Abstract

Proteins may serve as ideal CD8+ T cell immunogens for human immunodeficiency virus type 1 (HIV-1) if they can be delivered to and processed through the human leukocyte antigen class I pathway. This study shows that human blood monocyte–derived dendritic cells loaded with liposome-complexed HIV-1 proteins and matured with CD40 ligand can prime CD8+ T cells to HIV-1 in vitro. Whole HIV-1 protein in liposome may be an effective immunogen for HIV-1 vaccine protocols

Generation of antigen-specific, primary CD8+ cytotoxic T lymphocyte (CTL) responses represents a challenge to the development of protein-based vaccines for prevention of human immunodeficiency virus type 1 (HIV-1) infection. The conventional mechanism for antigenic processing of extracellular proteins is through an endocytic pathway for presentation by major histocompatibility complex (MHC) class II molecules to CD4+ T cells [1]. One potential strategy to induce vigorous MHC class I–restricted CTL responses with viral proteins involves specific targeting to the cytosol of antigen-presenting cells (APCs), particularly dendritic cells (DCs), with carriers such as liposome for accessing alternative MHC class I antigen-processing pathways (termed “cross-presentation”) [2]. We have previously demonstrated that lipofectin, a cationic liposome, provides a means of delivering exogenous proteins in vitro to human DCs for more efficient induction of HIV-1 specific, memory CD8+ CTL responses than protein alone [3]. The present study investigated whether DCs loaded with protein-liposome complexes can prime anti–HIV-1 CD8+ T cell responses

Materials and methods To obtain immature DCs (iDCs), CD14+ monocytes were positive selected from peripheral blood mononuclear cells (PBMC) of healthy donors using anti-CD14 monoclonal antibody (MAb)–coated magnetic microbeads (Miltenyi) and cultured in AIM V medium (GIBCO) containing 1000 U/mL of recombinant interleukin (IL)–4 and 1000 U/mL recombinant granulocyte-monocyte colony stimulating factor (GM-CSF; Schering-Plough) [4]. The number of viable DCs was determined by typical DC morphology in trypan blue dye–stained preparations. The purity of the DCs was ⩾90% on the basis of the expression of HLA-DR molecules and lack of expression of lineage markers by flow cytometry [4]. The maturation state of the DCs was determined by expression of CD80, CD83, and CD86

The toxicity of lipofectin, composed of DOTMA (N-[1-{2,3-dioleyloxy} propyl]–N, N, N-trimethylammonium chloride)/DOPE (dioleoyl-phosphotidylethanolamine) (GIBCO), was determined by loading iDCs with graded doses at 37°C for 0, 2, 4, and 8 h. The viability of the DCs by trypan blue dye exclusion was >90% when lipofectin was added at 15 μg/mL for up to 8 h, and decreased at higher concentrations of liposome. Therefore, 15 μg/mL lipofectin was used in this study.

The vaccinia virus (VV) constructs contained gag p24 or p55 for BH10 and HxB2 strains of HIV-1 (VV-Gag p24 and VV-Gag p55; Therion Biologics). VV-Env was vPE11 containing the env coding region of HIV-1 BH10 minus the signal sequence (B. Moss, National Institutes of Health [NIH], Bethesda, MD). The NYCBH strain of VV (VV-Vac; Therion) was used as control virus. Synthetic peptides representing previously identified HLA class I–restricted Gag and Env epitopes were prepared by the University of Pittsburgh Peptide Synthesis Facility

For preparation of APCs, 15 μg/mL lipofectin was mixed with 20 μg/mL HIV-1 p24 Gag or gp160 Env (IIIB; Protein Sciences) or p55 Gag (SF2; Chiron; AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, NIH) and incubated for 10 min at room temperature [3]. The lipofectin-protein mixture was added to 1×106 iDCs on day 7. The cells were incubated for 4 h at 37°C in 5% CO2 and treated with human CD40L trimer (1–2.5 μg/mL; Immunex) for 2 days to induce phenotypic and functional maturation (defined as mature DCs [mDCs]). Control samples were APCs only and APCs loaded with either liposome or HIV-1 protein. In some experiments, the iDCs were obtained from plastic adherent PBMC and matured with monocyte-conditioned medium (MCM) [4]

After incubation, the cells were washed and left untreated, or treated with psoralen (10 μg/mL; Calbiochem) and UV light (PUV) [5]. Cryopreserved PBMC were primed with antigen-loaded, autologous mDCs for 1 week and then stimulated with antigen-loaded, autologous PBMC as APCs for weeks 2–5 at an effector cell-to-APC ratio of 10:1 in RPMI 1640 medium containing antibiotics, 15% heat inactivated fetal calf serum, and recombinant human interleukin (IL)–2 (100 U/mL; Chiron). The effector cell-APC mixtures were seeded into round-bottom microtiter wells, with fresh medium and IL-2 added every 3–4 days

Antigen stimulated cells were harvested and washed after 4 or 5 weeks when there were adequate T cells for CTL assay. In some experiments, CD8+and CD4+ T cells were enriched to 98% and 90% purity, respectively, from cultured cells by negative selection using immunomagnetic beads (Dynal) [5]. The targets were autologous B-LCLs infected with the VV vectors or loaded with synthetic HIV-1 peptides (100 μg/mL) and labeled with 51Cr [4]. PBMC were typed for HLA class I by high resolution, reference strand mediated conformation analysis. To block CTL reactivity, target cells were incubated with anti-HLA class I (5 μg/mL; Accurate Chemical) or anti-HLA class II MAbs (5 μg/mL; Beckman Coulter) for 45 min at 37°C before adding to the effector cells. The CTL lysis assay was done in triplicate at effector/target (E/T) cell ratios of 40, 20, and 10. Specific lysis was calculated as 100 × [(experimental copies per minute {cpm} − spontaneous cpm)/(maximum cpm − spontaneous cpm)]. The percentage of antigen-specific lysis was calculated as (percentage lysis of VV-Gag or VV-Env infected or HIV-1 peptide loaded targets) − (percentage lysis of VV-vac infected or mock treated targets). The mean (±SE) lysis of the uninfected and VV-vac infected targets by the antigen-stimulated effectors at the various E/T ratios was 8%±3% (n=18) and 7%±1% (n=27), respectively. The coefficient of variation for cpm among replicate CTL values was 7%±1%. A positive CTL response was defined as ⩾10% antigen-specific lysis

Results Expression of costimulatory molecules CD80 and CD86, HLA-DR, and the Ig superfamily, DC surface maturation marker, CD83, was up-regulated on CD40L-treated DCs as compared to non-CD40L-treated DCs, indicating that the CD40L treatment induced DC maturation [4, 5]. Loading of iDCs or mDCs with either liposome, p24 Gag, or their combination did not alter expression of these surface molecules (data not shown)

We next established the parameters for in vitro priming by confirming the work of Zarling et al. [6], which used 1 week of priming of PBMC with HIV-1 Gag peptide-loaded, MCM-matured DCs, followed by weekly, sequential stimulations with Gag peptide-loaded PBMC to induce anti-Gag CD8+ CTL responses (data not shown). Priming with peptide-loaded iDCs did not induce CTL reactivity. In subsequent experiments, we used CD40L-matured DCs, because this treatment resulted in a more consistent maturation of the DCs [4]

Stimulation of PBMC from an HIV-1–negative subject with autologous mDCs that had been loaded with HIV-1 Gag p55 protein-liposome induced greater levels of anti–HIV-1 CD8+ T cell responses than mDCs loaded with HIV-1 Gag p55 protein alone (figure 1A). Similar priming was elicited by PUV-treated mDCs that had been loaded with Gag p55-liposome (figure 1B). The lytic reactivity was HLA class I restricted as shown by blocking of lysis by anti–HLA class I but not anti–HLA class II MAbs (figure 1B). To date, we have observed that CTLs specific for Gag or Env can be consistently primed by autologous mDCs loaded with these HIV-1 recombinant protein-liposome complexes in multiple PBMC samples obtained from 7 different subjects (data not shown).

Figure 1

Priming of human immunodeficiency virus type 1 (HIV-1)–specific CD8+ cytotoxic T lymphocyte (CTL) responses by different antigen-presenting cells (APCs). Priming of HIV-1–specific CTL in peripheral blood mononuclear cells (PBMC) from a HLA A*0201, B*27 subject by 1 week of stimulation with autologous mature dendritic cells (mDCs; either untreated or psoralen UV [PUV] treated), monocytes, or PBMC loaded with Gag p55-liposome, followed by 4 sequential weekly stimulations with autologous PBMC loaded with the same antigen (A). Results are the mean of triplicate values of antigen-specific lysis. The results show positive CTL responses (⩾10% specific lysis) to autologous B cell lines loaded with known CD8+ T cell epitopes for HLA A*0201 p24151–159 (TLNAWVKVV), and B*27 p24263–272 (KRWIILGLNK) and p1719–27(IRLRPGGKK), but not for HLA A*0201 p1777–85 (SLYNTVATL). The CTL activity was blocked by anti-HLA class I monoclonal antibody (MAb) but not anti-HLA class II MAb (B). E/T, effector/target cell ratio

Figure 1

Priming of human immunodeficiency virus type 1 (HIV-1)–specific CD8+ cytotoxic T lymphocyte (CTL) responses by different antigen-presenting cells (APCs). Priming of HIV-1–specific CTL in peripheral blood mononuclear cells (PBMC) from a HLA A*0201, B*27 subject by 1 week of stimulation with autologous mature dendritic cells (mDCs; either untreated or psoralen UV [PUV] treated), monocytes, or PBMC loaded with Gag p55-liposome, followed by 4 sequential weekly stimulations with autologous PBMC loaded with the same antigen (A). Results are the mean of triplicate values of antigen-specific lysis. The results show positive CTL responses (⩾10% specific lysis) to autologous B cell lines loaded with known CD8+ T cell epitopes for HLA A*0201 p24151–159 (TLNAWVKVV), and B*27 p24263–272 (KRWIILGLNK) and p1719–27(IRLRPGGKK), but not for HLA A*0201 p1777–85 (SLYNTVATL). The CTL activity was blocked by anti-HLA class I monoclonal antibody (MAb) but not anti-HLA class II MAb (B). E/T, effector/target cell ratio

The HLA class I specificity and breadth of this primary CD8+ T cell response was further demonstrated by killing of autologous targets expressing 3 of 4 HLA A*0201 and HLA B*27 HIV-1 peptides (figure 1A). We confirmed the specificity of CD8+ T cells primed by mDC loaded with either Gag p24 or p55 and liposome for p24151-159 (TLNAWVKVV) in 5 of 6 HLA A*0201 subjects. In further experiments, priming of CTLs from another HLA A*0201 person with mDCs loaded with gp160-liposome induced CTLs specific for HLA A*0201 epitopes gp120311–320 (RGPGRAFVTI) and gp41747–755 (RLVNGSLAL), but not gp120192–199 (KLTSCNTSN) or gp41818-827 (SLLNATDIAV) (data not shown). Similarly, CTLs from an HLA A*03 B*07 person that were primed by mDCs loaded with Gag p24- and Env gp160-liposome complexes were specific for HLA B*07 p24179–187 (ATPQDLNTM) and HLA A*03 gp41775–785 (RLRDLLLIVTR) epitopes, but not for the HLA B*07 p24148–156 (SPRTNAWV) epitope (data not shown).

We showed that Gag p55-liposome loaded mDCs more efficiently primed HIV-1–specific CD8+ CTLs than did monocytes or PBMC for recognition of targets expressing 4 different Gag epitopes (figure 1A). Furthermore, mDCs loaded with Gag p55-liposome were superior to iDCs for priming of CTL specific for targets expressing naturally processed Gag proteins (i.e., VV-p24 and VV-p55), as well as targets loaded with a known HLA A*0201 p24 epitope (p24151-159) (figure 2)

Figure 2

Priming of human immunodeficiency virus (HIV)-1–specific cytotoxic T lymphocyte (CTL) responses by immature dendritic cells (iDCs) and mature (m) DCs loaded with HIV-1 protein-liposome (lipo) complexes. Peripheral blood mononuclear cells (PBMC) from an HLA A*0201 subject were primed for HIV-1–specific CTL by 1 week of stimulation with autologous iDCs and mDCs loaded with Gag p24-lipo complexes, followed by 4 sequential weekly stimulations with autologous PBMC loaded with the same antigen. Targets were autologous B cell ines infected with vaccinia virus expressing Gag p55 or p24, or loaded with an HIV-1 peptide (p24151–159 [TLNAWVKVV]) representing a known CTL epitope for HLA A*0201. Results are the mean of triplicate values of antigen-specific lysis. E/T, effector/target cell ratio

Figure 2

Priming of human immunodeficiency virus (HIV)-1–specific cytotoxic T lymphocyte (CTL) responses by immature dendritic cells (iDCs) and mature (m) DCs loaded with HIV-1 protein-liposome (lipo) complexes. Peripheral blood mononuclear cells (PBMC) from an HLA A*0201 subject were primed for HIV-1–specific CTL by 1 week of stimulation with autologous iDCs and mDCs loaded with Gag p24-lipo complexes, followed by 4 sequential weekly stimulations with autologous PBMC loaded with the same antigen. Targets were autologous B cell ines infected with vaccinia virus expressing Gag p55 or p24, or loaded with an HIV-1 peptide (p24151–159 [TLNAWVKVV]) representing a known CTL epitope for HLA A*0201. Results are the mean of triplicate values of antigen-specific lysis. E/T, effector/target cell ratio

Discussion We were consistently able to prime anti–HIV-1 CD8+ T cell responses with mDCs loaded with HIV-1 protein-liposome complexes but not with mDCs loaded with HIV-1 proteins alone. In vitro priming of CD8+ T cells to HIV-1 was first demonstrated by Engleman et al. [7, 8] who showed that multiple, sequential stimulations of naive CD8+ T cells with autologous or allogeneic, HLA class I–matched DCs freshly isolated from blood samples and loaded with HIV-1 Gag and Env peptides, could activate HIV-1–specific CTL. More recently, monocyte-derived iDCs transduced with an env- and nef-deleted HIV-1 vector pseudotyped by VSV G protein [9], HIV gag-encoded mRNA-transfected DCs [10], or iDCs that were matured with MCM or CD40L and loaded with HIV-1 peptides representing HLA class I epitopes [6, 11], have been used to prime anti–HIV-1 CTL in vitro. There is only one report of in vitro priming of CD8+ T cells with whole HIV-1 protein, where Nef-specific CD8+ T cells were primed by repetitive stimulation of PBMC with γ-irradiated iDCs loaded with HIV-1 Nef protein [12]. In this case, optimal T cell priming required addition of interferon (IFN)–α

We found that priming of anti–HIV-1 CD8+ T cell responses in vitro required liposome as a delivery vehicle. We believe that this was due, in part, to the unique fusogenic property of the DOPE portion of lipofectin that enhances access of exogenous protein from the endosomes/lysosomes into the cytosol of DCs by destabilizing the endosomes and allowing the release of HIV-1 protein into the cytosol [13]. Moreover, a single stimulation of PBMC for one week with HIV-1 protein-liposome loaded mDCs, followed by multiple, weekly stimulations with protein-liposome loaded PBMC, was sufficient to induce a primary, anti–HIV-1 CD8+ T cell response. Initial stimulation with antigen-loaded mDCs was essential for priming, as iDCs, monocytes or PBMC loaded with liposome-complexed HIV-1 protein were poor primers of HIV-1 specific T cells. The mDCs expressed enhanced levels of costimulatory and HLA class II molecules, and these were not altered by addition of liposome-protein complexes. Maturation of DCs with CD40L increases surface expression of HLA and costimulatory molecules and release of cytokines including IL-12 p70 (when combined with IFN-α stimulation) and IL-15 from mDCs that enhance antigen-specific CD8+ T cell responses [14]. Such costimulatory CD40/CD40L interactions lower the threshold of antigen required to activate primary T cell responses, and may augment survival of antigen-activated T cells

Our data, that PUV-treated mDCs could prime anti–HIV-1 CD8+ T cell responses, are concordant with our previous findings among HIV-1–infected persons [4]. The PUV-treated DCs are likely undergoing apoptosis, which suggest that viable DCs are not essential for triggering of T cell priming. This may relate to the high levels of costimulatory molecule expression that is retained on PUV-treated DCs (X.-L.H., Z.F., and C.R.R., unpublished data)

Priming of CD8+ T cells with mDCs loaded with HIV-1 proteins and liposome elicited CTL reactivity against multiple HIV-1 HLA class I epitopes. The levels of lysis were within the range expected for CTL primed by a multi-epitope protein immunogen, using single epitope, individual peptides as targets. Lack of recognition of some HIV-1 peptides by the HIV-1 protein-primed T cells could reflect immunodominance among the peptides for their HLA class I alleles, differences in proteasomal cleavage and heterogeneity of peptide binding due to polymorphism of HLA class I alleles [2]

We reported elsewhere activation of HIV-1 specific, CD8+ T cells from HIV-1–infected persons in vitro by cross-presentation of liposome-complexed HIV-1 proteins by DCs [3]. These studies suggest that complexing of HIV-1 protein with liposome promotes processing of the protein through an alternative, cytosolic pathway for cross-presentation of peptides by HLA class I molecules to both naive and memory CD8+ T cells. Our studies, together with a recent report showing in vitro priming of Gag-specific CD4+ T cells by mDCs loaded with uncomplexed HIV-1 p24 protein [15], suggest that a protein-based antigen approach could be used to prime or boost both anti–HIV-1 CD8+ and CD4+ T cell activity in prophylactic and therapeutic vaccine protocols

Acknowledgments

We thank K. Picha (Immunex; Seattle, WA) for assistance in providing the CD40L, S. Narula (Schering-Plough; Kenilworth, NJ) for providing hIL-4 and hGM-CSF, and J. Malenka for administrative assistance

References

1.
Watts
C
Amigorena
S
Antigen traffic pathways in dendritic cells
Traffic
 , 
2000
, vol. 
1
 (pg. 
312
-
7
)
2.
Yewdell
JW
Norbury
CC
Bennink
JR
Mechanisms of exogenous antigen presentation by MHC class I molecules in vitro and in vivo: implications for generating CD8+ T cell responses to infectious agents, tumors, transplants, and vaccines
Adv Immunol
 , 
1999
, vol. 
73
 (pg. 
1
-
77
)
3.
Zheng
L
Huang
XL
Fan
Z
Borowski
L
Wilson
CC
Rinaldo
CR
Jr
Delivery of liposome‐encapsulated HIV type 1 proteins to human dendritic cells for stimulation of HIV type 1–specific memory cytotoxic T lymphocyte responses
AIDS Res Hum Retroviruses
 , 
1999
, vol. 
15
 (pg. 
1011
-
20
)
4.
Fan
Z
Huang
XL
Borowski
L
Mellors
JW
Rinaldo
CR
Jr
Restoration of anti–human immunodeficiency virus type 1 (HIV‐1) responses in CD8+ T cells from late‐stage patients on prolonged antiretroviral therapy by stimulation in vitro with HIV‐1 protein‐loaded dendritic cells
J Virol
 , 
2001
, vol. 
75
 (pg. 
4413
-
9
)
5.
Zhao
XQ
Huang
XL
Gupta
P
, et al.  . 
Induction of anti–human immunodeficiency virus type 1 (HIV‐1) CD8+ and CD4+ T‐cell reactivity by dendritic cells loaded with HIV‐1 X4–infected apoptotic cells
J Virol
 , 
2002
, vol. 
76
 (pg. 
3007
-
14
)
6.
Zarling
AL
Johnson
JG
Hoffman
RW
Lee
DR
Induction of primary human CD8+ T lymphocyte responses in vitro using dendritic cells
J Immunol
 , 
1999
, vol. 
162
 (pg. 
5197
-
204
)
7.
Mehta‐Damani
A
Markowicz
S
Engleman
EG
Generation of antigen‐specific CD8+ CTLs from naive precursors
J Immunol
 , 
1994
, vol. 
153
 (pg. 
996
-
1003
)
8.
Dupuis
M
Peshwa
MV
Benike
C
, et al.  . 
Allogeneic dendritic cell induction of HIV‐specific cytotoxic T lymphocyte responses from T cells of HIV type 1–infected and uninfected individuals
AIDS Res Hum Retroviruses
 , 
1997
, vol. 
13
 (pg. 
33
-
9
)
9.
Gruber
A
Kan‐Mitchell
J
Kuhen
KL
Mukai
T
Wong‐Staal
F
Dendritic cells transduced by multiply deleted HIV‐1 vectors exhibit normal phenotypes and functions and elicit an HIV‐specific cytotoxic T‐lymphocyte response in vitro
Blood
 , 
2000
, vol. 
96
 (pg. 
1327
-
33
)
10.
Weissman
D
Ni
H
Scales
D
, et al.  . 
HIV Gag mRNA transfection of dendritic cells (DC) delivers encoded antigen to MHC Class I and II molecules, causes DC maturation, and induces a potent human in vitro primary immune response
J Immunol
 , 
2000
, vol. 
165
 (pg. 
4710
-
7
)
11.
von Bergwelt‐Baildon
MS
Vonderheide
RH
Maecker
B
, et al.  . 
Human primary and memory cytotoxic T lymphocyte responses are efficiently induced by means of CD40‐activated B cells as antigen‐presenting cells: potential for clinical application
Blood
 , 
2002
, vol. 
99
 (pg. 
3319
-
25
)
12.
Wilson
CC
Olson
WC
Tuting
T
Rinaldo
CR
Lotze
MT
Storkus
WJ
HIV‐1–specific CTL responses primed in vitro by blood‐derived dendritic cells and Th1‐biasing cytokines
J Immunol
 , 
1999
, vol. 
162
 (pg. 
3070
-
8
)
13.
Chesnoy
S
Huang
L
Structure and function of lipid‐DNA complexes for gene delivery
Annu Rev Biophys Biomol Struct
 , 
2000
, vol. 
29
 (pg. 
27
-
47
)
14.
van Kooten
C
Banchereau
J
CD40‐CD40 ligand
J Leuko Biol
 , 
2000
, vol. 
67
 (pg. 
2
-
17
)
15.
Schlienger
K
Craighead
N
Lee
KP
Levine
BL
June
CH
Efficient priming of protein antigen‐specific human CD4+ T cells by monocyte‐derived dendritic cells
Blood
 , 
2000
, vol. 
96
 (pg. 
3490
-
8
)
Presented in part: Dendritic Cells at the Host-Pathogen Interface Conference, sponsored by the National Institute of Allergy and Infectious Diseases, Warrenton, Virginia, 7 May 2002
Informed consent was obtained from the study subjects, and this study was done in accordance with experimentation guidelines of the US Department of Health and Human Services and the Institutional Review Board of the University of Pittsburgh.
Potential conflict of interest: E.K.T. is an employee of Immunex Corporation, who supplied the CD40 ligand for this study.
Financial support: National Institutes of Health (grants R01 AI41870, U01 AI37984, U01 AI35041, R01 CA24553, and T32 AI 07487).
Present affiliation: Department of Immunology, Section of Immunobiology, Yale University School of Medicine, New Haven, Connecticut.