patients: A National Cancer Institute of Canada Clinical Trials Group study. J Clin Oncol 8:385-389, 1990 (20) OKEN MM, CREECH RH, TORMEY DC, ET AL: Toxicity andresponsecrite-

ria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 5:649655, 1982 (21) ETTINGER DS, FrNKELSTEm DM, ABELOFF MD, ET AL: A randomized comparison of standard chemotherapy versus alternating chemotherapy and maintenance versus no maintenance therapy for extensive-stage small-cell lung cancer A phase III study of the Eastern Cooperative Oncology Group. J Clin Oncol 8:230-240, 1990 (22) ADAMS EG, CRAMPTON SL, BHUYAN BK: Effect

of 7-con-O-

(23) NEIL GL, KUENTZEL SL, MCGOVREN JP: Treatment of mouse tumors with 7-con-O-methylnogarol and other analogs of the anthracycline antibiotic, nogalamycin. Cancer Treat Rep 63:1971-1978, 1979 (24) MCGOVREN JP, NEIL GL, DELINGER RH, ET AL: Chronic cardiotoxicity

studies in rabbits with 7-con-O-methymogarol, a new anthiacycline antitumor agent. Cancer Res 39:4349-4855, 1979 (25) BRODER LE, COHEN MH, SELAWRY OS: Treatment of bronchogenic car-

cinoma, n . Small cell. Cancer Treat Rev 4:219-260, 1977 (26) EVANS WK, OSOBA D, FELD R, ET AL: Etoposide (VP-16) and cisplatin:

An effective treatment for relapse in small-cell rung cancer. J Clin Oncol 3:65-71, 1985

methylnogarol on DNA synthesis, survival, and cell cycle progression of Chinese hamster ovary cells. Cancer Res 41:4981-4987, 1981

Judy Kantor, Kari Irvine, Scott Abrams, Howard Kaufman, Judy DiPietro, Jeffrey Schlom* Background: Human carcinoembryonic antigen (CEA) is a 180-kd glycoprotein expressed in human colorectal, gastric, pancreatic, breast, and non—small-cell lung carcinomas. Previous studies have demonstrated enhanced immune responses to other antigens presented with vaccinia virus proteins via a recombinant vaccinia virus construct. In addition, we have developed a recombinant CEA-vaccinia virus construct, designated rV(WR)-CEA, and have demonstrated humoral anti-CEA responses in mice after immunization with that virus. Purpose: The goals of this study were (a) to construct a recombinant CEA-vaccinia vaccine in a less virulent vaccinia strain that is potentially safe and effective for treatment of patients whose tumors express CEA and (b) to evaluate the ability of the recombinant CEA-vaccinia vaccine to prevent and reverse tumor growth in mice and to elicit cell-mediated and humoral anti-CEA immune responses. Methods: Using the New York City strain of vaccinia virus, which is used in smallpox vaccination and is more attenuated for humans than rVfWR), we derived a recombinant CEA-vaccinia construct, designated rV(NYC)-CEA. The ability of this construct to induce antitumor immunity was evaluated in mice receiving subcutaneous injections of murine colon adenocarcinoma cells expressing the human CEA gene. Results: Administration of rV(NYC)-CEA in mice induced strong anti-CEA antibody responses, as well as CEA-specific cell-mediated responses, including delayed-type hypersensitivity, lymphoproliferative, and cytotoxic responses. Vaccination of mice with the rV(NYC)-CEA rendered them resistant to the growth of subsequently transplanted CEA-expressing tumors. Moreover, when mice were vaccinated 7 days after tumor cell injection, tumor growth was either greatly reduced or eliminated. No toxic effects were observed in any of the mice. Conclusion: These studies demonstrate that antitumor activity can be induced with the use of a recombinant CEA-vaccinia virus construct derived from an

1084

attenuated vaccinia strain, and they reveal the range of cell-mediated and humoral responses induced by this recombinant vaccine. [J Natl Cancer Inst 84:1084-1091, 1992]

Vaccinia virus, a member of the poxvirus family, has been safely and successfully used as a live vaccine for the prevention and eradication of smallpox. Recently, a recombinant vaccinia virus expressing human immunodeficiency virus (HIV) envelope glycoprotein has been evaluated in healthy subjects for its safety and ability to elicit humoral and cellular immune responses to HIV (7) as determined using in vitro assays. The use of vaccines for active immunotherapy of cancer has been investigated utilizing several types of immunogens. Most of these studies have used cell extracts (2,3), oncolysates (4,5), and x-irradiated cells (6~) as immunogens. The vast majority of these studies involved melanoma cells. Recently, a recombinant vaccinia virus expressing the human melanoma-associated antigen p97 was developed and, upon immunization, was found to protect mice from challenge with melanoma cells expressing p97 (7). In addition, two monkeys immunized with p97 recombinant virus demonstrated humoral and cell-mediated responses to p97 antigen, including delayed-type hypersensitivity (DTH) (8). Phase I clinical trials are now underway with the recombinant p97 vaccinia virus (Hu SL: personal communication). Recombinant vaccinia virus constructs with other candidate human tumor-associated antigens have also been recently developed and tested in animal models (9,10).

Received December 17, 1991; revised April 3, 1992; accepted April 2 1 , 1992. Laboratory of Tumor Immunology and Biology, Division of Cancer Biology, Diagnosis, and Centers, National Cancer Institute, Bethesda, Md. Present address: H. Kaufman, Boston University Hospital, Boston, Mass. 'Correspondence to: Jeffrey Schlom, Ph.D., National Institutes of Health, Bldg. 10, Rm. 8B07, Bethesda, MD 20892.

Journal of the National Cancer Institute

Downloaded from http://jnci.oxfordjournals.org/ at Florida Atlantic University on January 28, 2015

Antitumor Activity and Immune Responses Induced by a Recombinant Carcinoembryonic Antigen-Vaccinia Virus Vaccine

Vol. 84, No. 14, July 15, 1992

such mice. All mice immunized with the rV(WR)-CEA construct showed reduced growth of the CEA-transduced colon carcinoma cells (20). In view of these initial findings with the rV(WR)-CEA construct, the studies reported here were designed with two end points: (a) to construct and use a vaccinia virus strain that is less virulent than V(WR) in humans and still obtain biologic effects and (b) to better define the host's humoral and cellmediated immune responses to this recombinant CEA-vaccinia virus construct. To this end, we used the New York City Board of Health strain of vaccinia virus, designated V(NYQ, which has been successfully utilized in patients.

Materials and Methods Cells The MC-38 murine colonic adenocarcinoma cell line (22) was supplied by the laboratory of Dr. Steve Rosenberg (National Cancer Institute, Bethesda, Md.). The derivative cell line expressing human CEA, designated MC-38-CEA-2, was made in our laboratory by the transduction of the human CEA gene by the retroviral expression vector pBNC (27). We used the anti-CEA MAb COL-1 and an identical isotype-matched control, the MAb UPC-10 to murine myeloma. Before being transplanted into the animals, cells were analyzed for the expression of CEA. BS-C-1 African green monkey kidney cells were obtained from the American Type Culture Collection, RockviUe, Md. Recombinant Vaccinia Virus The recombinant CEA-vaccinia virus from the WR vaccinia strain, designated rV(WR)-CEA, was derived as described previously (20). The recombinant CEA-vaccinia virus designated rV(NYC)-CEA was produced by homologous recombination of the pSC-11 plasmid containing the human CEA gene into the thymidine kinase gene of the New York City strain of vaccinia virus by a method described by Mackett et al. (23). The New York City Board of Health strain of vaccinia virus was obtained from the American Type Culture Collection (#VR-325). The plasmid was provided by Dr. B. Moss (National Institute of Allergy and Infectious Diseases, Bethesda, Md.) and contained the Escherichia coli Lac Z gene under the control of the vaccinia virus late-promoter p-11. Use of this plasmid provided a selection method for obtaining recombinant viral particles. Analysis of the recombinant virus was performed as described by Kaufman et al. (20). We detected the expressed protein by Western blot analysis using MAb COL-1. Virus for inoculation was grown in spinner cultures of HeLa cells, directly pelleted by centrifugation, and purified over a 20%-40% sucrose gradient (23). In Vivo Experiments Female C57BLV6 mice were obtained from the National Cancer Institute, Frederick Cancer Research and Development Center (Frederick, Md.). Six- to 8-week-old animals were given a subcutaneous injection of 2 x 105 MC-38-CEA-2 cells or of 2 x 105 MC-38 cells (which do not express CEA). Ten animals per group were inoculated by tail scarification with either 10 jtL of purified virus containing 1 x 107 plaqueforming units (PFU) of wild-type vaccinia virus (designated V-NYC) or recombinant rV(NYC)-CEA. Three immunizations ARTICLES 1085

Downloaded from http://jnci.oxfordjournals.org/ at Florida Atlantic University on January 28, 2015

Limited studies have been conducted to examine the ability of these recombinants to elicit active immune responses in carcinoma patients. One recent study, using intact x-irradiated colon carcinoma cells as immunogen, has shown some antitumor activity (77). A potential target for active immunotherapy, which may be useful for the treatment of several carcinoma types, is carcinoembryonic antigen (CEA) (72). We focused on this antigen because it is a biochemically well characterized glycoprotein, having a molecular mass of 180 kd, and is expressed at high levels on the surface of rumors of the gastrointestinal tract, i.e., carcinomas of the colon, rectum, stomach, and pancreas (73). CEA is also expressed on many mammary carcinomas (14) and on lung adenocarcinomas (75). Complementary DNA clones of the CEA gene have been isolated, and nucleotide sequences have been determined (76). The immunogenic nature of CEA in humans is unclear at this time. Studies have shown that sera of normal subjects and/ or of cancer patients contain CEA-immunoglobulin immune complexes (77). Many of these studies, however, have been conducted with polyclonal anti-CEA sera so the immune complexes detected may be to CEA-related antigens (18). There is no evidence of cell-mediated responses to CEA in humans. It thus appears that the copresentation of CEA with a strong immunogen such as vaccinia virus might represent a logical approach to inducing an anti-CEA response for tumor immunotherapy. Vaccinia virus is highly immunogenic and has been shown to stimulate both humoral and cell-mediated responses (79). Therefore, immunization with a recombinant CEAvaccinia virus could potentially provide a strong stimulus to the immune system. We (20) have recently reported the development of a recombinant vaccinia virus expressing human CEA [designated rV(WR)-CEA] as an initial step in the development of an agent with the potential to elicit a specific active immune response in patients whose tumors express CEA. The recombinant construct utilized was the WR strain of vaccinia virus, which is widely used for rodent model studies. We demonstrated, by Southern blot analysis and polymerase chain reaction, that a complete copy of the CEA-coding sequence was present in the recombinant CEA-vaccinia virus. This recombinant virus was also shown to express, on the cell surface of infected cells, a protein product recognized by the monoclonal antibody (MAb) COL-1, which is specific for CEA (75). Immunization of mice with rV(WR)-CEA was also shown to induce the production of antibodies to CEA (20). To determine if a recombinant CEA-vaccinia virus construct could elicit an antitumor immune response, an animal model needed to be developed. CEA is not expressed on any known animal tumors. Since an intact host immune response is necessary for these types of studies, human CEA-expressing tumors transplanted into athymic mice would not be suitable. Thus, an experimental model was developed (27) in which the human CEA gene was transduced into the established murine colon carcinoma cell line MC-38 (22). The CEA-transduced cells were shown to express human CEA on their cell surface, using a panel of anti-CEA MAbs, and the level of CEA expression was shown to be comparable to that of human colon carcinomas. The CEA-transduced and non—CEA-transduced MC-38 cells were both shown (Hand PH, Robbins PF, Salgaller ML, et al.: manuscript in preparation) to grow in immune competent C57BL/6 mice and led to the death of all

were given 14 days apart and according to the schedules described for each study. Animals were checked weekly for the presence of tumors. Tumors were measured by caliper in two dimensions, and the volumes were calculated using the formula (width2 x length)/2.

tracting the swelling of the footpads given the control MC-38 cells from the swelling of footpads challenged with the CEAexpressing cells. Results were expressed in " m i l , " which is equivalent to 0.0254 mm. The mean and standard deviation of the incremental swelling were calculated for each group. P values were determined using Student's t test of significance.

Vaccination With rV(NYC)-CEA Mice were vaccinated three times at 14-day intervals with 1 X 107 PFU of either rV(NYQ-CEA or V-NYC by tail scarification; 7 days after the last vaccination, 2 X 105 tumor cells were injected subcutaneously. Treatment of Tumor-Bearing Mice

Quantitation of CEA Antibody The humoral response to human CEA was determined by inoculating 6-week-old female C57BL/6 mice (10 mice per group) three times at 2-week intervals by tail scarification with 10 n-L of purified virus containing 1 X 10' PFU of V-NYC or rV(NYC)-CEA. Serum samples were collected at weekly intervals and analyzed for the presence of antibodies to CEA by the enzyme-linked immunosorbent assay (ELISA). Microtiter plates were coated overnight at 37 °C with 100 ng per well of purified CEA (International Enzymes, Inc., Fallbrook, Calif.). We used 5% bovine serum albumin in phosphate-buffered saline (PBS) to prevent nonspecific binding, followed by a 1-hour incubation with fivefold dilutions of test sera (starting at 1:50) or with the anti-CEA-specific MAb COL-1 (starting at 100 ng/50 M-L) (75). Antibodies immobilized on the wells were detected with goat antimouse IgG (Becton Dickinson, San Jose, Calif.) conjugated to horseradish peroxidase. The resulting complex was detected by the chromogen, o-phenylenediamine. Reactivity was measured in an ELISA microplate autoreader EL310 (Bio-Tek Instruments Inc., Winooski, Vt.) at an optical density of 490 nm. Titers were determined as the dilution factor when the optical density was equivalent to 0.5.

C57BL/6 mice were inoculated with V-NYC or rV(NYC)CEA by tail scarification, as described. Spleens were removed 28 days after the third exposure and were mechanically dispersed through stainless-steel mesh screens to isolate a singlecell suspension. Erythrocytes and dead cells were removed by centrifugation over a Ficoll-Hypaque gradient (density = 1.119 g/mL) (Sigma Chemical Co., St. Louis, Mo.). Subsequently, adherent cells were depleted from the mononuclear population by incubation and passage over nylon wool columns (Robbins Scientific Corp., Sunnyvale, Calif.), resulting in a T-cell-enriched fraction. Depletion of the adherent cells from the responder population before the lymphoproliferation assay helped to reduce the background autologous responses. T lymphocytes were washed and resuspended to the appropriate concentration in culture medium, which consisted of RPMI-1640 medium supplemented with 15 mW HEPES (pH 7.4), 5% heatinactivated fetal calf serum, 2 mAf L-glutamine, 0.1 mAf nonessential amino acids, 1 mM sodium pyruvate, 100 U/mL penicillin, 100 ^g/mL streptomycin (all from GIBCO BRL, Gaithersburg, Md.), and 5 X 10~5 M B-mercaptoethanol (Sigma Chemical Co.). T lymphocytes (1 X 105/well) were incubated in the presence of irradiated, normal syngeneic splenocytes (5 x 103/ well) as antigen-presenting cells with and without various stimuli. Stimuli included concanavalin A (Con A, Sigma Chemical Co.), UV irradiation inactivated V-NYC, purified CEA (International Enzymes Inc. or Vitro Diagnostics, Littleton, Colo.), or purified ovalbumin (Sigma Chemical Co.). Incubation was performed in 96-well, flat-bottom plates (Costar Corp., Cambridge, Mass.) for up to 3 days (Con A) or 5 days (antigens). Cultures were pulsed with [3H]thymidine (1 u,Ci/well) (Du Pont/NEN Research Products, Wilmington, Del.) for the final 18-24 hours of incubation. Cells were collected with a PHD cell harvester (Cambridge Technology, Cambridge, Mass.), and incorporated radioactivity was measured by liquid scintillation spectroscopy (LS 3801 counter; Beckman Instruments, Duarte, Calif.).

DTH Testing Cytotoxicity Assay In vivo, DTH provides a measure of cell-mediated immunity and, therefore, suggests a role for T lymphocytes in the induction and/or effector pathways. DTH assays were performed in mice by the measurement of footpad swelling. Ten mice per group were immunized two or three times at 14-day intervals by tail scarification with either 107 PFU of V-NYC or rV(NYC)-CEA. Six days after the last vaccination, 5 X 105 irradiated (20000 rad) MC-38 cells in 20 JJLL of PBS were injected into the left hind paw, and 20 M-L of PBS containing 5 X 10* x-irradiated CEA-transduced MC-38-CEA-2 cells was injected into the right hind paw. The thickness of the footpads was measured 48 hours later with a Mitutoyu Digimatic Micrometer (Mitutoyu Measuring Instruments, Maplewood, N.J.) in a single blind fashion. DTH was calculated by sub1086

Mice were inoculated with 1 X 107 PFU of V-NYC or rV(NYC)-CEA. After 14 days, they were given a second inoculation as a booster; 5 days later, their spleens were removed and processed as described above for the isolation of T lymphocytes. In parallel, MC-38 and MC-38-CEA-2 tumor cells were prepared for use as targets in a standard 5lCr release assay. Briefly, tumor cells (1-2 x 106 cells) were radiolabeled with 200 u,Ci of Na 2 51 CrO 4 (Amersham Corp., Arlington Heights, 111.) for 60 minutes at 37 °C, followed by thorough washing to remove unincorporated isotopes. T cells and targets (1 x lCrVwell), both resuspended in culture medium, were then combined at various effector-to-target ratios in 96-well, U-bottom plates (Costar Corp.). The plates were centrifuged at Journal of the National Cancer Institute

Downloaded from http://jnci.oxfordjournals.org/ at Florida Atlantic University on January 28, 2015

Studies were conducted to determine if inoculation of the rV(NYC)-CEA construct could inhibit the growth of established tumors, i.e., tumors transplanted 7 days prior to vaccinia virus treatment. Mice were innoculated with 1 x 10' PFU of V-NYC or rV(NYC)-CEA by tail scarification. As with the experiments described above, non-CEA-transduced and CEAtransduced rumors were used.

Lymphoproliferation Assay

lOOg for 5 minutes to initiate cell contact and were incubated for 4 or 16 hours at 37 °C with 5% CO 2 . After incubation, supernatants were collected, using a Supernatant Collection System (Skatron, Inc., Sterling, Va.), and radioactivity was quantitated in a gamma counter (Beckman Instruments). Spontaneous release of 51Cr was determined by incubation of targets in the absence of effectors, while maximum or total release of 51Cr was determined by incubation of targets in 0.1% Triton X-100. Percentage of specific release of 5ICr was determined by the following equation: percent specific release = [(experimental — spontaneous)/(maximum — spontaneous)] x 100.

1 2 3 4 5 6 200-

m

• 97.4 —

Western Blotting

Results Western Blotting The molecular masses of the CEA product expressed by the rV(NYC)-CEA and rV(WR)-CEA constructs were determined by Western blot analysis, using the anti-CEA MAb COL-1 (Fig. 1). As controls, we used the 180-kd product of purified human CEA (lane 1) and CEA detected in an extract of the established human colon carcinoma cell line GEO (lane 2). Extracts of cells infected with V-NYC, rV(NYC)-CEA, V-WR, and rV(WR)-CEA were also transferred to membranes and analyzed with COL-1. As seen in lane 4, the cells infected with rV(NYC)-CEA expressed a 90-kd product, while the cells infected with wild-type V-NYC or V-WR showed no CEA present (lanes 3 and 5, respectively). Cells infected with the rV(WR)-CEA, however, expressed 70-, 90-, and 180-kd products that reacted with COL-1 (lane 6).

68 —

!

Fig. 1. Western blot analysis of CEA gene products produced by the recombinant vaccinia virus. A detergent lysate from vaccinia virus-infected BS-C-1 cells was subjected to electrophoresis on a 4%-12% gradient sodium dodecyl sulfate-polyacrylamide gel and transferred to a polyvinylidene difluoride membrane as described in the "Materials and Methods" section. Antigens were detected with the CEA-specific MAb COL-1, followed by l25I-labeled goat antimouse immunoglobulin. Lane 1 shows purified human CEA (1 |xg; The Binding Site, Inc., San Diego, Calif.); lane 2 shows an extract of the human adenocarcinoma cell line GEO (20 \ig). The other lanes show extracts of BSC-1 cells infected with V-NYC (lane 3), rV(NYC)-CEA (lane 4), V-WR (lane 5), and rV(WR)-CEA (lane 6). Numbers at the left side indicate the sizes (kd) and positions of migration of three molecular mass markers. Lanes 1 and 2 were exposed for 4 hours, and lanes 3-6 were exposed for 24 hours.

Two types of controls were used. First, the MC-38 colon carcinoma cells with and without the transduced human CEA gene were used to determine if antitumor effects were directed against CEA. Second, wild-type V-NYC was also used as an immunogen to determine if the immune responses generated were the consequence of the inserted human CEA gene. As seen in Fig. 2, A and B, neither the wild-type nor the recombinant vaccinia construct, when used as an immunogen, conferred any protection against growth of transplanted nonCEA-transduced tumor cells. Tumors from all 10 mice in each group grew rapidly at approximately the same rate. Nontransduced and CEA-transduced MC-38 tumors grew at similar rates in control animals receiving no vaccinia inoculation and at the same rate as tumors growing in mice which had received wild-type vaccinia (V-NYC) inoculations. Fig. 2, C and D, compare the efficacy of the V-NYC with that of the rV(NYC)-CEA to inhibit the transplantation of the CEA-transduced colon carcinoma cells. As seen in Fig. 2, C, in mice vaccinated with the wild-type V-NYC, eight of 10 transplanted tumors grew rapidly, and eventually all 10 tumors grew. In contrast, no tumors grew in any of the 10 mice vaccinated with the rV(NYC)-CEA construct. Furthermore, these mice immunized with rV(NYC)-CEA remained tumor free for 120 days following their first tumor challenge; at day 120, they were challenged with 1 x 106 CEA-transduced tumor cells, and again they remained tumor free throughout an additional 120-day observation period. No toxic effects due to the administration of rV(NYC)-CEA or V-NYC were observed.

Vaccination With rV(NYC)-CEA Studies were conducted to determine if the rV(NYC)-CEA construct, derived from the more attenuated NYC vaccinia strain, could prevent the establishment of tumor transplants. Vol. 84, No. 14, July 15, 1992

Treatment of Tumor-Bearing Mice Studies were conducted to determine if inoculation of the rV(NYC)-CEA construct could inhibit the growth of estabARTICLES 1087

Downloaded from http://jnci.oxfordjournals.org/ at Florida Atlantic University on January 28, 2015

A confluent monolayer of BS-C-1 cells was infected with V-NYC or rV(NYC)-CEA for 24-48 hours at a multiplicity of infection of 30 PFU per cell. The cells were then scraped and centrifuged for 5 minutes at 1800g at 4 °C. The supernatant was discarded, and the cell pellet was resuspended in 200 \hL of lysis buffer (100 mM Tris-HCl [pH 8.0], 100 mM NaCl, 0.5% Nonidet P40, and 0.2 mM phenylmethane sulfonyl fluoride) and set on ice for 10 minutes. Nuclei were separated by centrifugation for 10 minutes at 4 °C. The supernatant was removed, and an aliquot was boiled for 5 minutes in sodium dodecyl sulfate sample buffer as described by Laemmli (24) and applied to a 4%-12% gradient sodium dodecyl sulfatepolyacrylamide gel with a 3% stacking gel (NOVEX, Encinitas, Calif.). Proteins were transferred to a polyvinylidene difluoride membrane (0.45-p.m pore size) (Millipore Corporation, Bedford, Mass.), using a semidry blotter (Bio-Rad Laboratories, Richmond, Calif.) in 20% methanol, 0.039 M glycine, and 0.048 M Tris for 1 hour at 1 W/30 cm2. Following transfer, membranes were blocked in 4% Blotto (Instant Nonfat Dried Milk; Carnation Co., Los Angeles, Calif.) in PBS and incubated for 1 hour with the MAb COL-1 (1 (ig/mL). Bound antibody was detected using l25I-labeled goat antibody to mouse IgG (100 000 cpm/mL).

-

5000

Tumor: CEA neg V-NYC

Tumor: CEA neg rV(NYCKEA

4000 3000

Tumor: CEA po» rV(NYC)-CEA

n«10 14

21

28

35

42

49

0

7

14

21

28

35

42

49

Days Post Tumor Transplant

lished tumors, i.e., tumors transplanted 7 days prior to recombinant vaccinia virus treatment. As seen in Fig. 3, A and B, the growth rates of the MC-38 carcinoma cells (non-CEAtransduced) were similar regardless of whether the V-NYC or rV(NYC)-CEA constructs were used for treatment. Similar tumor growth rates were also seen in mice bearing the CEAtransduced MC-38 cells when treated with the wild-type V-NYC (Fig. 3, C). In contrast, greatly reduced tumor growth was seen in all 10 tumor-bearing mice treated with the rV(NYC)-CEA construct (Fig. 3, D). Moreover, three animals in this group that failed to develop tumors for 4 months were again challenged with 1 X 106 CEA-transduced MC-38 cells and remained tumor free through an additional 120-day observation period. Non-CEA-transduced MC-38 tumors implanted at the same time on the contralateral side grew at the site. No toxic effects due to the administration of the rV(NYC)-CEA were observed. . . Immune Responses Anti-CEA antibody response. Studies were undertaken to define the type of immune responses elicited by the administration of the recombinant vaccinia construct. As seen in Table 1, serum titers ranged from 1:700 to 1:5250 (average, 1:2255) in mice administered rV(NYC)-CEA, while titers at or below 1:200 (average *£l:82) were observed in all 14 mice inoculated with V-NYC and in all 24 preinoculation sera. All sera in the preinoculation groups and in groups inoculated with either vaccinia construct were also negative or weakly positive for reactivity to ovalbumin (Table 1), with the exception of one mouse 1088

with a titer of 1:750. The dynamics of increase in antibody titer following inoculation with rV(NYC)-CEA were also monitored. Following the first inoculation, there was a modest rise in anti-CEA titer, which greatly increased after the second and third inoculations with rV(NYC)-CEA (data not shown). DTH. Cell-mediated immune responses to the rV(NYC)CEA construct were measured using assays for DTH, lymphoproliferation, and cytotoxicity. DTH reactions in mice inoculated with rV(NYC)-CEA or V-NYC were measured after challenge with x-irradiated tumor cells. One week after the last vaccinia inoculation, one footpad was given an injection of irradiated nontransduced MC-38 cells, and the other footpad was given the irradiated CEA-transduced MC-38 cells. As seen in Fig. 4, little or no differences between footpads were noted in mice given injections of control PBS solution; from these results, an arbitrary value of 17 mil (see "Materials and Methods" section) was used as a normal cutoff. Similar negative results were obtained when mice were inoculated with the V-NYC construct twice. Two of 10 mice given two injections of rV(NYC)-CEA showed some differential swelling in the footpad that was given an injection of CEA-transduced tumor cells. Mice given three injections of the V-NYC construct showed little or no DTH response to the CEA-transduced tumors, while the majority of mice (14 of 20) given three injections of the rV(NYC)-CEA construct demonstrated a differential DTH reactivity to the CEA-transduced tumor cells. The difference in DTH results following three injections of rV(NYC)-CEA versus the V-NYC construct was statistically significant (P