying mechanism of the “adjuvant effect” ofthe Ty21a bacterial ghosts is likely to be complex, we exploredpotential mechanisms by examining the ability of Ty21a BGs toelicit the expression of specific signature cytokines and the poten-

tial involvement of TLR pathways. Experiments using cell lines(THP-1 cells and RAW264.7 cells), human PBMCs, and dendriticcells in vitro have suggested that BGs stimulate antigen presentingcells (APCs) to produce higher levels of Th2-polarizing cytokines

5738 J. Wen et al. / Vaccine 30 (2012) 5733– 5739

Fig. 5. The role of TLR4 and TLR5 in the Ty21a BG-stimulated activation of RAW264.7 cells. The levels of TLR4 and TLR5 mRNA in RAW264.7 cells treated with BGs or PBSwere observed by RT-PCR: (A) the agarose gel analysis of the PCR products of TLR4 (upper panel), TLR5 (middle panel) and �-actin (lower panel); (B) the ratio of TLR4 signalto the �-actin signal from the gel; (C) the ratio of TLR5 signal to the �-actin signal from the gel. (D) Effect of anti-TLR4 or TLR5 antibodies on the production of IL-10 fromR quantw

atrusBTcnntp

fvowtmtioctitr

a

AW264.7 cells treated with BGs. IL-10 level in the cell culture supernatants were

ith different Ab to that of BG-treated cells.

nd relatively lower levels of Th1-polarizing cytokines [20,33],hereby enabling BG-mediated delivery of DNA vaccine to elicitobust humoral and cellular immune responses. Our in vitro resultssing murine macrophage cell line RAW264.7 demonstrated someimilar and some different cytokine patterns: both the BG-DNA andGs alone stimulated RAW264.7 cells to produce high levels of theh2-poalrizing cytokine, IL-10, but similar levels of Th1-polarizedytokines (IFN� and IL-12) compared with stimulation with theaked DNA vaccine. Similar results were also found in the vacci-ated mice in vivo (data not show). This may help to explain whyhe Ty21a BG increased Ab responses and suggests its role as arospective adjuvant.

We then asked how Ty21a BGs could activate APCs. Differentrom the traditional physical or chemical methods used to inacti-ate bacteria, BGs were prepared by gentle lysis to allow retentionf all the structures on the cell envelope, such as LPS and flagella,hich can be recognized by TLR4 and TLR5 [30,31], implying that

his adjuvant role could be linked to the activation of these twoolecules. We found that levels of TLR4 and TLR5 mRNA from BG-

reated RAW264.7 cells were different from those of untreated cells,ndicating that both TLR4 and TLR5 were involved in the activationf the cells due to the BGs treatment. Subsequently, blocking theells with specific mAbs to TLR4 significantly inhibited the produc-ion of IL-10. These data suggest that Ty21a BGs activate the cellsn part via the TLR4 pathway and, possibly, via TLR5 pathway. Fur-

her studies are needed to demine whether other PPRs (patternecognition receptor) are involved in activating APCs.

In summary, we demonstrated that the Ty21a BG is a novel, safe,nd effective delivery vehicle for DNA vaccines; and it can enhance

itated by ELISA, and values were expressed as the ratio of the IL-10 of cells treated

the ability of a DNA vaccine to elicit peripheral and mucosal anti-body immune responses, presumably through the upregulation ofthe Th2 immune response via the TLR4 and possibly TLR5 path-ways. We therefore suggest that Ty21a BGs be further explored asa novel tool to improve the delivery of existing HIV vaccines forefficient induction of mucosal immunity. Evidently, such a systemis also suited for vaccine development against other pathogens.

Acknowledgements

This work was supported by grants from NSFC (#30972392 and#81173098) and National 863 Program of China (#2007AA02Z414).

Conflict of interest statement: There is no conflict of financialinterest.

References

[1] Buchbinder SP, Mehrotra DV, Duerr A, Fitzgerald DW, Mogg R, Li D, et al. Effi-cacy assessment of a cell-mediated immunity HIV-1 vaccine (the step study): adouble-blinded, randomized, placebo-controlled, test-of-concept trial. Lancet2008;372:1881–93.

[2] Fitzgerald DW, Janes H, Robertson M, Coombs R, Frank I, Gilbert P, et al. An Ad-5-vectored HIV-1 vaccine elicits cell-mediated immunity but does not affectdisease progression in HIV-1-infected male subjects: results from randomizedplacebo-controlled trial (The Step Study). J Infect Dis 2011;203:765–72.

[3] Flynn NM, Forthal DN, Harro CD, Judson FN, Mayer KH, Para MF, et al. Placebo-controlled phase 3 trial of a recombinant glycoprotein 120 vaccine to prevent

HIV-1 infection. J Infect Dis 2005;191:654–65.

[4] Gilbert PB, Peterson ML, Follmann D, Hudgens MG, Francis DP, Gurwith M, et al.Correlation between immunologic responses to a recombinant glycoprotein120 vaccine and incidence of HIV-1 infection in a phase 3 HIV-1 preventivevaccine trial. J Infect Dis 2005;191:666–77.

ne 30

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

J. Wen et al. / Vacci

[5] Rerks-Ngarm S, Pitisuttihum P, Nitayaphan S, Kaewkungwal J, Chiu J, ParisR, et al. Vaccination with ALVAC and AIDSVAX to prevent HIV-1 infection inThailand. N Engl J Med 2009;361:2209–20.

[6] Ortiz AM, Klatt NR, Li B, Yi Y, Tabb B, Hao XP, et al. Depletion of CD4+ T cellsabrogates post-peak decline of viremia in SIV-infected rhesus macaques. J ClinInvest 2011;121:4433–45.

[7] Walker BD, Ahmed R, Plotkin S. Use both arms to bit HIV. Nat Med2011;17:1194–5.

[8] Mcmichael AJ, Borrow P, Tomaras GD, Goonetilleke N, Haynes BF. The immuneresponse during acute HIV-1 infection: clues for vaccine development. Nat RevImmunol 2010;10:11–23.

[9] Haase AT. Targeting early infection to prevent HIV-1 mucosal transmission.Nature 2010;464:217–23.

10] Jones DH, Corris S, McDonalds S, Clegg JC, Farrar GH. Poly (dl-lactide-co-glycolide)-encapsulated plasmid DNA elicits systemic and mucosalantibody responses to encoded protein after oral administration. Vaccine1997;15:814–7.

11] Berzofsky JA. A push–pull vaccine strategy using Toll-like recep-tor ligands, IL-15, and blockade of negative regulation to improvethe quality and quantity of T cell immune responses. Vaccine 2011,http://dx.doi.org/10.1016/j.vaccine.2011.11.034.

12] Van Roey GA, Arias MA, Tregoning JS, Rowe G, Shattock RJ. Thymic stromal lym-phopoitin (TSLP) acts as a potent mucosal adjuvant for HIV-1 gp140 vaccinationin mice. Eur J Immunol 2012;42:353–63.

13] Vecino WH, Morin PM, Agha R, Jacobs Jr WR, Fennelly GJ. Mucosal DNA vaccina-tion with highly attenuated Shigella is superior to attenuated Salmonella andcomparable to intramuscular DNA vaccination for T cells against HV. ImmunolLett 2002;82:197–204.

14] Woo PCY, Tsoi HW, Leung HCH, Wong LP, Wong SSY, Chan E, et al. Enhancementby ampicillin of antibody responses induced by protein antigen and a DNAvaccine carried by live-attenuated Salmonella enterica serovar Typhi. Clin DiagnLab Immunol 2000;7:586–99.

15] Huang X, Liu L, Ren L, Qiu C, Wan Y, Xu J. Mucosal priming with replica-tive Tiantan vaccinia and systemic boosting with DNA vaccine raisedstrong mucosal and systemic HIV-specific immune responses. Vaccine2007;25:8874–84.

16] Huang X, Lu B, Yu W, Fang Q, Liu L, Zhuang K, et al. A novel replication-competent vaccinia vector MVTT is superior to MVA for inducing high levels ofneutralizing antibody via mucosal vaccination. PLoS ONE 2009;4:e4180.

17] Kibler KV, Gomez CE, Perdiguero B, Wong S, Huynh T, Holechek S, et al.

Improved NYVAC-based vaccine vectors. PLoS ONE 2011;6:e25674.

18] Garcia-Arriaza J, Najera JL, Gomez CE, Tewabe N, Sorzano COS, Calandra T,et al. A candidate HIV/AIDS vaccine (MVA-B) lacking vaccinia virus geneC6L enhances memory HIV-1-speicific T-cell responses. PLoS ONE 2011;6:e24244.

[

(2012) 5733– 5739 5739

19] Kudela P, Koller VJ, Lubitz W. Bacterial ghosts (BGs)—advanced antigen anddrug delivery system. Vaccine 2010;28:5760–7.

20] Ebensen T, Pauker S, Link C, Kudela P, de Domenico C, Lubitz W, et al. Bac-terial ghosts are an efficient delivery system for DNA vaccines. J Immunol2004;172:6858–65.

21] Kudela P, Paukner S, Mayr UB, Cholujova D, Schwarczova Z, Sedlak J, et al. Bac-terial ghosts as a novel advanced targeting vehicles for DNA delivery to thehuman monocyte-derived dendritic cells. J Immunother 2005;28:136–43.

22] Paukner S, Kudela P, Kohl G, Schlapp T, Friedrichs S, Lubitz W. DNA-loadedbacterial ghosts differently mediate reporter gene transfer and expression inmacrophages. Mol Ther 2005;11:215–23.

23] Barrell BG, Air GM, Hutchison III CA. Overlapping genes in bacteriophagePhiX174. Nature 1976;264:34–41.

24] Jechlinger W, Szostak MP, Witte A, Lubitz W. Altered temperature inductionsensitivity of the lambda PR/cI857 system for controlled gene E expression inEscherichia coli. FEMS Microbiol Lett 1999;173:347–52.

25] Darji A, Guzman CA, Gerstal B, Wachholz P, Timmis KN, Wehland J, et al.Oral somatic transgene vaccination using attenuated S. typhimurium. Cell1997;91:765–75.

26] Vecino WH, Morin PM, Agha R, Jacobs Jr WR, Fennelly GJ. Mucosal DNA vaccina-tion with highly attenuated Shigella is superior to attenuated Salmonella andcomparable to intramuscular DNA vaccination for T cells against HIV. ImmunolLett 2002;82:197–204.

27] Baumler AJ, Tsolis RM, Heffron F. The lpf fimbrial operon mediates adhesionof Salmonella typhimurium to murine Peyer’s patches. Proc Natl Acad Sci USA1996;93:279–83.

28] Jones BD, Ghori N, Falkow S. Salmonella typhimurium initiates murine infec-tion by penetrating and destroying the specialized epithelial M cells of Peyer’spatches. J Exp Med 1994;180:15–23.

29] Kou Z, Quinn M, Chen H, Rodrigo SWW, Rose RC, Schlesinger JJ, et al. Mono-cytes, but not T or B Cells, are the principal target cells for dengue virus(DV) infection among human peripheral blood mononuclear cells. J Med Virol2008;80:134–46.

30] Takeuchi O, Akira S. Pattern recognition receptors and inflammation. Cell2010;140:805–20.

31] Treanor JJ, Taylor DN, Tussey L, Hay C, Nolan C, Fitzgerald T, et al. Safety andimmunogenicity of a recombinant hemagglutinin influenza-flagellin fusionvaccine (VAX125) in healthy young adults. Vaccine 2010;28:8268–74.

32] Yu B, Yang M, Wong HYB, Watt RM, Song E, Zheng BJ, et al. A method to generaterecombinant Salmonella typhi Ty21a strains expressing multiple heterologous

genes using an improved recombineering strategy. Appl Microbiol Biotechnol2011;91:177–88.

33] Haslberger AG, Kohl G, Felnerova D, Mayr UB, Furst-Ladani S, Lubitz W. Acti-vation, stimulation and uptake of bacteria ghosts in antigen presenting cells. JBiotechnol 2000;83:57–66.