Abstract
α-Fetoprotein (AFP) is a potential target for immunotherapy in hepatocellular carcinoma; both the murine and human T-cell repertoires can recognize AFP-derived epitopes in the context of the MHC. Protective immunity can be generated with AFP-engineered dendritic cell-based vaccines. We now report a DNA-based immunization strategy using a prime-boost approach: coadministration of plasmid DNA encoding murine AFP and murine granulocyte-macrophage colony-stimulating factor followed by boosting with an AFP-expressing nonreplicating adenoviral vector. This immunization strategy can elicit a high frequency of Th1-type AFP-specific cells leading to tumor protective immunity in mice at levels comparable with AFP-engineered dendritic cells. This cell-free mode of immunization is better suited for large-scale vaccine efforts for patients with hepatocellular carcinoma.
INTRODUCTION
HCC4 is a major cause of cancer death with more than 1.2 million global annual incidences. The incidence of HCC is rising rapidly in both Asian and Western countries because of the global pandemic of hepatitis B and C infections (1) . Surgery and liver transplantation are the only effective treatments, but most HCC patients are not eligible because of diagnosis at a late stage or underlying liver insufficiency in the setting of cirrhosis. We have reported previously that AFP can serve as target for T-cell immunotherapy (2, 3, 4) , an observation confirmed by others (5) .
Strong AFP-specific T-cell immunity can be generated by DCs engineered to express AFP. However, as a treatment strategy for large numbers of patients, DC-based approaches are cumbersome and expensive. An approach not requiring ex vivo cell manipulation would be more easily applied to large populations. We observed that repetitive AFP plasmid immunization generated weaker and less reproducible antitumor protection than engineered DCs (4) . We sought to optimize a DNA-based vaccine approach. A number of groups have shown that superior levels of T-cell immunity can be generated using a heterogeneous prime-boost strategy in which animals are primed with a plasmid vector and boosted with a viral vector encoding the same antigen (6, 7, 8, 9, 10, 11, 12, 13, 14) . In many of these vaccine models, expression of GM-CSF with the plasmid prime enhanced immunogenicity, presumably through the attraction of host APCs (15 , 16) . We report that priming mice with plasmids encoding AFP and GM-CSF and boosting with adenovirus encoding AFP elicits robust T-cell responses to defined AFP class I peptide epitopes and strong protective immunity.
MATERIALS AND METHODS
Mice and Cell Lines.
HLA-A2.1/Kb transgenic mice (17 , 18) were provided by Dr. Linda Sherman (Scripps Research Institute, La Jolla, CA) and are currently bred by the animal facility of the Department of Radiation Oncology at UCLA. C57BL/6 mice (H-2b) were originally purchased from The Jackson Laboratory (Bar Harbor, ME) and are currently bred by the animal facility at UCLA. All of the mice were handled in accordance with the animal care policy of UCLA. All of the cell lines were maintained in complete medium, RPMI 1640 (Life Technologies, Inc.) with 10% fetal bovine serum (Omega Scientific, Tarzana, CA) and penicillin/streptomycin/fungizone (Life Technologies, Inc.). Transfected cell lines were maintained under constant G418 selection (0.5mg/ml; Life Technologies, Inc.) in complete medium. Jurkat cells stably transfected with the chimeric class I molecule A2.1/Kb and cotransfected with either hAFP (Jurkat A2.1/hAFP) or MART-1 (Jurkat A2.1/MART-1) were developed as described previously (2) . EL4 is a spontaneous murine lymphoma that arose in C57BL/6 mice and was obtained from the American Type Culture Collection (Rockville, MD). EL4 cells stably transfected with mAFP (EL4 mAFP) was developed as described previously (4) .
Plasmid, Recombinant Adenovirus, and DC Immunizations.
AdVhAFP and AdVmAFP were generated as described previously (2 , 4) . The hAFP, mAFP, hMART-1, and mGM-CSF plasmids used in this study were constructed using the pVAX1 expression vector (Invitrogen, Carlsbad, CA). This backbone contains the kanamycin resistance gene and no extraneous expression cassettes, being suitable for clinical testing. All of the DNA for injection was prepared using an endo-free Maxi Prep kit (Qiagen, Valencia, CA) to eliminate endotoxin. Plasmids and adenoviral vectors were stored at −80°C. Plasmids were diluted in saline to 100 μg/100 μl, and AdVmAFP and AdVhAFP were adjusted to 1 × 109 pfu/100 μl for injection before injecting into the left anterior tibialis muscle of mice. Priming with plasmids was performed with either 100 μg of AFP plasmids or 100 μg mixed with 100 μg pmGM. PVAX1 plasmids lacking inserts were used as controls and were added to other groups to equalize the total amount of injected plasmid. A 0.3 insulin syringe with a 25-gauge 0.5-inch-long needle was used for the i.m. injections. Mice were boosted i.m. with adenovirus 2 weeks after priming with DNA. The dose of virus was delivered in a total volume of 100 μl in normal saline. DCs were differentiated from murine bone marrow progenitor cells following the Inaba method (19) with modifications (20) . On day 8 of culture, DCs were transduced with AdVhAFP or AdVmAFP as described (2 , 4) and washed three times with PBS before 5 × 105 cells were injected s.c. into mice.
ELISPOT Assay.
The ELISPOT assay was used to measure the frequency of cells producing the cytokine IFN-γ in splenocytes harvested from immunized mice as described previously (2 , 21) . Two weeks after immunization, splenocytes were harvested and restimulated directly in anti-IFN-γ monoclonal antibody (PharMingen) coated ELISPOT plate wells in vitro with either 4 μg/ml of HLA-A2.1-restricted AFP peptide(s) or 4 μg/ml the control peptide MART27–35 in RPMI 1640 containing 10% fetal bovine serum, 10 units/ml of human interleukin-2 and 50 μm of 2-mercaptoethanol. The plates (Millipore, Bedford, MA) were incubated with restimulated cells or fresh splenocytes (in duplicate at three dilutions) at 37°C for 24 h. The plates were then washed and incubated with a biotin-conjugated secondary antibody and then developed. The colored spots, representing cytokine producing cells, were counted under a dissecting microscope.
Peptide Synthesis.
Peptides were synthesized at the UCLA Peptide Synthesis Facility (Dr. Joseph Reeve, Jr., Director) using standard f-moc technology. Stock solutions were prepared in DMSO (10 mg or 5 mg/ml) and were kept in −80°C until used.
In Vivo Tumor Challenge.
Tumor challenge was performed 2 weeks after the last immunization with 105 cells in a single cell suspension per animal from tumors progressively growing in syngeneic mice. Tumor cells for challenge were washed after enzymatic digestion and resuspended in 0.2 ml of PBS/animal to be injected s.c. into the left flank. All of the tumor challenge doses were > 70% viable as determined by trypan blue exclusion. The size of tumors was assessed three times weekly using calipers. Tumor volume was approximated by the following calculation: 4/3 π r3 (r = radius). The Student t test or the Mann-Whitney rank-sum test was performed to interpret the significance of differences between final tumor volumes of the different groups of animals (presented as mean +/− SE in the figures).
RESULTS
Prime-Boost Vaccines Generate AFP T-Cell Responses in Transgenic Mice.
We have shown previously that A2.1/Kb tg mice will recognize several human AFP-derived peptide epitopes in the context of A2.1 when immunized with AdVhAFP/DC (3 , 21) . We used these mice as an immunological testing platform to screen various DNA-based vaccine combinations. A T-cell response to the immunodominant epitope AFP158–166 [FMNKFIYEI] was measured using a IFN-γ ELISPOT assay.
As shown in Fig. 1⇓ , immunization of A2.1/Kb tg mice with AdVhAFP/DC elicits strong T-cell responses to this A2.1 epitope. An i.m. vaccination with AdVhAFP alone produced a small response, as did vaccination with hAFP plasmid alone. However, a combination of these two vectors, in which mice were primed with 100 μg of phAFP and boosted with 109 pfu AdVhAFP elicited a greater response, and one that was comparable with DC immunization if a plasmid encoding mGM-CSF was included in the prime. A variety of controls demonstrate the need for heterologous vectors encoding the same antigen.
Frequency of IFN-γ-producing splenic cells as measured by ELISPOT assay. Splenocytes from immunized mice were harvested 2 weeks after the last immunization and restimulated in vitro with AFP158–166. Priming with phAFP/pmGM and boosting with AdVhAFP resulted in significantly higher level of IFN-γ-producing cells than injecting phAFP (P = 0.0001), phAFP/pmGM (P = 0.0001), AdVhAFP (P = 0.003), phAFP-AdVhAFP (P = 0.03), or phAFP-AdVLuc (P = 0.004). The difference between AdVhAFP/DC and phAFP/pmGM-AdVhAFP is not statistically significant (P = 0.3). Priming with phAFP/pmGM-CSF was performed with 100 μg of phAFP and 100 μg of pmGM-CSF in 200 μl of saline, and injected into the left and right anterior tibialis muscle (100 μl each). Empty vector (100 μg) was injected alone with phAFP when priming without pmGM-CSF to equalize the amount of plasmid used. Boosting with either AdVhAFP or AdVLuc was performed with 109 pfu in 100 μl of saline. Spots from restimulation with the control MART27–35 peptide were less than 20/106 cells in all groups and were subtracted. Similar results were observed in three of three independent experiments; bars, ± SE.
AFP plasmid DNA administered i.m. weekly or biweekly, two to four times, produced only a minor enhancement of AFP-specific responses. We also did not observe any enhancement with the coadministration of phAFP with plasmids expressing murine IL-12 or B7.1 (data not shown). A reduction in the dose of AdVhAFP boost from 109 pfu to 108 reduced this response to one third (and to background with 107 and 106 pfu). An increase to 1010 did not improve the response as measured by ELISPOT (data not shown).
Prime-Boost Immunization Generates an AFP-Specific Multiepitopic Response.
In addition to AFP158–166, three other hAFP immunodominant epitopes (AFP542–550, AFP325–334, and AFP137–145) identified previously were also recognized after prime-boost immunization of A2.1/Kb tg mice (Fig. 2)⇓ . The magnitude of the response to each of the epitopes varied from experiment to experiment, although restimulation with AFP158–166 in general resulted in the greatest number of spots (Figs. 1⇓ and 2⇓ ). These data suggest that the AFP gene delivered by our prime-boost vaccination was expressed, processed, and presented in an efficient manner to the T-cell repertoire of A2.1/Kb tg mice.
Frequency of IFN-γ-producing cells was determined in splenocytes 2 weeks after the last immunization and restimulated in vitro with each of the four immunodominant AFP epitopes (AFP158–166, AFP542–550, AFP325–334, and AFP137–145). Prime-boost activated splenocytes recognized all four HLA-A2.1-restricted human AFP epitopes. As expected, splenocytes from AdVhAFP/DC immunized mice also recognized the four epitopes. Spots from restimulation with the negative control MART27–35 peptide were less than 20/106 cells in all groups and were subtracted; bars, ± SE.
Recognition of Cells Endogenously Expressing AFP.
To demonstrate that the prime-boost immunization strategy stimulates T cells, which can recognize tumor cells expressing AFP endogenously, splenocytes from immunized mice were restimulated with stably transfected Jurkat A2.1/Kb/hAFP cells instead of individual peptides. As show in Fig. 3⇓ , prime-boost vaccination stimulated cells that recognized endogenously expressed and processed AFP at a level similar to AFP-transduced DC vaccination.
Frequency of IFN-γ-producing cells after splenocytes restimulated with an AFP-expressing cell line. Splenocytes from prime-boost and AdVhAFP/DC immunized recognized irradiated Jurkat A2.1/hAFP cells after restimulated in vitro for 24-h. phAFP/pmGM-AdVhAFP-immunized mice produced significantly higher number of spots than mice immunized with phAFP/pmGM alone (P = 0.01) and AdVhAFP alone (P = 0.03). The difference between AdVhAFP/DC- and phAFP/pmGM-AdVhAFP-immunized mice is not significant (P = 0.35). Untreated control mice produced no IFN-γ spots. Spots from restimulation with the control JurkatA2.1/MART27–35 cells were less than 30/106 cells in all groups and were subtracted; bars, ± SE.
Prime-Boost Immunization Protects Mice From in Vivo Tumor Challenge.
Having identified a potentially superior method of DNA-based immunization using the A2.1/Kb tg mouse platform, we determined whether AFP-specific antitumor protection could be generated in immunocompetent mice. C57BL/6 mice were primed with a coinjection of pmAFP, and pmGM-CSF followed 2 weeks later with a boost of AdVmAFP. These mice were challenged with EL4 mAFP (EL4 lymphoma cell line stably transfected with mAFP) or the parental EL4 cell line. Fig. 4a⇓ shows that prime-boost immunizations resulted in significant protection compared with untreated mice (P < 0.05 on day 18) or mice injected with pmAFP/pmGM alone (P < 0.05 on day 18). AdVmAFP vaccination alone was also unable to confer protection against EL4 mAFP challenge (Fig. 4a)⇓ . The degree of protection in prime-boost immunized mice was comparable with AdVmAFP/DC immunized mice. This protection was AFP specific, because there is no benefit in mice challenged with EL4 (Fig. 4b)⇓ . In eight replicate experiments, complete protection from tumor challenge with EL4 mAFP using the prime-boost immunization method was achieved in 15 of 40 mice (37.5%; P < 0.05 versus control group) compared with 18 of 40 (45%) mice in AdVmAFP/DC immunized mice (P < 0.05 versus control group) after 40 days of observation. In vivo protection correlated with IFN-γ ELISPOT responses using splenocytes restimulated with EL4 mAFP (Fig. 4c)⇓ .
DNA prime and adenovirus boost immunization generate AFP-specific immunity. a, delayed tumor growth observed in prime-boost immunized mice after EL4 mAFP challenge. Mice received one i.m. injection of 100 μg of pmAFP admixed with 100 μg of pmGM followed by one i.m. injection of AdVmAFP (109 pfu) 2 weeks later (▴), pmAFP/pmGM alone (▪), AdVmAFP alone (•), AdVmAFP/DC (▵), or no treatment (□). Two weeks after last immunization, mice were challenged with 105 EL4 mAFP cells. Tumor sizes were significantly smaller in prime-boost and AdVmAFP/DC immunized mice than untreated control mice (P < 0.05) and mice treated with pmAFP/pmGM alone (P < 0.05) or AdVmAFP alone (P < 0.05). b, prime-boost immunization has no effect in mice from a challenge with parental EL4 tumor. Tumors were progressed at approximately the same rate in prime-boost immunized mice (▪) and untreated control mice (□). c, frequency of IFN-γ-producing splenic cells retrieved from mice immunized the prime-boost method after restimulated in vitro with irradiated EL4 mAFP for 24 h had higher IFN-γ release as measured by ELISPOT assay than control (P = 0.009), pmAFP/pmGM (P = 0.0006), or AdVmAFP-immunized (P = 0.002) mice. AdVmAFP/DC-immunized mice had a higher frequency of splenic IFN-γ-producing cells than in prime-boost immunized mice (P = 0.02); bars, ± SE.
DISCUSSION
AFP is a transcriptionally regulated protein expressed by most hepatocellular carcinomas (22) . We originally demonstrated that both murine and human T-cell repertories could recognize AFP (2, 3, 4) despite being exposed to high plasma levels of this oncofetal protein during embryonic development. A complete analysis of 74 hAFP-derived peptides revealed four immunodominant and ten subdominant epitopes restricted by HLA-A2.1 (2 , 3) . Several of these immunodominant peptides could be eluted from the surface of an A2.1/AFP-positive HCC cell line (3) . AFP can serve as a tumor rejection antigen in a murine model of DC-based genetic immunotherapy. DCs engineered to express murine AFP induced potent T-cell responses in mice, as evidenced by the generation of AFP-specific CTL, cytokine-producing T cells, and protective immunity (4) . We observed that plasmid AFP immunization resulted in detectable but low levels of AFP-specific T-cell responses and poorly reproducible protective immunity (4) .
DNA-based vaccination (both naked plasmid and viral vector) is a potentially powerful method to immunize against microbial and viral antigens through both humoral antibody and cell mediated responses (23) . The mechanism of generation of T-cell immunity involves local target cell transfection and protein antigen production, which is taken up by host APC leading to cross-presentation in the draining lymph nodes; in addition, direct DNA transfection of peripheral tissue APC has also been demonstrated (24, 25, 26, 27) . The plasmid backbone in this approach is derived from bacterial DNA, which contains immunostimulatory nucleotide sequences (hypomethylated cytidine phosphate guanosine motifs) that have the ability to attract APC to the vaccination site and induce a type 1 immune response in humans (28) . However, naked DNA immunization has shown a limited ability to immunize to tumor antigens (29) . Therefore, several strategies to enhance its ability to generate immune responses have been tested, including the coinjection of muscle damaging agents like snake venom or bupivacaine (because naked DNA transfection in regenerating muscle was found to be more efficient; Refs. 30 , 31 ), the coadministration of GM-CSF to enhance the attraction of APC (32) , and the coinjection of plasmids carrying immunostimulatory molecules like CD40-L, IL-2, and IL-12 (24 , 33, 34, 35) .
Additional enhancement of the T-cell stimulatory effect of plasmid DNA can be accomplished by boosting with viral vectors expressing the same antigen. The possible mechanism for this boosting effect is the presentation of the same antigen in a highly immunogenic milieu generated by the viral epitopes. The prime-boost strategy has been tested in small and large animal models resulting in high levels of circulating antigen-specific cells in peripheral blood (up to 20–30% by MHC tetramer analysis; Refs. 8 , 9 , 15 , 36, 37, 38, 39, 40, 41 ).
DC-based antitumor vaccines have demonstrated a powerful ability to generate tumor-specific immune responses both in preclinical (42) and clinical settings (43, 44, 45, 46) . However, the need for costly cell culture procedures preclude their wide availability for clinical use, and minor variations in the culture technique or antigen loading may yield suboptimal, even tolerizing, vaccines (47 , 48) . Therefore, alternative cell-free vaccines with a similar ability to generate responses to tumor antigens would be desirable for the development of widely clinically applicable tumor vaccines. Our data suggests that that sequential immunization with plasmid DNA and adenovirus encoding AFP can generate effective antitumor immunity in mice, which may be suitable for large-scale clinical testing as a vaccine for subjects with HCC.
Footnotes
-
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
-
↵1 Supported in part by NIH/National Cancer Institute Grants RO1 CA79976, RO1 CA77623, and K12 CA76905; the Monkarsh Fund; the Naify Fund; and the Stacy and Evelyn Kesselman Research Fund. W. S. M. is supported by NIH training Grant T32 CA75956 and a fellowship from the Jonsson Cancer Foundation. A. R. is a recipient of an American Society of Clinical Oncology (ASCO) Career Development Award.
-
↵2 Present address: Mylan School of Pharmacy, Duquesne University, Pittsburgh, PA 15282.
-
↵3 To whom requests for reprints should be addressed, at Division of Surgical Oncology, Room 54-140 CHS, UCLA School of Medicine, 10833 Le Conte Avenue, Los Angeles, CA 90095-1782. Phone: (310) 825-2644; Fax: (310) 825-7575; E-mail: jeconomou{at}mednet.ucla.edu
-
↵4 The abbreviations used are: HCC, hepatocellular carcinoma; AFP, α-fetoprotein; mAFP, murine α-fetoprotein; DC, dendritic cell; GM-CSF, granulocyte-macrophage colony stimulating factor; ELISPOT, enzyme-linked immunospot; APC, antigen presenting cell; UCLA, University of California, Los Angeles; phAFP, plasmid DNA encoding human AFP; hAFP, human α-fetoprotein; IL, interleukin; pfu, plasmid forming unit(s); pmAFP, DNA encoding murine AFP; AdVhAFP, adenovirus encoding human AFP; AdVmAFP, adenovirus encoding murine AFP; mGM-CSF, murine granulocyte-macrophage colony stimulating factor.
- Received July 12, 2001.
- Accepted October 18, 2001.
- ©2001 American Association for Cancer Research.