Humoral and cellular immune responses of dogs immunized ... - Nature

Gene Therapy (1998) 5, 865–868  1998 Stockton Press All rights reserved 0969-7128/98 $12.00 http://www.stockton-press.co.uk/gt

Humoral and cellular immune responses of dogs immunized with a nucleic acid vaccine encoding human carcinoembryonic antigen BF Smith1, HJ Baker1, DT Curiel2, W Jiang1 and RM Conry2 1

Scott-Ritchey Research Center, Auburn University, Auburn, AL; and 2Comprehensive Cancer Center, University of Alabama, Birmingham, AL, USA

A nucleic acid vaccine encoding human carcinoembryonic antigen (CEA) was administered to 10 juvenile dogs, 10– 15 weeks of age, by four parenteral routes. The routes tested were intramuscular injection using a needle and syringe, intramuscular injection using the Biojector needleless injection device, intradermal injection or intravenous injection. All groups received 150 ␮g of plasmid DNA on weeks 0, 4, 7 and 13. All dogs were bled weekly for 17 weeks and tested for antibody against human CEA. Dogs given plasmid intramuscularly either by needle and syringe or Biojector showed significant antibody responses by week 9 which peaked by week 15. Dogs receiving plasmid intra-

venously showed slight, unsustained increases in antibody titers while dogs receiving plasmid intradermally had significant titers, but at levels approximately one log less than those induced by intramuscular injection. The five dogs immunized by intramuscular delivery of plasmid DNA were examined for cellular immune responses to human CEA by lymphoblast transformation (LBT) assay. All five showed significant CEA-specific lymphoproliferation when compared with unvaccinated dogs. Physical examination, clinical chemistry, hematology and histopathology examinations revealed no abnormalities associated with nucleic acid immunization.

Keywords: DNA; vaccine; CEA; canine; cellular; humoral

Introduction For several decades, the concept of immunization against tumor antigens has been pursued as a method of therapy for cancer.1 Initially, this consisted of injecting killed or irradiated tumor cells in an attempt to induce an immune response. With the identification of specific tumor-associated antigens, these antigens have been proposed as specific immunogens.2 Immune responses to these therapies have been variable and they have not as yet been widely implemented clinically. The discovery that nucleic acid expression constructs, when injected intramuscularly into mice, could generate an immune response to the product of the transferred gene, has had a dramatic impact on vaccinology.3–5 The ability to utilize nucleic acid immunization to induce specific antitumor immune responses in cancer patients would have several benefits over conventional immunization methods. Nucleic acid immunization has been shown to generate both cellular and humoral immunity.6,7 Recombinant protein and peptide vaccines frequently generate only humoral immunity failing to induce cellular immunity which is critical in immune-mediated destruction of tumor. Injection of irradiated tumor cells carries some risk of re-implanting viable cells if irradiation is inadequate. Nucleic acid vaccines contain no living organisms, have not been demonCorrespondence: BF Smith, Scott-Ritchey Research Center, College of Veterinary Medicine, Auburn, AL 36849, USA Received 17 November 1997; revised 28 January 1998; accepted 10 February 1998

strated to integrate into host chromosomes and should have a high degree of safety associated with their administration.8 While the initial experiments with nucleic acid vaccines demonstrated efficacy when used intramuscularly, subsequent reports have indicated that other routes such as intradermal and intravenous administration are efficacious.9,10 We have previously demonstrated that dogs immunized with a plasmid encoding the nucleocapsid protein of canine parvovirus are protected from virulent viral challenge.11 We therefore examined the ability of a nucleic acid vaccine encoding the human tumorassociated antigen, CEA, to generate both humoral and cellular immune responses when administered by four different parenteral routes, in a nonrodent outbred species, the domestic dog.

Results Administration of the plasmid in solution resulted in momentary slight discomfort in dogs injected intramuscularly, similar to that seen following intramuscular administration of commercial conventional vaccines. Skin blebs were observed in all intradermal injections indicating correct needle placement. No hyperemia or other response was seen in the area of these skin blebs. No abnormal clinical signs were observed following injection and no local reactions were observed at the injection sites. In order to evaluate humoral response to the product of the transferred CEA gene, antibody responses to CEA were monitored weekly (Table 1). Titers were first observed to rise above preinjection levels in three of five

Nucleic acid vaccination with CEA in dogs BF Smith et al

866 Table 1 Anti-CEA antibody responses in dogs injected with a CEA encoding plasmid Week

0

6

7

8

9

10

11

12

14

15

16

17

i.m.

10 0 1 30 0

4 47 49 35 22

0 13 13 31 20

58 400 55 117 200

82 1500 45 140 144

60 1500 37 96 95

67 1900 28 95 91

84 1300 26 87 120

1400 6700 1000 530 1100

1300 10 000 5080 1300 1600

1100 8400 1620 550 1200

740 7200 1299 380 1100

i.d.

21 11

3 2

0 0

8 0

24 48

16 0

30 43

23 25

450 46

490 400

520 360

360 310

i.v.

21 9 2

20 40 5

0 0 52

4 8 0

0 2 0

4 9 0

15 0 19

29 3 39

30 0 16

7 0 15

32 3 8

5 0 0

Bio

Dogs were injected on weeks 0, 4, 7 and 13. Antibody to human CEA was quantified by RIA with results reported as nanograms of 125 I-CEA bound per milliliter of serum. Values greater than 32 represent positive results. i.m., intramuscular injection; Bio, Biojector 2000 needleless intramuscular injection; i.d., intradermal injection; i.v., intravenous injection.

intramuscular (i.m.) injection group dogs at week 6 and in one intravenous (i.v.) group dog at week 6 and a second at week 7. All of these responses subsequently fell below baseline levels. Following the third injection at week 7, all five dogs immunized i.m. showed a rapid elevation in antibody levels. Only one dog in each of the intravenous and intradermal (i.d.) groups showed a brief elevation in anti-CEA antibody to significant levels. A fourth injection at week 13 resulted in an approximately five-fold increase in the antibody levels of the i.m. injection groups, and an elevation to significant levels in both dogs who were immunized intradermally. The anti-CEA antibody response before immunization and 2 weeks after the second and third boost (weeks 9 and 15) is shown graphically in Figure 1. The anti-CEA antibody levels in all of the dogs which responded to the fourth injection remained relatively stable to the last sample at week 17. Due to high levels of antibody response in the intramu-

scular groups, these animals were examined for a cellular immune response. The animals were killed at week 25, 2 weeks after a fifth injection and spleen tissue was removed at necropsy for lymphoblastic transformation assay. Isolated splenic lymphocytes from all five dogs showed increasing proliferation in response to increasing doses of human CEA (Table 2). Lymphocytes from two naive, age-matched dogs showed no evidence of lymphoblastic transformation in response to CEA. The naive dog splenocytes demonstrated robust responses to a mitogen, indicating that these cells were functionally intact. None of the animals demonstrated lymphoproliferation in response to the negative control, BSA. Complete necropsies, clinical hematology and serum chemistries were performed on the five dogs receiving the intramuscular plasmid injections. No gross lesions were noted during the post-mortem examination and histologic examination of muscle, lung, intestine, liver, kidney, spleen and gonad revealed no lesions. Complete

Figure 1 Anti-CEA antibody quantification before immunization and 2 weeks following the 2nd and 3rd boost with a plasmid encoding human CEA. All dogs were injected with 150 ␮g pGT37 DNA on weeks 0, 4, 7 and 13. Results are reported in nanograms 125I-CEA bound per milliliter of serum and are shown for weeks 0 (pre-injection), 9 and 15. The dashed line at 32 ng/ml bound represents the threshold for a positive result. IM, Intramuscular injection; Bio, Biojector 2000 needleless intramuscular injection; ID, intradermal injection; IV, intravenous injection.


867 Table 2 CEA-specific lymphoproliferative responses among dogs receiving intramuscular pGT37 Needle injection Dog 1 Control CEA 30 ␮g CEA 10 ␮g CEA 3 ␮g CEA 1 ␮g BSA PHA

140 ± 50 7200 ± 800 5800 ± 1000 5100 ± 200 3100 ± 400 110 ± 50 72 000 ± 1300

Dog 2 60 ± 10 2200 ± 200 1400 ± 300 800 ± 70 840 ± 200 20 ± 4 81 000 ± 1300

Biojector Dog 3 90 ± 10 4000 ± 200 1600 ± 300 630 ± 80 610 ± 50 100 ± 30 77 000 ± 1900

Dog 4 90 ± 30 5400 ± 500 2800 ± 600 2700 ± 500 2400 ± 100 160 ± 50 86 000 ± 3300

Naive Dog 5

30 ± 5 4100 ± 400 4300 ± 400 1400 ± 200 1700 ± 200 30 ± 4 79 000 ± 1000

Dog 11 35 ± 5 40 ± 5 30 ± 5 40 ± 10 40 ± 10 50 ± 10 70 000 ± 3000

Dog 12 100 ± 30 140 ± 30 80 ± 10 70 ± 10 80 ± 10 90 ± 20 76 000 ± 2000

Dogs received 150 ␮g of a plasmid encoding human CEA on weeks 0, 4, 7, 13 and 23 with splenic mononuclear cells harvested 2 weeks after the last injection. Values are mean c.p.m. ± s.e.m. for quadruplicate wells with values greater than a three-fold increase compared with that animal’s control wells representing a positive response.

blood counts revealed no abnormalities. Serum chemistries showed slight elevations of serum alkaline phosphatase and phosphorus, and a slight depression of globulin from the laboratory’s normal range values in every animal. However, the values from the vaccinated dogs are consistent with those of nonvaccinated dogs less than 6 months of age, and are therefore due to the age of the subject animals.12

Discussion These results indicate that it is possible to detect humoral and cellular immune responses to a tumor-associated antigen gene product in the dog using a nucleic acid vaccine, when the DNA is administered intramuscularly or intradermally. Intravenous delivery of this dose of plasmid DNA produced little or no detectable antibody response. Similar immune responses have been reported in a number of species other than inbred laboratory mice, including rabbits, rats, chickens, fish, cattle and primates.7,13,14 In addition, phase I clinical trials with human subjects indicate that humans respond in a similar fashion.14 All dogs in the intramuscularly and intradermally immunized groups developed significant antiCEA antibody responses. The dogs injected intramuscularly, either with a needle and syringe or the Biojector 2000 (Bioject, Portland, OR, USA), showed the greatest antibody response, while those receiving DNA intradermally had titers approximately 1 log lower. Two of the three dogs receiving DNA intravenously showed brief modest elevations of anti-CEA antibody levels of borderline significance. Therefore, it appears that intramuscular administration was the most efficient means of inducing a humoral immune response in dogs. Subsequent experiments with other nucleic acid immunogens in dogs have shown that the dose of DNA of 150 ␮g is significantly lower than optimal. Doses ranging from 200 ␮g to 800 ␮g have produced protective immunity to parvoviral infection in dogs with a single i.m. injection, while doses of 500 ␮g have been reported as effective in protecting calves against bovine herpes virus infection.11,15 Increasing the dose of DNA might also improve the responses in the animals receiving intravenous and intradermal injections. Recent experiments utilizing unmethylated CpG motifs indicate that increased amounts of DNA not specifically coding for

antigen may have a nonspecific immunostimulatory activity.16 Additionally, both vector backbones and cloned antigen genes possess CpG motifs within their sequence, which when expressed in bacteria, are unmethylated and may serve as internal adjuvants. Another possibility might be that a greater number of immunizations or an increase in dose are required to overcome tolerance to self. While canine CEA has not been cloned, homology studies indicate that canine gene sequences bear a greater similarity to human genes than do murine genes. Thus, immunization of dogs against human CEA represents a greater challenge with respect to overcoming homology than immunization of mice to this antigen. Consequently, the positive immune response seen in dogs is an encouraging indication of the potential of CEA immunization in human patients. The presence of a cellular immune response in the dogs injected intramuscularly is encouraging, both for immunization against tumor-associated antigens and infectious disease agents. The ability to raise both humoral and cellular responses through nucleic acid immunization is one of the significant advantages of this technology. Demonstrating this effect in a large outbred species such as the dog indicates that this technology may be successfully transferred to other outbred mammals, including humans.

Materials and methods Animals Outbred dogs from two litters were used. The dogs were 10 and 15 weeks of age at the start of the experiment and weighed approximately 10 to 15 kg each. The dogs were housed in a HEPA filtered positive pressure ventilated facility to exclude adventitious antigens. Plasmid The plasmid pGT37, which was used for immunization, has been described elsewhere.17–19 Briefly, the plasmid contains the human carcinoembryonic antigen cDNA driven by the immediate–early promoter of the cytomegalovirus (CMV). The bacterial kanamycin resistance gene is present and used for selection of E. coli carrying the plasmid. DNA for injection was purified from bacterial fermentations by anion exchange chromatography (Qiagen, Santa Clara, CA, USA).


868

Injections One hundred and fifty micrograms of the plasmid pGT37 was delivered in 0.5 ml sterile phosphate-buffered saline (PBS). Intramuscular injections were performed in the right quadriceps muscle using a 25 gauge needle in three dogs, and a needleless, CO2-powered injection device (Biojector 2000, No. 3 syringe, Bioject, Portland, OR, USA) in two dogs. Three dogs received pGT37 injected intravenously into the external jugular vein and two dogs received the plasmid as an intradermal bleb, administered with a 27 gauge needle in the skin of the right lower abdominal quadrant. Serum samples were collected from each dog before injection and weekly following injection. The dogs were reinjected with the same dose of DNA and by the same route as their original injection on weeks 4, 7 and 13. The five dogs receiving DNA intramuscularly were reinjected again on week 23 and the animals were killed on week 25 for complete necropsy, during which splenic tissue was obtained for lymphoblastic transformation assays. Anti-CEA antibody assay Antibody response specific for human CEA was quantified using the double antibody technique of Midgley.20,21 Canine sera were diluted 1:10 with phosphate-buffered saline containing 0.1% pig skin gelatin (gel-PBS). Two hundred microliters of each sample were incubated in duplicate with 100 ␮l of 125I-labeled CEA at 1 ␮g/ml (approximately 125 000 c.p.m. per 100 ng) and 700 ␮l gelPBS. After a 16-h incubation at room temperature, 1 ml of rabbit anti-dog IgG heavy and light chain (Bethyl Laboratories, Montgomery, TX, USA) diluted 1:40 in PBS containing 10% polyethylene glycol was added to each sample and incubated at room temperature for 15 minutes to precipitate anti-CEA antibody complexed with 125I-CEA. Samples were centrifuged at 1700 g for 20 min at 4°C and precipitated radioactivity was determined with an automatic gamma counter (ICN Micromedic Systems, Huntsville, AL, USA) interfaced with an IBM System 2 computer. Background nonspecific binding of approximately 2% of the available 125I-CEA was subtracted from c.p.m. bound and nanograms of CEA bound per milliliter of sera was calculated from the known specific activity of the 125I-CEA. A positive response (⬎32 ng/ml) was defined as exceeding 2 standard deviations above the mean of 10 normal dog sera. CEA lymphoblast transformation assay This assay was performed as previously described.17–19 Briefly canine splenic mononuclear cells were isolated by mincing the spleen with a scalpel followed by passage through a 100 ␮m sterile filter to obtain a single cell suspension. Erythrocytes were removed by ammonium chloride lysis. Mononuclear cells were then obtained by Ficoll (Histopaque 1077; Sigma, St Louis, MO, USA) gradient centrifugation and added at 1 × 105 cells per well. Stimulated cells received human CEA purified from human colon cancer liver metastases (Vitro Diagnostics, Littleton, CO, USA) over a range of concentrations (1–30 ␮g/ml); bovine serum albumin (30 ␮g/ml) as a negative control antigen; or PHA at 5 ␮g/ml as a positive control mitogen. The range of CEA concentrations described above provided optimal stimulation in our previous studies of mice immunized with pGT37. Control cells

received media alone. Cells in all wells were cultured in a total volume of 200 ␮l of complete RPMI with 10% fetal calf serum. After 5 days in culture, cells were plated with tritiated thymidine (1 ␮Ci per well) overnight, harvested with a Skatron automatic cell harvester (Skatron, Sterling, VA, USA) and incorporated radioactivity was determined using a Packard Matrix 9600 direct beta scintillation counter (Packard Instrument, Meriden, CT, USA). The Packard Matrix 9600 counts samples dry at approximately five-fold lower efficiency than standard beta scintillation counters requiring scintillation fluid. The mean c.p.m. for quadruplicate wells correlates with the cellular proliferation rate.

References 1 Hellstrom I, Hellstrom KE. Tumor immunology: an overview. In: Bystryn J-C, Ferrone S, Livingston P (eds). Specific Immunotherapy of Cancer with Vaccines. New York Academy of Sciences: New York, 1993, pp 24–33. 2 Hodge JW. Carcinoembryonic antigen as a target for cancer vaccines. Cancer Immunol Immunother 1996; 43: 127–134. 3 Wolff JA et al. Direct gene transfer into mouse muscle in vivo. Science 1990; 247: 1465–1468. 4 Fynan EF et al. DNA vaccines: a novel approach to immunization. Int J Immunopharmacol 1995; 17: 79–83. 5 Donnelly JJ, Ulmer JB, Liu MA. DNA vaccines. Life Sci 1997; 60: 163–172. 6 Pardoll DM, Beckerleg AM. Exposing the immunology of naked DNA vaccines. Immunity 1995; 3: 165–169. 7 Donnelly JJ, Ulmer JB, Liu MA. Immunization with DNA. J Immunol Meth 1994; 176: 145–152. 8 Ertle HC, Xiang ZQ. Genetic immunization. Viral Immunol 1996; 9: 1–9. 9 Robinson HL, Hunt LA, Webster RG. Protection against a lethal influenza virus challenge by immunization with a haemagglutinin-expressing plasmid. Vaccine 1993; 9: 957–960. 10 Raz E et al. Intradermal gene immunization: the possible role of DNA uptake in the induction of cellular immunity to viruses. Proc Natl Acad Sci USA 1994; 91: 9519–9523. 11 Jiang W et al. Nucleic acid immunization protects dogs against challenge with virulent canine parvovirus. Vaccine 1998; 16: 601–607. 12 Wolford ST et al. Effect of age on serum chemistry profile, electrophoresis and thyroid hormones in Beagle dogs 2 weeks to 1 year of age. Vet Clin Pathol 1988; 17: 35–42. 13 Manickan E, Karem KL, Rouse BT. DNA vaccines – a modern gimmick or a boon to vaccinology? Crit Rev Immunol 1997; 17: 139–154. 14 Chattergoon M, Boyer J, Weiner DB. Genetic immunization: a new era in vaccines and immune therapeutics. FASEB J 1997; 11: 753–763. 15 Babiuk LA et al. DNA immunization with bovine herpes virus1 genes. Ann NY Acad Sci 1995; 772: 47–63. 16 Krieg AM et al. CpG motifs in bacterial DNA trigger direct Bcell activation. Nature 1995; 374: 564–549. 17 Conry RM et al. A carcinoembryonic antigen polynucleotide vaccine for human clinical use. Cancer Gene Ther 1995; 2: 33–38. 18 Conry RM et al. A carcinoembryonic antigen polynucleotide vaccine has in vivo antitumor activity. Gene Therapy 1995; 2: 59–65. 19 Conry RM et al. Immune response to a carcinoembryonic antigen polynucleotide vaccine. Cancer Res 1994; 54: 1164–1168. 20 LoBuglio AF et al. Mouse/human chimeric monoclonal antibody in man: kinetics and immune response. Proc Natl Acad Sci USA 1989; 86: 4220–4224. 21 Midgley A. Radioimmunoassay: a method for human chorionic gonadotropin and human luteinizing hormone. Endocrinology 1966; 79: 10–14.