document

NATURE BIOTECHNOLOGY VOLUME 16 NOVEMBER 1998 1045

Adenovirus gene therapy is limited by the induction of an immuneresponse to the virus or the gene-therapy protein product1–4. A spe-cific T-cell response to the adenovirus results in the failure of read-ministration of the gene therapy5,6. Previous attempts to reduce theT-cell response to the adenovirus during gene therapy, includingblockade of major histocompatibility class I and II antigens, reduc-tion in the antigenicity of the adenovirus, and prevention of cos-timulation of T cells, either have not fully eliminated the responseor invoke general immunosuppression1,6–10.

An ideal strategy for elimination of the immune response would beinduction of peripheral T-cell tolerance that is specific for the adenovi-ral vector. T-cell activity is modulated by several mechanisms thatinduce apoptosis of the participating cells to downregulate theresponse. Clonal deletion of antigen-specific T cells, which is mediatedby apoptosis, is an important mechanism in the maintenance ofperipheral T-cell tolerance11–14. Activation-induced cell death in T cells,in which apoptosis of the T cells is mediated by upregulation of Fasand Fas ligand, also contributes to downregulation of the T-cellresponse15–18. Antigen presenting cells (APCs), which process and pre-sent antigen to the T cells, express elevated levels of Fas ligand afteractivation. The outcome of the interaction of the APCs with the T cellscan be either highly immunogenic or tolerogenic depending on theinteractions of costimulatory molecules, cytokines, and the expressionof Fas ligand. The increased expression of Fas ligand on activated APCsresults in apoptosis of the T cells during antigen presentation—aprocess that is thought to be a critical means of down-modulating T-cell responses19,20. Activation-induced cell death of T cells plays animportant role in the maintenance of tolerance to adenovirus21.

Fas ligand can create immune-privileged sites and prevent graftrejection by inducing apoptosis in T cells entering the site11,12,14. Thesefindings suggest a strategy in which introduction of APCs express-ing high levels of Fas ligand together with a specific antigen mightinduce specific, systemic tolerance to the antigen. We show thatAPCs, which express Fas ligand and processed adenovirus antigens,can directly induce apoptosis of Fas-positive T cells resulting in ade-novirus-specific T-cell tolerance. High levels of Fas ligand and aden-ovirus antigens were induced in APCs by coinfection withAdLoxpfasL and AxCANcre. Pretreatment of recipient mice with theadenovirus-transfected APCs that produce Fas ligand resulted ininduction of T-cell tolerance to the adenovirus. The decreased T-cellresponse to the adenovirus was demonstrated by decreased cytokineproduction, a decreased cytotoxic T-cell response, inhibition ofclonal expansion of CD3+ T cells, and prolonged expression of thelacZ transgene after administration of AdCMVlacZ. Induction of T-cell tolerance to adenovirus required production of Fas ligand bythe APCs and did not occur with adenovirus-transfected, controlAPCs. T-cell tolerance also required production of Fas by the T cellsof recipient mice, as lpr/lpr mice could not be tolerized. The T-celltolerance was antigen-specific as there was a normal T-cell responseto murine cytomegalovirus (MCMV) in tolerized mice.

ResultsCoinfection with AdLoxpfasL and AxCANCre results in high lev-els of Fas ligand production and induction of apoptosis in A20target cells. A modified AdLoxpfasL adenovirus yielded high titerviral production in 293 cells22 and allowed control of Fas ligand

RESEARCH

Induction of specific T-cell tolerance byadenovirus-transfected, Fas

ligand–producing antigen presenting cellsHuang-Ge Zhang1, Di Liu1, Yuji Heike2, PingAr Yang1, Zheng Wang1, Xiaoyun Wang1, David T. Curiel2,

Tong Zhou1, and John D. Mountz1,3*1 The University of Alabama at Birmingham, Department of Medicine, Division of Clinical Immunology and Rheumatology, Birmingham, AL 35294.

2 Gene Therapy Program, University of Alabama at Birmingham, Birmingham, AL 35294. 3Veterans Administration Medical Center, Birmingham, AL 35233.*Corresponding author (e-mail: [email protected]).

Received 8 April 1998; accepted 25 September 1998

A major problem associated with adenovirus gene therapy is the T cell-mediated immune response,which is elicited by inoculation of the adenovirus vector and leads to rapid clearance of the virus and lossof transgene expression. In this study, the immune response to adenovirus was prevented by induction ofspecific T-cell tolerance by pretreatment with adenovirus-infected antigen-presenting cells (APC) thatexpress Fas ligand. Compared with control-treated mice, the tolerized mice showed prolonged expres-sion of lacZ upon administration of AdCMVlacZ 1 week after tolerance induction. In contrast to the con-trol mice, the tolerized mice did not display proliferation of CD3+ T cells in the spleen in response toAdCMVlacZ. Tolerance induction also was indicated by the lower production of interferon-g and inter-leukin-2 by peripheral T cells isolated from AdCMVlacZ-challenged tolerized mice than by AdCMVlacZ-challenged control-treated mice. The T-cell tolerance was specific for the adenovirus as the T-cellresponses to irrelative murine cytomegalovirus remained unimpaired. Our results indicate that aden-ovirus-specific T-cell tolerance can be induced by APCs that coexpress Fas ligand and adenovirus anti-gens. We propose that this new strategy can be used to induce tolerance to adenovirus vector gene ther-apy with resultant prolonged expression of the transgene.

Keywords: gene therapy, apoptosis, immune response

1998 Nature America Inc. • http://biotech.nature.com

1998

Nat

ure

Am

eric

a In

c. •

http

://bi

otec

h.na

ture

.com

production as Fas ligand is not produced in the absence of coinfec-tion with AxCANcre. This strategy was used to induce high levels ofFas ligand production in an APC cell line derived from Fas-mutantB6-lpr/lpr mice. This modified APC cell line could induce apopto-sis of A20 target cells (Fig. 1). The lytic activity of the APC trans-fected with AdLoxpfasL+AxCANCre (APC-AdfasL) against A20target cells was 10-fold higher than that of APCs transfected byelectroporation with a pcDNA3-fasL expression vector, and 100-fold higher than lipopolysaccharide (1 µg/ml)-activated APCs.These results demonstrate that coinfection of both AdLoxpfasL andAxCANCre into the Fas-deficient APCs results in high levels offunctional Fas ligand production.

Prolonged production of LacZ in the liver after APC-AdfasLtherapy. Expression of adenovirus gene therapy in the liver is limiteddue to an acute inflammatory response and a chronic cytotoxic T-cell response23,24. To determine whether treatment with APC trans-fected with AdLoxpfasL+AxCANcre (APC-AdfasL) leads to prolon-gation of transgene expression, APC-AdfasL–treated and APC trans-fected with AdLoxpfasL+AdCMVluc (APC-AdControl)–treatedmice were inoculated with AdCMVlacZ (1×1010 pfu). The levels oflacZ gene expression in the liver decreased rapidly in mice treatedwith the APC-AdControl (Fig. 2). In contrast, in mice treated withAPC-AdfasL, the levels of lacZ gene expression did not undergo adecrease and were sustained for at least 50 days after gene delivery.These results indicate that pretreatment with APC-AdfasL signifi-cantly prolongs AdCMVlacZ transgene expression.

Decreased T-cell expansion in APC-AdfasL–treated mice. Wild-type B6 mice were treated with APCs, APC-AdControl, or APC-AdfasL every 3 days until five doses were given, and then challengedintravenously with AdCMVlacZ (1×1010 pfu). Three days later,frozen sections of spleen were analyzed by immunohistochemicalstaining with anti-CD3. Compared with control mice receivingAPCs (Fig. 3A), there was a clonal expansion of CD3+ T cells in thespleens of mice treated with APC-AdControl after challenge (Fig.3B), which was not observed in the spleens of APC-AdfasL tolerizedmice (Fig. 3C). These results suggest that APC-AdfasL induces toler-ance of splenic T cells to subsequent stimulation.

Adenovirus-transfected APCs migrate to the spleen. To deter-mine whether adenovirus-transfected APCs migrated to thespleen, wild-type B6 mice were treated with a single dose of 1 x 106

APC-AdCMVGFP. Two days later, mice were killed and the spleenand liver were analyzed for green fluorescent protein–positive(GFP+) cells by fluorescence microscopy. There were high numbersof GFP+ cells located in the pericortical region of the spleen (Fig.4A). In contrast, few APCs migrated to the liver (Fig. 4B).

To determine whether treatment with APC-fasL resulted inapoptosis of liver cells, wild-type B6 mice were treated with a singledose of APC-fasL (106 cells) or directly coinjected with AdLoxpfasL+AxCANcre (109 pfu). There was no obvious apoptosis of the liver24 h after APC-fasL treatment as determined by Hoechst staining(Fig. 4C). In contrast, intravenous injection of AdLoxpfasL+AxCANcre into B6-+/+ mice resulted in extensive apoptosis of liver12 h later (Fig. 4D).

Decreased cytotoxic response of T cells to AdCMVGFP-trans-fected target cells after tolerance with Fas ligand–producing APCs.Mice were tolerized in vivo with APC-AdfasL or APC-AdControland then stimulated with AdCMVlacZ. Seven days later, splenic Tcells were purified and their ability to kill AdCMVGFP-transfectedAPC target cells was determined at different effector/target (E/T)ratios. After stimulation with AdCMVlacZ, T cells from mice thathad been tolerized with APC-AdControl demonstrated high in vitrocytotoxic activity against APC transfected with AdCMVGFP (Fig.5). In contrast, mice that had been tolerized with APC-AdfasL andsubsequently immunized with AdCMVlacZ exhibited low cytotoxicactivity against the AdCMVGFP-transfected APCs. Low cytotoxic

T-cell response by APC-AdfasL tolerized mice at day 7 after chal-lenge with AdCMVlacZ correlates with the low cytotoxic responsein vivo and prolonged production of LacZ in the liver.

T-cell tolerance demonstrated by decreased interferon-g andinterleukin-2 production. To further evaluate the effect of pretreat-ment with APC-AdfasL on T-cell activity and tolerance induction,the production of interleukin-2 (IL-2) and interferon-g (IFN-g) bythe splenic T cells was determined. Thirty days after treatment witheither APC-AdfasL or APC-AdControl, mice were killed and thespleen cells stimulated for 48 h with either APC or APC transfectedwith AdCMVlacZ. T cells from APC-AdControl–treated wild-typemice produced high levels of IL-2 (Fig. 6A) and IFN-g (Fig. 6B) inresponse to adenovirus-transfected APCs but not untransfectedAPCs. In contrast, T cells from APC-AdfasL–treated wild-type miceproduced low levels of IL-2 (Fig. 6A) and IFN-g (Fig. 6B) in responseto adenovirus-transfected APCs. These results indicate that APC-AdfasL results in long-term, systemic tolerization of T cells in vivoand induces nonresponsiveness in these T cells upon challenge 4weeks after treatment.

1046 NATURE BIOTECHNOLOGY VOLUME 16 NOVEMBER 1998

RESEARCH

Figure 1. FasL production by APCs. APCs from B6-lpr/lpr mice werecotransfected with AdLoxpfasL + AxCANcre (p), transfected withpcDNA3fasL (P) or stimulated with lipopolysaccharide (l). FasLproduction was determined by APC-induced lysis of a 51Cr-labeledA20 cell line.

Figure 2. AdlacZ transgene expression after APC-AdfasL followed byAdCMVlacZ. Wild-type B6 mice were treated with 1×106 of APCscotransfected with AdLoxpfasL+AxCANcre (APC-AdfasL) (L) orAPCs cotransfected with AdLoxpfasL+AdCMVluc (APC-AdControl)(l) or phosphate-buffered saline (L) every 3 days until five doseswere given. After 7 days, mice were inoculated intravenously with 1×1010 pfu AdCMVlacZ and b-Gal was determined up to 50 days later.The error bars indicate the mean ± standard error of the mean (SEM)for three mice analyzed separately in triplicate assay.


1998

Nat

ure

Am

eric

a In

c. •

http

://bi

otec

h.na

ture

.com


To determine whether induction of tolerance by the APC-AdfasLis mediated by Fas ligand interaction with Fas, the activity of spleencells from Fas-deficient B6-lpr/lpr mice that had been tolerized withAPC-AdfasL was evaluated. On challenge with APCs transfectedwith AdCMVlacZ, the spleen cells from tolerized B6-lpr/lpr miceproduced high levels of IL-2 and IFN-γ at 48 h (Fig. 6C and D).These results indicate that induction of tolerance by APC-AdfasL isdependent on functional expression of Fas in the recipients.

APC-AdfasL induces specific tolerance to adenovirus. Todetermine whether the T-cell tolerance induced by APC-AdfasLwas specific for adenoviral vector rather than general immune sup-

pression of the response to viral infection, the T-cell response ofAPC-AdfasL and APC-AdControl tolerized mice to MCMV infec-tion was evaluated. B6-+/+ mice were treated with APC-AdfasL asdescribed above for induction of tolerance, and then challenged 7days later with either adenovirus or MCMV (Fig 7). Although therewas reduction in the T-cell response to adenoviral vector, the T-cell

RESEARCH

Figure 3. Splenic T cells after APC-AdfasLfollowed by AdCMVLacZ. Wild-type C57BL/6 micewere treated with 1×106 of (A) APC, (B) APC-AdControl, (C) APC-AdfasL every 3 days until fivedoses were given and challenged with AdCMVlacZ(1×1010 pfu). Three days after AdCMVlacZchallenge, mice were killed and CD3+ T cells in thespleen were identified. The results arerepresentative of five mice/group.

A B C

100 µm

Figure 4. Migration of APC-AdCMVGFP in vivo. Wild-type C57BL/6 mice were treated with APC-AdCMVGFP, and analyzed 48 h later for GFP+

cells by fluorescence microscopy of (A) spleen and (B) liver. Apoptosis of hepatocytes in wild-type C57BL/6 mice was analyzed 24 h afterintravenous administration of (C) 1×106 APC-AdfasL and (D) AdLoxpfasL+AxCANcre (1×109 pfu each).

A B C D

100 µm 25 µm

Figure 5. Cytotoxic T-cell response to APC+Ad after APC-AdfasL.Wild-type C57BL/6 mice were treated with 1×106 APC-AdfasL (P) orAPC-AdControl (p) every 3 days until five doses were given and werechallenged with AdCMVlacZ 7 days later. Mice were killed and the T-cell cytotoxic response against APCs transfected with AdCMVGFPwas determined. The error bars indicate the mean±SEM for threemice analyzed separately in triplicate assays.

Figure 6. Cytokine production by T cells stimulated with APC+Adafter APC-AdfasL. (A and B) Wild-type B6 and (C and D) B6-lpr/lprmice were treated with APC-AdfasL or APC-AdControl cells (1×106).Thirty days later, the mice were sacrificed and spleen cells werestimulated for 48 h in vitro with adenovirus transfected (L) oruntransfected (l) irradiated APCs. (A and C) IL-2 and (B and D) IFN-gin the supernatant was determined by ELISA.

A B

C D


1998

Nat

ure

Am

eric

a In

c. •

http

://bi

otec

h.na

ture

.com

response to MCMV was not impaired as demonstrated by the com-parable levels of IL-2 produced by the T cells from both APC-AdControl and APC-AdfasL–treated mice. This result indicatesthat inhibition of the T-cell response in APC-AdfasL tolerized miceis specific for adenoviral vector.

DiscussionProduction of Fas ligand can be used to induce specific tolerance byapoptosis and clonal deletion of antigen-reactive T cells15,16. Fas lig-and can create immune-privileged sites and prevent graft rejectionby inducing apoptosis of the T cells11,25, but Fas ligand can also causelocal tissue damage26–32. An AdLoxpfasL vector has been produced22,but direct administration of this Ad/fasL was toxic to the liver andislet cells. We have shown that APCs producing Fas ligand can safe-ly induce apoptosis of T cells that produce Fas, resulting in antigen-specific T-cell tolerance without significant liver toxicity.

AdLoxpfasL coinfection with AxCANcre results in high levels ofFas ligand production in almost 100% of transfected APCs. Onereason for this high efficiency of infection is that both viruses canbe grown to high titers in the 293 cells as there is no Fas ligand pro-duction by AdLoxpfasL, which requires coinfection with anAxCANcre virus22. Second, the two-virus system was used to infectan APC cell line derived from Fas mutant C57BL/6-lpr/lpr mice.Therefore, these APCs can produce high levels of Fas ligand with-out undergoing autocrine suicide. For applications to human genetherapy, Fas production by APCs could be down-modulated usingan sFv that recognizes Fas and inhibits cell-surface expression. Thismethod has been described for trapping expressed peptides in theendosome33,34. A second approach that we have successfully used toprevent Fas ligand apoptosis in normal murine APCs is coinfectionwith AdCrmA+AdLoxpfasL+AxCANcre. There is a high efficiencyof infection of these APCs with adenovirus. This is in contrast tolow efficiency transfection of DNA into APCs using lipofectin(1–5%) or electroporation (8–10%).

The mechanism of tolerance induction by APC-AdfasL ismigration to the spleen and prevention of a subsequent CD3+ T-cellexpansion following administration of AdCMVlacZ. This preventssubsequent generation of cytotoxic T cells that migrate to the liverand mediate an immune response after adenovirus administration.The absence of cytotoxic T cells at 7 days postinfection withAdCMVlacZ correlated with a prolonged production of LacZ in theliver of tolerized mice compared with nontolerized mice.Adenovirus expression of Fas ligand within an APC may be usefulas pretreatment for systemic tolerization against administration ofan adenovirus/gene therapy product.

Tolerance induction by APCs transfected with adenovirus andexpressing high levels of Fas ligand is specific for adenovirus, butnot MCMV. Other methods for induction of tolerance to, orimmunosuppression of, adenovirus gene therapy are associatedwith a more generalized immunosuppressed state, which would beundesirable for long-term gene therapy use. The APC-AdfasLtolerizing technique abrogates the ability of the recipient torespond to the tolerizing virus used to infect the APC, but does notaffect the response to another virus.

Experimental protocolAnimals. Six- to 10-week-old, female C57BL/6-+/+ and C57BL/6-lpr/lprmice were obtained from the Jackson Laboratory (Bar Harbor, MA). Micewere maintained in pathogen-free conditions.

Construction of Fas ligand expression adenovirus vector. Briefly, a 10.4 kbshuttle vector containing the fragment of adenovirus from 0 map unit to 1 mapunit followed by the 1.6 kb chicken b-actin promoter plus CMV enhancer. Thiswas followed by two Loxp sites separated by a Neo resistant gene plus a 0.3 kbbovine growth hormone poly A tail. The full-length 0.9 kb FasL was cloneddownstream from the bovine growth hormone poly A tail, which was followedby an SV40 polyA tail and by the 9.8–16.1 map units of adenovirus22.

MCMV virus. MCMV Virus Smith strain was obtained from the AmericanType Culture Collection (Rockville, MD). The viruses were titrated as dupli-cates in log10 dilutions on subconfluent primary murine embryo fibroblasts in12-well plates. Seven days later, monolayers were stained with neutral red andthe number of plaques counted. The supernatant was dispensed intoaliquots, which were stored at -80°C and used as the MCMV stock virus pool(3 x 107 pfu/ml).

Infection of APCs for Fas ligand expression22. Murine B6-lpr/lpr APCswere transfected with either AdLoxpfasL+AxCANcre (APC-AdfasL) orAdLoxpfasL+AdCMVluc (APC-AdControl) at 5 pfu/cell of each virus for 1 hat 37°C, and the transfected cells incubated at 37°C for a further 24 h.Presence of murine Fas ligand and adenoviral antigens on the surface of B6-lpr/lpr APCs was identified using an indirect immunofluorescent assay, andthe functional ability of Fas ligand in mediating killing was evaluated using a51Cr-release assay as described22.

Analysis of Fas ligand production by APCs transfected with AdLoxpfasL+AxCANcre. Fas ligand cytotoxic activity was assayed as described22. Fas lig-and production was determined by the ability of the transfected APCs toinduce apoptosis of a 51Cr labeled, Fas-sensitive cell line, A20. Target cells (1×106), which are sensitive to cytotoxic lysis, were incubated with 20 µCi of[51Cr]-sodium chromate in 100 ml of RPMI-1640 containing 10% fetal calfserum at 37°C for 1 h. After washing with medium, these cells were used astarget cells. Effector cells were prepared from B6-lpr/lpr APCs transfectedwith AdLoxpfasL+AxCANcre as described above. These effector cells werethen incubated with [51Cr]-labeled target cells (1×104) at different E/T ratiosin a total volume of 200 µl of the medium. Release of 51Cr into the supernatantwas assessed 6 h later using a b-counter. The percentage of specific 51Crrelease was calculated as follows:

(experimental 51Cr release – spontaneous 51Cr release)% specific lysis =

(maximum 51Cr release – spontaneous 51Cr release)

The spontaneous release of 51Cr using these assays has routinely been 8–12%of the maximum release. Analysis of adenovirus-specific cytotoxic T-cellactivity using AdCMVGFP-transfected target cells. The adenovirus shuttlevector construct was produced by cloning the enhanced GFP gene (Clontech,Palo Alto, CA) into the HindIII-XbaI site of pCA13 (Microbix, Canada). Thiswas cotransfected with pJM17 to produce recombinant AdCMVGFP.AdCMVGFP was plaque purified by three rounds of selection in order toinfect APCs for use as target cells in an analysis of cytotoxic effector T cellsfrom mice treated with APC-AdfasL and APC-AdControl. Effector T cellswere prepared from spleen of AdCMVlacZ-immunized and nonimmunizedmice. These effector cells were then incubated with AdCMVGFP-transfectedtarget cells (1×105) at different E/T ratios in round-bottom microtiter platesin a total volume of 200 µl of the medium for 48 h, and GFP-positive APCwere sorted using fluorescence-activated cell sorting analysis. The percentageof specific cytotoxicity was calculated as follows:

(100% – experimental % GFP+ – spontaneous % GFP+)% specific lysis =

(100% – maximum % GFP+ – spontaneous % GFP+)

1048 NATURE BIOTECHNOLOGY VOLUME 16 NOVEMBER 1998

RESEARCH

Figure 7. IL-2 production by T cells stimulated with APC+MCMV afterAPC-AdfasL. Wild-type mice were treated with either APC-AdfasL orAPC-AdControl. Seven days later, mice were challenged in vivo witheither AdCMVlacZ or MCMV. Seven days later splenic T cells werestimulated in vitro with (l) APCs, (L) AdCMVlacZ-transfected APCs,or (L) MCMV-infected APCs. IL-2 production in the supernatants wasdetermined by ELISA.


1998

Nat

ure

Am

eric

a In

c. •

http

://bi

otec

h.na

ture

.com


Administration of APC-AdfasL for induction of tolerance. Ten-week-oldC57BL/6-+/+ mice were injected intravenously with 1×106 of the APCs cotrans-fected with AdLoxpfasL+AxCANcre (APC-AdfasL) or APCs cotransfected withAdLoxpfasL+AdCMVluc (APC-AdControl) or with phosphate-buffered salineevery 3 days until five doses were given. On day 7 after the final injection, micewere challenged with AdCMVlacZ and the T-cell cytotoxic response againstAPCs transfected with adenovirus was determined 1 week after challenge.

Analysis of the immune response to adenovirus and MCMV after toler-ance induction. One week after tolerance induction, mice were treated withAdCMVlacZ (1×1010 pfu given intravenously) or MCMV (1×105 pfu givenintravenously). After an additional 7 days, purified splenic T cells were stimu-lated in vitro with APCs alone, or APCs that had been incubated either withMCMV or AdCMVlacZ. After 48 h the supernate was collected and analyzedfor IL-2 expression.

Quantitation of b-galactosidase expression in liver. b-galactosidase (b-gal) activity was determined as described35. Freshly isolated liver tissue washomogenized for 10 s in a tissumizer in 1 ml of b-gal buffer (Tropix, Bedford,MA). The homogenate was centrifuged at 12,500 G for 10 min at 4°C, and thesupernatant was heated for 60 min at 48°C to inactivate the endogenouseukaryotic b-galactosidase activity. The sample was then centrifuged at12,500 G for 5 min, and 10 µl of the supernatant was assayed for b-galactosi-dase activity using the Galacto-light (Tropix) chemiluminescent reporterassay. The reaction was carried out for 10 min at room temperature and b-galactosidase activity was assayed using a luminomiter (Monolight 500;Wallac, Gaithersburg, MD). The protein concentration was determined bythe Bradford assay (Bio-Rad, Hercules, CA). The activity is expressed as therelative light units/min/mg of total protein in the liver.

Cytokine production in vitro in response to APCs transfected with aden-ovirus. B6-lpr/lpr APCs were transfected with AdCMVlacZ (10 pfu/cell) for 1h in 1 ml of media and then diluted by addition of 10 ml of RPMI-1640 sup-plemented with 10% fetal bovine serum and culture continued at 37°C for 24h. Before use as stimulator cells, the APCs were γ-irradiated, and 1×105 APCswere mixed with T cells isolated from the spleen of tolerized mice. The mixedcells were incubated in 96-well plates for 2 days at 37°C. The supernatantswere collected and induction of IL-2 and IFN-g was determined using anELISA assay kit (R & D Systems, Minneapolis, MN).

Histopathological examination of tissue sections. Animals were killed bycervical dislocation. Organs were removed and fixed in neutral 10% forma-lin/phosphate-buffered saline for 24 h. Tissues were then embedded in paraffinblocks, sectioned (10 µm thickness), and stained with hematoxylin and eosin.Paraffin-embedded tissue sections were deparaffinized and treated with 3%H2O2 at room temperature for 15 min. After washing three times with neutralphosphate-buffered saline, tissues were stained with an antibody against anti-CD3 (Dako Corporation, Carpinteria, CA) following standard avidin-biotinconjugate immunohistochemical techniques according to the manufacturer’smanual (Dako). A peroxidase-conjugated secondary antibody was then appliedto the sections at room temperature for 2 h. Positive staining was visualizedusing diaminobenzidine substrate (Dako). Apoptosis analysis of the liver byHoechst staining was carried out as previously described22.

Analysis of APC-AdCMVGFP localization. APC-AdCMVGFP was producedas described above. B6-+/+ mice (five mice/group) were injected intravenouslywith 106 APC-AdCMVGFP or control APCs. Mice were killed 48 h later andfrozen sections of the liver and spleen were analyzed by fluorescence microscopy.

Statistical analysis. The two-tailed Student’s t-test was used for statisticalanalysis when two different groups of samples were compared. The one-fac-tor analysis of variance test was used when more than two groups of sampleswere compared. A p value of less than 0.05 was considered significant.

AcknowledgmentsThis work was supported in part by VA Career Development and Merit ReviewAward, National Institutes of Health Grants NO1-AR-6-2224 and RO1-AR-42547 from NIAMS to J.D.M., and by a grant from Sankyo, Inc. T.Z. is the recip-ient of an Arthritis Foundation Investigator Award.

1. Yang, Y. and Wilson, J.M. 1995. Clearance of adenovirus-infected hepatocytesby MHC class I-restricted CD4+ CTLs in vivo. J. Immunol. 155:2564–2570.

2. Christ. M., Lusky, M., Stoeckel, F., Dreyer, D., Dieterle, A., Michou, A.I. et al. 1997.Gene therapy with recombinant adenovirus vectors: evaluation of the hostimmune response. Immunol Lett. 57:19–25.

3. Yang, Y., Li, Q., Ertl, H.C., and Wilson, J.M. 1995. Cellular and humoral immuneresponses to viral antigens create barriers to lung-directed gene therapy withrecombinant adenoviruses. J. Virol. 69:2004–2015.

4. Gilgenkrantz, H., Duboc, D., Juillard, V., Couton, D., Pavirani, A., Guillet, J.G. et al.

1995. Transient expression of genes transferred in vivo into heart using first-genera-tion adenoviral vectors: role of the immune response. Hum. Gene Ther. 6:1265–1274.

5. Yang, Y., Nunes, F.A., Berencsi, K., Furth, E.E., Gonczol, E., and Wilson, J.M.1994. Cellular immunity to viral antigens limits E1-deleted adenoviruses for genetherapy. Proc. Natl. Acad. Sci. USA 91:4407–4411.

6. Juillard, V., Villefroy, P., Godfrin, D., Pavirani, A., Venet, A., and Guillet, J.G. 1995.Long-term humoral and cellular immunity induced by a single immunization withreplication-defective adenovirus recombinant vector. Eur. J. Immunol. 25:3467–3473.

7. Yang, Y., Su, Q., Grewal, I.S., Schilz, R., Flavell, R.A., and Wilson, J.M. 1996.Transient subversion of CD40 ligand function diminishes immune responses toadenovirus vectors in mouse liver and lung tissues. J. Virol. 70:6370–6377.

8. Schowalter, D.B., Meuse, L., Wilson, C.B., Linsley, P.S., and Kay, MA. 1997.Constitutive expression of murine CTLA4Ig from a recombinant adenovirus vec-tor results in prolonged transgene expression. Gene Ther. 4:853–860.

9. Qin, L., Ding, Y., Pahud, D.R., Robson, N.D., Shaked, A., and Bromberg, J.S. 1997.Adenovirus-mediated gene transfer of viral interleukin-10 inhibits the immuneresponse to both alloantigen and adenoviral antigen. Hum. Gene Ther. 8:1365–1374.

10. Zsengeller, Z.K., Boivin, G.P., Sawchuk, S.S., Trapnell, B.C., Whitsett, J.A., and Hirsch,R. 1997. Anti-T cell receptor antibody prolongs transgene expression and reduces lunginflammation after adenovirus-mediated gene transfer. Hum. Gene Ther. 8:935–941.

11. Bellgrau, D., Gold, D., Selawry, H., Moore, J., Franzusoff, A., and Duke, R.C.1995. A role for CD95 ligand in preventing graft rejection. Nature 377:630–632.

12. French, L.E., Hahne, M., Viard, I., Radlgruber, G., Zanone, R., Becker, K. et al.1996. Fas and Fas ligand in embryos and adult mice: ligand expression in sever-al immune-privileged tissues and coexpression in adult tissues characterized byapoptotic cell turnover. J. Cell Biol. 199:335–343.

13. Lee, J., Richburg, J.H., Younkin, S.C., and Boekelheide, K. The Fas system is a keyregulator of germ cell apoptosis in the testis. 1997. Endocrinology 138:2081–2088.

14. Griffith, T.S., Yu, X., Herndon, J.M., Green, D.R., and Ferguson, T.A. 1996. CD95-induced apoptosis of lymphocytes in an immune privileged site induces immuno-logical tolerance. Immunity. 5:7–16.

15. Watanabe-Fukunaga, R., Brannan, C.I., Copeland, N.G., Jenkins, N.A., andNagata, S. 1992. Lymphoproliferation disorder in mice explained by defects inFas antigen that mediates apoptosis. Nature. 356:314–317.

16. Zhou, T, Bluethmann, H., Zhang, J., Edwards, C.K., and Mountz, J.D. 1992.Defective maintenance of T cell tolerance to a superantigen in MRL-lpr/lpr mice.J. Exp. Med. 176:1063–1072.

17. Suda, T., Takahashi, T., Golstein, P., and Nagata, S. 1993. Molecular cloning andexpression of the Fas ligand, a novel member of the tumor necrosis factor family.Cell 75:1169–1178.

18. Wu, J., Zhou, T., Zhang, J., He, J., Gause, W.C., and Mountz, J.D. 1994.Correction of accelerated autoimmune disease by early replacement of themutated lpr gene with the normal Fas apoptosis gene in the T cells of transgenicMRL-lpr/lpr mice. Proc. Natl. Acad. Sci. USA 91:2344–2348.

19. Suda, T., Okazaki, T., Naito, Y., Yokota, T., Arai, N., Ozaki, S. et al 1995. Expressionof the Fas ligand in cells of T cell lineage. J. Immunol. 154:3806–3813.

20. Cheng, J., Liu, C., Yang, P., Zhou, T., and Mountz, J.D. 1997. Increased lympho-cyte apoptosis in Fas ligand transgenic mice. J. Immunol. 159:674–684.

21. DeMatteo, R.P., Raper, S.E., Ahn, M., Fisher, K.J., Burke, C., Radu, A. et al. 1995.Gene transfer to the thymus. A means of abrogating the immune response torecombinant adenovirus. Ann. Surg. 222:229–239.

22. Zhang, H.G., Bilbao, G., Zhou, T., Contreras, J.L., Gomez-Navarro, J., Feng, M.et al. 1998. Application of a Fas ligand encoding a recombinant adenovirus vec-tor for prolongation of transgene expression. J. Virol. 72:2483–2490.

23. Zsengeller, Z.K., Wert, S.E., Hull, W.M., Hu, X., Yei, S., Trapnell, B.C. et al. 1995.Persistence of replication-deficient adenovirus-mediated gene transfer in lungsof immune-deficient (nu/nu) mice. Hum. Gene Ther. 6:457–467.

24. Zhang, H.-G., Zhou, T., Yang, P., Edwards, C.K., Curiel, D.T., and Mountz, J.D.1998. Inhibition of tumor necrosis factor alpha decreases inflammation and pro-longs adenovirus gene expression in lung and liver. Hum. Gene Ther. 9:1875–1884.

25. Lau, H.T., Yu, M., Fontana, A., and Stoeckert, C.J. Jr. Prevention of islet allograft rejec-tion with engineered myoblasts expressing FasL in mice. 1995. Science 273:109–112.

26. Griffith, T.S., Brunner, T., Fletcher, S.M., Green, D.R., and Ferguson, T.A. 1995. Fas lig-and-induced apoptosis as a mechanism of immune privilege. Science 270:1189–1192.

27. Kondo, T., Suda, T., Fukuyama, H., Adachi, M., and Nagata, S. 1997. Essentialroles of the Fas ligand in the development of hepatitis. Nat. Med. 3:409–419.

28. Giordano, C., Stassi, G., De Maria, R., Todaro, M., Richlusa, P., Papoff, G. et al.1997. Potential involvement of Fas and its ligand in the pathogenesis ofHashimoto’s thyroiditis. Science 275:960–963.

29. Larsen, C.P., Alexander, D.Z., Hendrix, R., Ritchie, S.C., and Pearson, T.C. 1995.Fas-mediated cytotoxicity. Transplantation 60:221–224.

30. Muruve, D.A., Nicolson, A.G., Manfro, R.C., Strom, T.B., Sukhatme, V.P., andLibermann, T.A. 1997. Adenovirus-mediated expression of Fas ligand induceshepatic apoptosis after systemic administration and apoptosis of ex vivo-infect-ed pancreatic islet allografts and isografts. Hum. Gene Ther. 8:955–963.

31. Muruve, D.A., Manfro, R.C., Strom, T.B., and Libermann, T.A. 1997. Ex vivo ade-novirus-mediated gene delivery leads to long-term expression in pancreatic islettransplants. Transplantation 64:542–546.

32. Sigalla J., David A., Anegon I., Fiche M., Huvelin JM., Boeffard F. et al. 1997.Adenovirus-mediated gene transfer into isolated mouse adult pancreatic islets:normal beta-cell function despite induction of an anti-adenovirus immuneresponse. Hum. Gene Ther. 8:1625–1634.

33. Piche, A., Grim, J., Rancourt, C., Gomez-Navarro, J., Reed, J.C., and Curiel, D.T.1998. Modulation of Bcl-2 protein levels by an intracellular anti-Bcl-2 single-chain antibody increases drug-induced cytotoxicity in the breast cancer cell lineMCF-7. Cancer Res. 58:2134–2140.

34. Grim, J., Deshane, J., Siegal, G.P., Alvarez, R.D., DiFiore, P., and Curiel, D.T.1998. The level of erbB2 expression predicts sensitivity to the cytotoxic effects ofan intracellular anti-erbB2 sFv. J. Mol. Med. 76:451–458.

35. Young, D.C., Kingsley, S.D., Ryan, K.A., and Dutko, F.J. 1993. Selective inactiva-tion of eukaryotic β-galactosidase in assays for inhibitors of HIV-1 TAT using bac-teria β-galactosidase as a reporter enzyme. Anal. Biochem. 215:24–30.

RESEARCH1998 Nature America Inc. • http://biotech.nature.com

1998

Nat

ure

Am

eric

a In

c. •

http

://bi

otec

h.na

ture

.com

document

Documents