Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Malignant mesothelioma (MM) is an aggressive tumor of the mesothelium associated with previous exposure to asbestos, and has become a common and difficult clinical problem in many parts of the world. The latency period of this disease following initial exposure to asbestos is between 15 and 40 yr, and the frequency of the disease is expected to increase (1, 2), with the total number of new cases peaking by early in the 21st century (3). The mean survival time from diagnosis of MM averages 9 mo (4). Unfortunately, standard therapies, including surgery, radiotherapy, and chemotherapy, have not altered the prognosis in this malignancy. An urgent need therefore exists to investigate new types of therapies that may aid in the eradication of MM.

We have derived considerable data, both laboratory-based (5) and during clinical trials (8), supporting the notion that MM is susceptible to immunologic intervention. We have previously shown, in experiments with a murine model of MM that mimics the human counterpart of the disease in terms of histologic, cell biologic, and immunobiologic features (9, 10), that the systemic administration of interleukin (IL)-12 has profound antitumor properties against a murine MM in vivo (11). IL-12 has a number of immunostimulatory effects that are potentially useful in cancer therapy, including enhancement of natural killer (NK) and cytotoxic T-lymphocyte (CTL) function (12), promotion of the development T-helper (Th1)- type cells (15), and induction of interferon (IFN)- and tumor necrosis factor (TNF) (16, 17). In fact, systemic administration of IL-12 has profound antitumor activity in vivo in a number of other models (18).

The systemic administration of IL-12, however, requires the use of supraphysiologic concentrations of this cytokine in order to produce adequate levels at the tumor site. Unfortunately, such high levels can cause severe toxicity (25), and in order to circumvent the complications of systemic administration of IL-12, we have therefore shifted our focus toward generating local intratumoral IL-12 secretion. We have shown that the injection of recombinant IL-12 into MM has profound antitumor effects (11). An important observation in these studies, however, was that IL-12 had to be present continuously in the tumor; cessation of injections was followed by tumor regrowth.

Because repeated injection of IL-12 is not feasible in humans, we turned our attention to gene transfer approaches to achieve prolonged cytokine production within the tumor in MM. With use of such an approach, it was crucial to determine whether the continuous, local, paracrine production of IL-12 led to a local reduction in tumor growth in addition to a systemic antitumor effector T-cell response with a concomitant response to untransfected tumor tissue at a distal site. Several gene-transfer techniques have been used to generate paracrine secretion of IL-12. In vivo transduction of other tumors with recombinant vaccinia virus or adenovirus encoding the IL-12 genes has been shown to result in local production of IL-12 and prevention of tumor growth (26). An alternative strategy, using genetically engineered fibroblasts to secrete IL-12 at the tumor site, also resulted in inhibition of tumor growth (29, 30).

Although the foregoing studies highlight the beneficial effects of local production of IL-12, potential limitations to such therapies exist. Cytokine-gene delivery through viral vectors can be hampered if a strong antiviral immune response is induced that results in elimination of the virus. Genetically engineered fibroblasts may only be useful for treating tumors that share target antigens with the transfected fibroblast. In order to develop a clearly defined system in which the effects of local production of IL-12 could be characterized, and to avoid the problems associated with the approaches previously described, we genetically engineered a solid, nonimmunogenic MM tumor cell to produce IL-12. This study shows that local production of IL-12, generated by an MM-IL-12 transfectant, induced systemic immunity that could prevent growth of the parental tumor and, importantly, that CD8⁺ effectors were generated which exhibited antitumor effects against a second, untransfected tumor growing at a distal site.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Mice

BALB/c and BALB/c-nu/nu mice were obtained from the Animal Resource Center (Perth, Western Australia) and maintained under standard conditions at the animal facility of the Department of Medicine, University of Western Australia.

Cell Lines

The establishment of the BALB/c mouse-derived MM tumor cell line AB1 has been described previously (23). Cell lines were maintained in RPMI 1640 medium (GIBCO, Grand Island, NY) supplemented with 5% fetal bovine serum (GIBCO), 20 mM 4-(2-hydroxyethyl)-1-piperazine- N-2-ethanesulfonic acid, 200 mM L-glutamine (GIBCO), 0.05 mM mercaptoethanol, 100 µg/ml gentamicin, and 120 µg/ml of penicillin. Additionally, cell lines transfected with the bacterial neomycin phosphotransferase gene were maintained in medium containing G418 (Geneticin; GIBCO; 400 µg/ml). All cell cultures were grown at 37°C in a 5% CO₂ humidified atmosphere.

DNA Constructs and Transfection Procedures

Murine complementary DNA (cDNA) encoding IL-12 subunits p35 and p40 in the bluescript SK+ plasmid were kindly provided by Ueli Gubler (Hoffman LaRoche, Nutley, NJ). The p40 and p35 subunits were subcloned into the expression vectors pHApr-1-neo (kindly provided by Dr. J. Allison, University of California, Berkeley, CA) and kCMVintPolyli, respectively. The MM tumor cell line AB1 was cotransfected with both expression vectors, using cationic lipid 1,2-dioleyl-3-trimethylammonium propane (DOTAP) (Boehringer Mannheim, Mannheim, Germany). For this procedure, 3 × 10⁵ tumor cells were seeded into six-well plates and allowed to recover for 24 h. DNA (10 µg of p35 and 10 µg of p40) was mixed with 80 µg of DOTAP, and tumor cells were incubated with the DNA/ transfection mixture for 24 h in OPTI-MEM medium (GIBCO). Cells were then trypsinized, transferred into 80-cm² flasks (Becton Dickinson Labware, Franklin Lakes, NJ), and maintained in normal medium for 48 h before the beginning of selection with G418 at 400 µg/ml for 14-21 d. Resistant colonies were pooled and cloned by limiting dilution. Control AB1 transfectants (AB1-Neo) were transfected with the pHApr-1-neo plasmid, which contained the bacterial neomycin phosphotransferase gene only. The AB1-IL-12 clone used in the study produced 10 ng of IL-12/10⁶ cells/48 h.

Screening and Quantification of IL-12 Production

The IL-12 transfectants were screened for the production of biologically active IL-12 through use of an IL-12 capture bioassay described elsewhere (31). Briefly, 96-well flat-bottom tissue culture plates (Becton Dickinson Labware) were coated with rat monoclonal antimurine IL-12 antibody C15.1 (100 µl/well, 20 µg/ml in 0.1 M bicarbonate/carbonate buffer, pH 9.5, for 24 h at 4°C). Hybridomas C15.1 and C17.8 were kindly provided by Dr. G. Trinchieri of The Wistar Institute, Philadelphia, PA. Plates were washed with phosphate-buffered saline (PBS) and blocked with 5% fetal calf serum (FCS)/PBS (200 µl/well; 1 h at room temperature [RT]). After several washes, supernatants from IL-12 transfectants or IL-12 reference standards (a generous gift of Dr. S. Wolf of the Genetics Institute, Cambridge, MA) were added (100 µl/well; 5 h at 37°C) and the plates were washed several times before BALB/c splenocytes were added (5 × 10⁵ cells/well). The production of interferon (IFN)- after 48 h was measured with a commercially available enzyme-linked immunosorbent assay (ELISA) for IFN- (PharMingen, San Diego, CA) as instructed by the manufacturer. The production of IL-12 was quantified through a commercially available sandwich-type ELISA (PharMingen). Supernatant from 10⁶ cells was collected after a 48-h incubation period and stored at 20°C until assayed. Briefly, plates were coated with anti-IL-12 antibody (Red-T/G297-289; PharMingen) (8 µg/ml at 4°C overnight) and blocked with 10% FCS/ PBS (1 h at 37°C) after which IL-12 standards and supernatants were added for an overnight incubation at 4°C. Monoclonal anti-IL-12 antibody C17.8 was biotinylated and used at 4 µg/ml (1 h at RT), followed by an incubation with streptavidin-alkaline phosphatase (1:1,000; Dako, Carpinteria, CA; 0.5 h at RT). Between each incubation step, plates were washed several times in 0.5% Tween/PBS. After addition of p-nitrophenylphosphate, plates were developed for 15-45 min and absorbance was read at 405 nm.

In Vitro Depletions

Lymphocytes (5 × 10⁷ cells/ml) were incubated on ice for 30 min with supernatant containing anti-CD4 or anti-CD8 antibody (obtained from hybridomas RL172 and 3.168, respectively, and kindly made available by Dr. W. Heath, Melbourne, Victoria, Australia). For lysis, the labeled lymphocytes were washed, resuspended at 5 × 10⁷ cells/ ml, and incubated for 30-45 min at 37°C with a 1:10 dilution of rabbit complement (C-Six Diagnostics Inc., Mequoun, WI). Analysis with a FACScan flow cytometer (Becton Dickinson) revealed that the lysis procedure depleted more than 98% of each lymphocyte subset.

Tumor Challenge Experiments

Exponentially growing tumor cell lines were harvested by brief trypsinization, washed, and resuspended in PBS. BALB/c mice were injected in the left flank with transfected or parental AB1 cells. The tumor inoculum and time of challenge or rechallenge varied in different experiments, and are specified in the figure legends. In the Winn assay, lymphocytes from the draining lymph nodes (LNs) of mice primed with AB1 or with the IL-12 transfectant, mixed at various ratios with 1 × 10⁵ AB1 cells, were injected into naive BALB/c recipients. Tumor size was measured every 2-4 d with calipers, and mice were killed when the tumor reached 10 mm in diameter. The final measurement, made on mice that were killed, was carried over to the subsequent time points, and data are presented as mean tumor area, which is the product on the two largest perpendicular diameters of a tumor.

Immunohistologic Staining

Tissues from various sites were removed, placed in compound-embedding medium, and snap frozen by placing the embedding mold on dry ice. Ten-micrometer-thick sections were cut, collected on poly-L-lysine-coated slides, and allowed to air dry for 1 h. Slides were stored at 4°C (for a maximum of 2 d) before staining. Routine hematoxylin staining was performed on tissue. Before immunostaining, sections were fixed with 1% paraformaldehyde (5 min), blocked with a biotin/avidin kit (Vector Laboratories, Burlingame, CA), and further blocked 1% bovine serum albumin (15 min). Sections were then incubated with primary antibody, rat anti-CD4 (GK1.5) or rat anti-CD8 (53.6.72) antibody for 1 h, with secondary antibody (biotinylated rabbit antirat-Ig; Jackson Immuno-Research Laboratories, Inc., West Grove, PA) for 30 min, with streptavidin/horseradish peroxidase (Dako) for 30 min, and with diaminobenzidine/H₂O₂ (Sigma, St. Louis, MO) for 5-10 min. Slides were washed for 5 min in PBS between each incubation step.

Anti-CD3 Stimulation

Ninety-six-well flat-bottom plates (Falcon) were precoated with 100 µl of rat antimouse CD3 antibody KT3.2 (1 mg/ ml in PBS) overnight at 4°C. Single-cell suspensions of LN cells were added to a final number of 2 × 10⁵ cells/well. Supernatants were collected after 48 h and stored at 20°C until assayed with a commercially based ELISA for IFN- (PharMingen).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Paracrine Secretion of IL-12 Renders MM-IL-12-Transfectant Immunogenic

The AB1 MM cell line was transfected with the IL-12 genes, and two clones were identified that subsequently failed to form tumors in normal mice (AB1-IL-12-2A and AB1- IL-12-1.4), even when given at relatively large doses (10⁶ cells). This report details the immunity induced by cells of one of these two clones, AB1-IL-12-2A, henceforth designated AB1-IL-12, which produced 10 ng IL-12/10⁶ cells/48 h.

Initial experiments investigated whether the transfection procedure altered the tumorgenicity of this clone. As shown in Figure 1, nude mice, which are devoid of T cells, developed AB1-IL-12 tumors at a similar, albeit slightly slower, rate than that of the control parental line transfected with the neomycin resistance gene only (AB1-Neo). When the tumor cells were cultured in vitro, the doubling time for AB1-IL-12 cells, of 42.36 ± 3.6 h, was almost double that of the AB1-Neo transfectant cells, at 24.3 ± 2.6 h. However, there was no direct relationship between in vitro and in vivo growth rates when 10 IL-12 transfectants were compared; thus, for example, another IL-12 transfectant, AB1-IL-12-4, had the same in vitro doubling time as AB1-Neo cells, but grew more rapidly in nude mice (data not shown). Nevertheless, these data indicate that the IL-12 transfectant retained tumorgenicity and was not rejected by non-immune-specific mechanisms.

View larger version (18K): [in this window] [in a new window]   Figure 1.   AB1-IL-12 transfectant forms tumors in nude mice. BALB/c-nu/nu mice were inoculated subcutaneously with AB1 tumor cells (106) that were transfected with the neomycin resistance gene only (AB1-Neo; n = 4) or with the genes encoding IL-12 (AB1-IL-12; n = 5), and tumor growth was monitored. The data shown are from a single experiment that is representative of three different experiments.

To determine whether AB1-IL-12 cells could form tumors in immunocompetent hosts, we inoculated BALB/c mice with 10⁶ cells of either AB1-IL-12 cells or the control AB1 cell line. AB1-IL-12 cells failed to form progressive tumors, whereas all mice challenged with the control AB1-Neo cell line developed rapidly growing tumors by Day 35 after inoculation (Figure 2A). This lack of tumor growth was dependent upon the production of IL-12, since administration of neutralizing IL-12 antibodies reversed the phenomenon (Figure 2B). Interestingly, up to 60% of mice inoculated with AB1-IL-12 cells developed small nodules, less than 1 mm in diameter and unmeasurable with calipers, at the site of tumor injection on Days 7-14, but these resolved within several days of occurrence. Immunohistologic analysis of these nodules on Day 10 after tumor inoculation revealed sparse tumorlike tissue that was heavily infiltrated with CD4⁺ and CD8⁺ cells (Figures 3A and 3B). This pattern of lymphocytic infiltration was in stark contrast to that observed in the parental tumor at the same stage, in which few if any T cells were observed (Figures 3C and 3D). A paucity of lymphocytic infiltration is characteristic of MM throughout tumor growth. These data, in association with those obtained in nude mice, imply that the IL-12 transfectant prevents tumor growth through T-cell-dependent mechanisms. In addition, animals inoculated with the IL-12 transfectant did not exhibit any signs of ill health, which have been associated with systemic administration of IL-12.

View larger version (16K): [in this window] [in a new window]   Figure 2.   AB1-IL-12 transfectants fail to grow in immune competent mice, and this is directly due to IL-12 production. BALB/c mice were inoculated subcutaneously with AB1 tumor cells (106) that were transfected with the neomycin resistance gene only (AB1-Neo; n = 10) or with the genes encoding IL-12 (AB1-IL-12; n = 10), and tumor growth was monitored (A). The data shown are from a single experiment that is representative of three different experiments. Two groups of BALB/c mice (n = 5) were inoculated with 106 AB1-IL-12 cells; the experimental group was given 50 µl of monoclonal anti-IL-12 antibody (C17.8) intraperitoneally and subcutaneously at the site of tumor injection on Days 0, 2, 5, 9, 12, 18, 25, and 32. Control mice received saline only (B).

View larger version (156K): [in this window] [in a new window]   Figure 3.   T-cell infiltrate of parental AB1 and AB1-IL-12 tumors. BALB/c mice were inoculated subcutaneously with 106 AB1-IL-12 transfectant cells alone (A, B), 106 cells of parental AB1 alone (C, D), or 106 cells of parental AB1 and AB1-IL-12 at different sites (E, F ). (A, C, and E ) Staining with anti-CD4 antibody. (B, D, and F ) Staining with anti-CD8 antibody.

AB1-IL-12 Induces Long-Term Immunity against Parental Tumor

The association between T cells and protection against tumor growth led us to investigate whether mice inoculated with AB1-IL-12 showed long-lasting and cross-reactive immunity to the parental tumor. Animals previously inoculated with 10⁶ AB1-IL-12 cells were rechallenged 3 mo later in the opposite flank with 10⁵ cells of parental AB1, and tumor growth was monitored. Mice that had previously rejected AB1-IL-12 subsequently showed either no growth of the parental tumor (60% of rechallenged recipients) or delayed tumor growth (i.e., palpable tumor at Day 60 as compared with Day 35 in naive animals) (Figure 4). These data indicate that prior immunization with AB1-IL-12 induces a memory response that has the capacity to inhibit or eradicate the parental tumor, further supporting the notion that IL-12 production by MM cells induces a tumor-specific immune response.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Mice previously challenged with AB1-IL-12 were protected from subsequent parental tumor. Naive BALB/c mice (n = 10) or mice that had been challenged with 106 cells of AB1-IL-12 3 mo earlier (n = 10) were inoculated subcutaneously with parental AB1 cells (105). Data are from a single experiment and are representative of two different experiments.

Paracrine Secretion of IL-12 Inhibits Growth of Parental AB1

To further elucidate the biologic effects of paracrine IL-12, we tested whether AB1-IL-12 could inhibit the concomitant growth of parental AB1 cells. Mice were inoculated with a total of 10⁶ tumor cells, consisting of a mixture of IL-12-secreting and parental AB1 tumor cells at three different ratios (1:4, 1:1, 4:1). Coinoculation of the IL-12 transfectant with the parental tumor inhibited the growth of parental tumor in a dose-dependent manner (i.e., the inocula containing a higher percentage of IL-12-producing tumor cells showed a lower incidence of tumor engraftment) (Figure 5). At the highest ratio of AB1-IL-12:AB1, 60% of the animals were protected from AB1 tumor growth. This result was not simply due to a smaller inoculum of parental AB1 cells, since this tumor grew at the same rate in control animals irrespective of inoculation dose (Figure 5).

View larger version (31K): [in this window] [in a new window]   Figure 5.   The presence of AB1-IL-12 protected against parental AB1 tumor development in a dose-dependent manner. BALB/c mice were inoculated subcutaneously with a total of 106 tumor cells containing different ratios of the IL-12-producing AB1 transfectant (as indicated). Mice in the control groups were inoculated with the corresponding number of parental AB1 tumor cells alone. All groups consisted of 10 animals, and the data presented are from a single experiment that is representative of two different experiments.

Effects of MM-IL-12 Production on an Established Distal, Untransfected MM Tumor

The experiments previously described showed that the local secretion of IL-12 enhanced protection against the parental tumor locally. We next determined whether this response was sufficient to affect tumor growth at a distant site. To this end, one group of mice was injected subcutaneously in the left flank with 10⁶ AB1-IL-12 cells and in the right flank with 10⁴ AB1 cells. The control group, inoculated with 10⁴ AB1 cells, started to develop tumors by Day 20, and all mice had tumors by Day 40. In contrast, mice inoculated with AB1-IL-12 demonstrated either delayed onset of the distal parental AB1 or, in 60% of the recipients, complete absence of tumor growth (Figure 6). However, this protective effect was only seen when the parental AB1 tumor burden was very low (10⁴ cells). At greater inocula of AB1 cells (10⁵ or 10⁶ cells), tumors developed at the same rate as in animals that were not coinoculated with AB1-IL-12 (data not shown). Immunohistologic analysis of the AB1 tumors that eventually emerged in mice inoculated distally with AB1-IL-12 revealed more CD4⁺ and CD8⁺ T cells infiltrating the tumor mass (Figures 3E and 3F) than in AB1 tumors from mice that were not concomitantly inoculated (Figures 3C and 3D). However, the degree of CD4⁺ and CD8⁺ T-cell infiltration was dramatically less than that seen in the transient tumors formed by AB1-IL-12 cells, which were subsequently rejected (Figures 3A and 3B).

View larger version (16K): [in this window] [in a new window]   Figure 6.   Animals inoculated with AB1-IL-12 are protected against parental AB1 tumor at a distal site. BALB/c mice were challenged subcutaneously in the left flank with 104 AB1 cells and the test group was also given 106 AB1-IL-12 cells, but in the opposite flank. Groups consisted of 10 animals, and the data presented are representative of two experiments.

Identifying the Effector Mechanism Induced by MM-IL-12

To elucidate the effector cell involved in the immunity induced by AB1-IL-12 transfectant cells, we tested the ability of LN cells from animals inoculated with the transfectant cells to subsequently protect against parental tumor in a Winn assay. Initially, mice were inoculated subcutaneously in the left flank with 10⁶ AB1-IL-12 cells and were monitored for 3 mo to ensure lack of tumor growth. At this time point, the mice were reinoculated with 10⁶ transfectant cells, and 7 d later, lymphocytes were extracted from the draining LNs, mixed with 10⁵ parental AB1 tumor cells at a ratio of 80:1, and injected subcutaneously into naive BALB/c mice. BALB/c mice given LN cells from a naive animal or from control mice that had received only AB1 cells developed tumors by Days 20-22. In contrast, 100% of mice coinoculated with LN cells from the AB1-IL-12 challenged mice failed to develop tumors (Figure 7). Depletion of CD8⁺ lymphocytes prior to mixing with AB1 cells resulted in rapid tumor formation, albeit this was slightly delayed as compared with that in control mice inoculated with tumor cells alone. The depletion of CD4⁺ cells from the inoculum had little effect on protection against tumor growth, since the animals so treated remained tumor free for the entire period of the experiment. These data indicate that the anti-MM effect was largely due to the development of CD8⁺ effector cells.

View larger version (26K): [in this window] [in a new window]   Figure 7.   CD8+ cells generated by multiple primings with AB1- IL-12 protected against parental AB1 tumor in the Winn assay. Mice that were primed with 106 AB1-IL-12 cells were rechallenged with 106 AB1-IL-12 cells 3 mo later. Inguinal LN cells from these mice or from naive mice were mixed with viable AB1 cells at a ratio of 80:1, and the mixture was injected subcutaneously into BALB/c mice, after which growth of AB1 tumors was monitored. Groups consisted of five animals, and the data presented are representative of two experiments.

In addition, LN cells from animals inoculated subcutaneously 7 d earlier with AB1 cells were isolated and induced to proliferate by cross-linking through CD3, and their subsequent IFN- production was tested with an ELISA. In two of three independent experiments, higher levels of IFN- were produced in animals inoculated with AB1-IL-12 than in those given the control tumor line (Figure 8).

View larger version (15K): [in this window] [in a new window]   Figure 8.   Lymphocytes from the draining LNs of mice inoculated with AB1-IL-12 produced IFN-. LN cells from AB1-IL-12- inoculated mice were activated in vitro by plate-bound anti-CD3, and IFN- production was measured with an ELISA.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Substantial recent information suggests that immunotherapy might be a useful therapeutic option in MM, which is encouraging considering the frustrating failure of all standard treatment modalities to offer hope to patients with this disease. One of the most powerful experimental approaches used to date against tumors has been administration of the cytokine IL-12, which induces a number of antitumor effects, such as the induction of tumor-specific immune responses and antiangiogenic factors. In fact, we have previously shown that systemic administration of IL-12 reduces the incidence of murine MM tumors (11). Unfortunately, IL-12 in therapeutic doses, particularly those associated with systemic administration, can cause severe toxic side effects. Because of this we reasoned that local delivery of IL-12 may be of greater therapeutic value, and that gene transfer may be the most feasible and powerful way to achieve this. Indeed, the data presented here show that paracrine secretion of IL-12 induced both local and distal anti-MM responses without any associated toxicity.

That the antitumor effects in our study were the result of induction of systemic T-cell immunity was supported by four observations. First, AB1-IL-12 transfectant cells did not produce progressive tumors in normal mice but did so in nude mice, thus implicating T-cells in tumor rejection. Importantly, tumor rejection in our model correlated with the degree of T-cell infiltration. The transient tumors that developed in a proportion of mice challenged with AB1- IL-12 had large numbers of infiltrating CD4⁺ and CD8⁺ cells, in contrast to mice injected with parental AB1 tumor only, which had very few CD4⁺ or CD8⁺ cells infiltrating the tumor mass. This is consistent with our previous finding that intratumoral injection of IL-12 was associated with large numbers of T-cells infiltrating the MM tumors, whereas progressively growing tumors had sparse T-cell infiltrates (11). This correlation between regression and cellular infiltration with IL-12 treatment has been reported in other systems (30, 32). Second, mice challenged with AB1-IL-12 were protected against rechallenge with parental AB1 tumor, indicating that a memory response had been established. Third, memory CD8⁺ effector cells generated by multiple priming with AB1-IL-12 completely prevented the growth of parental AB1 when coinoculated into naive recipients. We have not observed such memory responses using the irradiated parental AB1 cell line in a vaccine protocol. Moreover, we found evidence of systemic antitumor activity at a distal site.

These immune effects do not negate the possibility of other factors contributing toward tumor rejection. As mentioned earlier, IL-12 is associated with antiangiogenic effects via the induction of IP-10 (33), the expression of which is inducible by IFN- (37). This mechanism is difficult to demonstrate histologically, since tumor growth does not occur in animals inoculated with AB1-IL-12 cells, although as shown in Figure 3, some animals develop a small transitory nodule at the site of tumor inoculation. Histologically, these nodules consist of sparse tumor cells infiltrating the subcutaneous fatty tissue amidst a network of infiltrating lymphocytes and macrophages, without any of the organization that is seen in nonrejecting tumors. It is therefore possible that even at this early stage, blood vessel formation is altered, and in turn alters tumor progression. However, the growth of the transfectant in immune-deficient mice clearly intimates that any angiogenic effects are probably minor and secondary to the induction of specific T-cell immunity.

The antitumor activity of locally produced IL-12 is likely to affect varied aspects of immunity. It may enhance the induction of effector cells in the secondary lymphoid tissue, as well as affect the recruitment and/or function of antitumor effector cells and antigen-presenting cells (APCs) at the tumor site. The effector cells induced by administration of IL-12 generally appear to depend on the experimental model. It has been variously reported that both the CD4⁺ and CD8⁺ subsets are essential (22), that either CD4⁺ or CD8⁺ cells are sufficient (38), that only CD8⁺ cells are necessary (24, 39) or that CD4⁺ cells actually inhibit CD8⁺-cell-mediated protection (40). When specific subsets were depleted prior to systemic treatment with IL-12 in our study, it was found that both CD4⁺ and CD8⁺ cells were essential in mediating immunity to MM, but that NK cells were irrelevant (11). For long-term immunity, the generation of memory effectors from AB1-IL-12-inoculated mice showed that CD8⁺ cells were indeed the only subset required to stop subsequent tumor growth. As with systemic administration of IL-12, AB1-IL-12 cells were still rejected in experiments in which recipients were depleted of NK cells, indicating a minimal, if any, role for these cells in rejection of the transfectant cells (data not shown).

The anatomic location at which local production of IL-12 facilitates the priming of an immune response is not clear, but there are several possibilities. First, transfectant tumor cell products could reach the draining LNs and lead to effective priming. At the LN, tumor antigen may be presented in one of two ways. Uptake of tumor antigen and cross-presentation via the major histocompatibility complex (MHC) class I pathway has been demonstrated in other systems (41), and in our model this process may be rendered more effective in the context of IL-12 secretion. We have evidence that MM antigens are cross-presented to CD8⁺ cells in the LNs draining the tumor site (52). Alternatively, tumor antigen may be directly presented to LN T cells by the transfectant cells, the activation process again being facilitated by increased local IL-12 (42). A third possibility is that the transfectants directly activate tumor-reactive T cells at the transfectants' site of injection.

T cells from mice challenged with AB1-IL-12 produced significantly higher amounts of IFN- upon activation than did those inoculated with the parental tumor, indicating a more Th1-type response. This result correlates with the well-documented property of IL-12 as a potent inducer of IFN- production by NK and T cells (16, 17). IL-12 also augments the proliferation of activated NK and T cells, and enhances NK and CTL activity (12), properties that further contribute to the antitumor activity of this cytokine, although as already discussed, no role for NK cells has been shown in anti-MM immune responses. Apart from inducing Th1 differentiation, it has been shown that IL-12 treatment can reverse suppression of IFN- production by T cells and correlates with tumor rejection (21). It is interesting to note that in our murine model of nontransfected MM, tumor-infiltrating lymphocytes produce messenger RNA for IFN- in the early stages of tumor development, but that this becomes undetectable as tumors progress (43). This loss of IFN- production may be an important factor in the relatively weak immunogenic nature of this tumor, and the IL-12 transfectant may lead to reversal of this trait.

The local production of the IFN- by infiltrating T cells within the tumor milieu could have other positive effects on the tumor environment. IFN- upregulates MHC class I and II antigen expression on APCs (44) and on some other cell types, thus facilitating antigen presentation. Interestingly the transient AB1-IL-12 tumors in our study not only contained a large T-cell infiltrate, but also showed a more widespread pattern of class II MHC antigen expression than did parental tumors. It is unlikely that the AB1-IL-12 transfectant, via IFN- production, induced increased class II antigen expression on the tumor cells, since IFN- does not induce the upregulation of class II antigens on MM cells in vitro (data not shown). In addition, the immunohistology of MM tumors indicates that the number of tumor cells does not correlate with the degree of staining for MHC class II antigens, making it more likely that the class II-bearing cells within the tumor mass are infiltrating macrophages, as previously described (45), or possibly dendritic cells, which could potentiate CD4⁺ cell responses within the tumor. This is important because although we have shown that CD8⁺ cells are the effectors against MM tumors, CD4⁺ cells may provide help in maintaining or prolonging the CD8⁺-cell response, thereby contributing, albeit indirectly, to the efficacy of the antitumor response. IFN- is also one of the components involved in the induction of NO production by cells such as macrophages (46, 47). This compound has cytostatic/cytotoxic activity against tumor cells (48), and has been shown to directly contribute to the antitumor properties of IL-12 (51). Because macrophages are the predominant infiltrating leukocytes in MM tumors, it is possible that this pathway is also active in rejection of the IL-12 transfectant. Thus, once an immune response has been induced, the production of IL-12, either locally or in the LNs, may initiate a cascade of inflammatory signals that facilitates lymphocyte recruitment to the tumor site and enhances effector function.

In our model of MM, the importance of local production of IL-12 in facilitating tumor destruction is highlighted by the fact that in some instances, parental AB1 cells, at a site distal from an AB1-IL-12 inoculum, were not effectively destroyed despite the complete elimination of AB1- IL-12 cells. In our earlier studies with systemic administration of IL-12, we found that continuous IL-12 was required for complete tumor regression (11). Therefore, the failure of distal tumor to regress in this instance may be directly related to the efficacy of AB1-IL-12 rejection, which would limit the maintenance of IL-12 in the system. Furthermore, the dose of IL-12 produced within the tumor environment affects tumor development, as shown by our experiments involving mixing of parental and transfected cells, in which higher local production of IL-12 reduced the incidence of tumors. This finding agrees with other reports that the amount of IL-12 produced at the tumor site is critical in preventing tumor growth (30, 31).

The implications of our findings for the treatment of MM are 3-fold. First, our study shows that locally produced IL-12, engineered by gene-transfer methods, can prevent tumor growth by immunologic means without causing any obvious side effects. Hence, production of IL-12 through gene therapy may be an appropriate means to treat MM. Second, the amount of IL-12 produced at the tumor site will be important in determining the efficacy of the treatment. Third, the immune response against the primary tumor will only be effective in reducing distal tumors if the tumor burden is very low or if the efficacy of this approach is improved. To effectively eradicate all tumor burden, it may be necessary to combine IL-12 therapy with other immunostimulatory regimens.