Efficient induction of an antigen-specific, T helper type 1 immune response by interleukin-12-secreting fibroblasts

Abstract

To determine whether the paracrine secretion of interleukin (IL)-12 can efficiently convert immune responses characterized by high levels of synthesis of IL-4 and immunoglobulin E (IgE) into T helper 1 (Th1)-dominated responses, 3T3 fibroblasts were stably transfected to secrete IL-12 (480 units/10⁶ cells/48 hr). Their effects on the T helper cell-mediated immune response were investigated in ovalbumin (OVA)-primed mice. Free mouse recombinant IL-12 was included as a control group. IL-12-secreting fibroblasts (3T3/IL-12) were more effective than free recombinant IL-12 at increasing OVA-specific interferon-γ (IFN-γ) production and decreasing OVA-specific IL- 4 production in CD4⁺ T cells. In addition, injection with 3T3/IL-12 cells significantly increased anti-OVA immunoglobulin G2a (IgG2a) levels and decreased anti-OVA IgE levels in OVA-primed mice. This work suggests that IL-12-secreting fibroblasts can efficiently induce an antigen-specific Th1 response and may be beneficial in the treatment of diseases caused by undesirable T helper 2 (Th2)-dominated responses, including allergic diseases.

INTRODUCTION

T helper (Th) lymphocytes can be divided into two distinct subsets of effector cells – T helper 1 (Th1) and T helper 2 (Th2) – based on their functional capabilities and the profile of cytokines they produce.¹ The Th1 subset of CD4⁺ T cells secretes cytokines such as interferon-γ (IFN-γ) and tumour necrosis factor-α (TNF-α), and induces cell-mediated immune responses. The Th2 subset produces cytokines such as interleukin (IL)-4 and IL-5, which help B cells to proliferate and differentiate, and is associated with humoral immune responses.² Recent studies indicate that the ratio of Th1 and Th2 is closely correlated with the outcome of many diseases.^{3, 4} Polarized Th1-type and Th2-type responses play different roles in protection, Th1 being effective in the defence against intracellular pathogens and Th2 against intestinal nematodes.⁵ Moreover, Th1 responses predominate in organ-specific autoimmune disorders, acute allograft rejection, unexplained recurrent abortions and in some chronic inflammatory disorders of unknown aetiology.^{6, 7} In contrast, Th2 responses predominate in Omenn’s syndrome, transplantation tolerance, chronic graft versus host disease, systemic sclerosis and allergic diseases.⁸ The nature of Th1 or Th2 polarizing signals is not yet fully understood. However, the cytokines that are present in the environment of the CD4⁺ T cell at the time it encounters the antigen importantly regulate the differentiation of Th cells into either Th1 or Th2 subsets.⁹

IL-12, a heterodimeric cytokine secreted by macrophages (Mφ) and other antigen-presenting cells, is critical for the development of Th1 cells and initiation of the cell-mediated immune response.¹⁰ Recent evidence showed that administration of recombinant (r)IL-12 may be a key strategy in the treatment of Th2-dominated diseases such as infectious diseases and allergic diseases.^{11, 12} However, a single rIL-12 injection is not sufficient for a therapeutic effect and, in some cases, rIL-12 is effective only if administered at the time of parasite inoculation.¹³ Rempel et al. reported that administration in vivo of rIL-12 induced profound, but transient, commitment to Th1-associated patterns of cytokine and antibody production.¹⁴ Repeated injections of rIL-12 at relatively high doses showed severe toxicities, including an increase in transaminase concentration, pulmonary toxicity and leukopenia.^{15, 16} Alternatively, the cytokine concentration necessary for a therapeutic effect could be achieved by administration of genetically engineered cell lines containing a cytokine gene and constitutively producing cytokine. Previous reports showed that fibroblasts could be genetically modified and used for paracrine secretion of cytokines in tumour and infectious disease models.^{17, 18} Immunization with IL-12 gene-transfected fibroblasts delayed tumour formation and promoted antitumour immunity, ¹⁹ and clinical studies using cytokine-secreting fibroblasts are under investigation in cancer patients.

In this study we investigated the effects of long-lasting IL-12, maintained by IL-12-secreting fibroblasts, on an ongoing Th2-dominated immune response, and compared the results with those of free rIL-12. We demonstrated that IL-12-secreting fibroblasts were much more efficient than free rIL-12 in converting a Th2 response into a Th1-dominated response. This result suggests that fibroblasts may serve as a delivery system for paracrine secretion of IL-12 in the treatment of Th2-dominated diseases.

MATERIALS AND METHODS

Reagents, monoclonal antibodies (mAbs) and animals

Ovalbumin (OVA), phytohaemagglutinin (PHA) and alum were purchased from Sigma Chemical Co. (St. Louis, MO). Anti-mouse IL-4 (BVD4 and BVD6) and anti-mouse IFN-γ monoclonal antibodies (mAbs) (R4-6A2 and XMG1.2) were purified from ascitic fluids by ammonium sulphate precipitation followed by diethylaminoethyl–Sephacel chromatography (Sigma). mAb-secreting hybridomas and BALB/3T3 cells were obtained from the American Type Culture Collection (ATCC; Rockville, MD). The cells were maintained at 37° in a humidified 5% CO₂ atmosphere in RPMI-1640 or Dulbecco’s modified Eagle’s minimal essential medium (DMEM) containing 10% fetal bovine serum and antibiotics (growth medium). Anti-IL-12 p40 mAb, C17.8 (rat immunoglobulin G2a [IgG2a]), was kindly donated by Dr G. Trinchieri (Wistar Institute, Philadelphia, PA) and anti-IL-6 mAb (rat IgG2a) was obtained from PharMingen (San Diego, CA).

Six- to eight-week-old female BALB/c mice were obtained from the Charles River Laboratories (Wilmington, MA), and maintained in pathogen-limited conditions.

Construction of IL-12-secreting fibroblasts (3T3/IL-12)

Murine IL-12 p40 or IL-12 p35 genes were subcloned into a retroviral vector, pZipNeoSV(X), (originally from Dr R. Mulligan, Whitehead Institute for Biomedical Research, Cambridge, MA) termed subsequently as pZipNeoSV-p40 and pZipNeoSV-p35, respectively. The pZipNeoSV(X) vector contains the bacterial neomycin phosphotransferase gene under the control of the Moloney leukaemia virus long-terminal repeat. 3T3 fibroblasts were co-transfected with these two expression plasmids using a cationic liposome-mediated transfection method (lipofectin; Gibco BRL, Grand Island, NY). To obtain stably transfected clones, transfected cells were grown in growth medium containing G418 (500 μg/l) for ≈ 14 days, and resistant clones were propagated separately, with subsequent determination of IL-12 bioactivity in the culture supernatants. Control 3T3 cells containing only the neomycin phosphotransferase gene were prepared by transfection with pZipNeoSV(X) using the same procedure.

Injection of mice with the 3T3 cell constructs

Mice were injected intraperitoneally (i.p.) with 50 µg of OVA adsorbed onto 5 mg of Al(OH)₃ adjuvant. One week later, mice were injected i.p. with mixtures of 50 µg of OVA and one of each cell type (see the legends of Table 1, Fig. 1 and Fig. 2). Two weeks after this injection, mice were bled for the determination of anti-OVA Ab isotypes, and splenic CD4⁺ T cells were purified for cytokine assay.

Table 1.

Dose-dependent effect of interleukin-12 (IL-12)-secreting fibroblasts on the T helper 1 (Th1) cell-mediated immune response

Treatment*	IFN-γ production (ng/ml)	Anti-OVA IgG2a (µg/ml)
No vaccination	46·2 ± 2·7	0·8 ± 0·3
3T3/ZipNeo (1 × 107)	48·5 ± 4·5	1·1 ± 0·8
3T3/IL-12 (1 × 107)	178·2 ± 5·6	28·5 ± 2·5
[3T3/ZipNeo + 3T3/IL-12]	160·5 ± 3·8†	33·2 ± 4·1†
(1 × 106) (10%) (9 × 106) (90%)	118·7 ± 6·2†	20·6 ± 1·3†
(5 × 106) (50%) (5 × 106) (50%)	80·3 ± 2·1†	16·7 ± 0·8†
(9 × 106) (90%) (1 × 106) (10%)
(9·8 × 106) (98%) (0·2 × 106) (2%)	56·2 ± 6·7	3·8 ± 1·6

^*Mice were injected intraperitoneally (i.p.) with ovalbumin (OVA) in alum, followed by i.p. injection with 1 × 10⁷ 3T3/ZipNeo or 3T3/IL-12 cells, or with 3T3 cells consisting of different percentages of the 3T3/ZipNeo cells and 3T3/IL-12 cells. Experiments were repeated twice, with similar results obtained on each occasion.

^†P < 0·001, relative to a group injected with the 3T3/ZipNeo cells.

IFN-γ, interferon-γ.

Figure 1.

Production of anti-ovalbumin (OVA) antibody isotypes in OVA-primed BALB/c mice followed by injection with interleukin-12 (IL-12)-secreting fibroblasts. Mice were injected with OVA in alum. One week later, the mice were injected intraperitoneally (i.p.) with 1 × 10⁷ 3T3/IL-12 or 3T3/ZipNeo cells, or mouse recombinant (r)IL-12 (40 or 100 ng), or were not treated. Two weeks after injection, the anti-OVA isotypes in sera were determined by using isotype-specific enzyme-linked immunosorbent assay (ELISA). Data represent the mean ± SD values of five mice. *P < 0·05, relative to groups injected with the 3T3/ZipeNeo cells or rIL-12.

Figure 2.

Cytokine profiles of CD4⁺ T cells in ovalbumin (OVA)-primed BALB/c mice followed by injection with interleukin-12 (IL-12)-secreting fibroblasts. Mice were injected with OVA in alum. One week later, the mice were injected intraperitoneally (i.p.) with 1 × 10⁷ 3T3/IL-12 or 3T3/ZipNeo cells, or mouse recombinant (r)IL-12 (40, 100 ng), or not injected. Two weeks after the injection, splenic CD4⁺ T cells were isolated and stimulated with 100 µg/ml OVA for 3 days and the levels of interferon-γ (IFN-γ) and interleukin-4 (IL-4) in culture supernatants were measured by using enzyme-linked immunosorbent assay (ELISA). Data represent the mean ± SD values of triplicate determinations. The experiment was repeated three times, with similar results obtained on each occasion. *P < 0·001, relative to groups vaccinated with 3T3/ZipeNeo cells or rIL-12.

Preparation of splenic CD4⁺ T cells

Mice were killed and their spleens aseptically removed. Single-cell suspensions of spleen cells were prepared, and CD4⁺ T cells were purified by negative selection using magnetic activated cell sorter (MACS) analysis with a cocktail of biotinylated anti-mouse CD8, I-A, B220 and Mac-1 antibodies (Miltenyi, Sunnyvale, CA), as previously described.²⁰ Greater than 98% of the isolated cells were CD4⁺ T cells, as determined using cytofluorometry. For cytokine assay, 5 × 10⁵ spleen CD4⁺ T cells were stimulated in vitro in RPMI-1640 medium containing 100 µg/ml of OVA. After 3 days, supernatants were removed from each culture for measurement of cytokine levels.

Cytokine assays

The quantities of IL-4 and IFN-γ in culture supernatants were determined by sandwich enzyme-linked immunosorbent assays (ELISAs), using mAbs specific for each cytokine, as previously described.²¹ The mAbs for coating the plates and the biotinylated secondary mAbs were as follows: for IL-4, rat anti-mouse IL-4 (BVD4) and biotinylated anti-mouse IL-4 (BVD6); for IFN-γ, rat anti-mouse IFN-γ (R4-6A2) and biotinylated rat anti-mouse IFN-γ (XMG1.2). The lower limit of detection was 3 pg/ml for IL-4 and 125 pg/ml for IFN-γ. The biological activity of IL-12 produced by transfectants was determined by the ability to stimulate the proliferation of PHA-activated lymphocytes, as previously described.²² Briefly, culture supernatants were harvested after incubation of 10⁶ cells for 48 hr and then passed through a 0·2 µm membrane filter. Dilutions were transferred to 96-well microplates containing 10⁴ PHA-activated lymphocytes in a final volume of 200 µl. After 16 hr, the cells were pulsed with 0·5 µCi of ³H-thymidine and incorporation was measured 8 hr later. One unit of IL-12 gives half-maximal proliferation of lymphocytes under these experimental conditions. Results were extrapolated from a standard curve using a defined dose of mouse rIL-12 (specific activity, 1·2 × 10⁸ units/mg of protein; Hoffmann-La Roche, Nutley, NJ).

Determination of anti-OVA Ab isotypes

Mice were bled and the titres of OVA-specific Ab in the sera were measured by ELISAs using mAbs specific for each anti-OVA Ab isotype. Briefly, ELISA plates were coated with 5 µg/ml of OVA. After coating, serially diluted sera were added to the plates and incubated overnight at 4°. Anti-OVA IgG1 mAb, anti-OVA IgG2a mAb, and anti-OVA IgG from hyperimmunized mice were used as standards for quantification of IgG subclass and total IgG, respectively. After washing, horseradish peroxidase (HRP)-labelled anti-OVA IgG1, IgG2a and IgG were added and incubated for 2 hr at room temperature. After additional washing, o-phenylene diamine (OPD) substrate was added and developed, and the absorbance (A) was measured at 492 nm. Determination of OVA-specific immunoglobulin E (IgE) was performed by ELISA, using rat anti-mouse IgE (0·5 µg/ml) to coat the plates. After the samples were applied, the plates were washed and biotinylated OVA (5 µg/ml) was added. The plates were rewashed and developed with streptavidin-peroxidase in phosphate-buffered saline (PBS).

Statistical analyses

The Student’s t-test and one-way analysis of variance (anova) were used to determine the statistical differences between the various experimental and control groups. P-values < 0·05 were considered significant.

RESULTS

Stable secretion of biologically active IL-12 by IL-12 gene-transfected 3T3 fibroblasts (3T3/IL-12 cells)

IL-12-secreting 3T3 fibroblasts (3T3/IL-12 cells) were constructed as described in the Materials and methods. The culture supernatants obtained from 15 G418-resistant clones were tested for IL-12 secretion. Of these clones, clone 5 was chosen for use in the experiments because of its much higher expression of IL-12 (480 ± 20 units/10⁶ cells/48 hr) compared with the others. This activity, measured in a PHA blast assay, could be inhibited by a neutralizing anti-IL-12p40 mAb (C17.8; rat IgG2a) but not by a control isotype mAb (anti-IL-6; rat IgG2a), suggesting that the activity was IL-12 specific. Furthermore, an approximately equivalent quantity of IL-12 was formed after continuous passage in vitro for longer than 4 months (450 ± 12 units/10⁶ cells/48 hr), suggesting that IL-12 secretion by the 3T3/IL-12 cells was stable. This IL-12-secreting clone (3T3/IL-12) was used in all experiments, and the culture supernatant of the 3T3/IL-12 cells was harvested at the time of in vivo use in every experiment to confirm the expression of IL-12. The 3T3 cells transfected with pZipNeoSV(X) as a control vector (3T3/ZipNeo cells) did not form detectable amounts of IL-12.

Immunization with IL-12-secreting fibroblasts increased anti-OVA IgG2a production and decreased anti-OVA IgE production in OVA (alum)-primed mice

To test the effect of IL-12-secreting fibroblasts on the ongoing Th2-dominated immune response, BALB/c mice were injected with OVA in alum, which is known to induce a Th2-dominated response characterized by enhanced IL-4 production and elevated IgE.^{23, 24} Afterwards, the mice were injected with 1 × 10⁷ 3T3/IL-12 cells or 3T3/ZipNeo cells, or not injected. We also included groups of mice injected with free mouse rIL-12 (40 ng or 100 ng; 120 units/ng), to investigate the importance of persistent IL-12 concentrations maintained by the 3T3/IL-12 cells. Forty nanograms of free rIL-12 is biologically equivalent to the level of IL-12 produced by 1 × 10⁷ 3T3/IL-12 cells for 48 hr. At 2 weeks postinjection, mice were bled and anti-OVA Ab isotypes were determined. As seen in Fig. 1, OVA/IL-12 vaccination significantly increased the anti-OVA IgG2a Ab titre (from 0·8 µg/ml to 27·8 µg/ml), and decreased anti-OVA IgG1 (from 79 µg/ml to 58 µg/ml) and IgE production (from A₄₉₂ = 1·65 to A₄₉₂ = 0·46), compared with a group of unvaccinated mice. The IgG titres were not significantly different between groups. Injection with mouse rIL-12 was less effective at increasing anti-OVA IgG2a and decreasing anti-OVA IgE formation.

IL-12-secreting fibroblasts were more efficient than rIL-12 at increasing IFN-γ production and decreasing IL-4 production by CD4⁺ T cells

Experiments were performed to determine whether or not the functional characteristics shown by splenic CD4⁺ T cells were similar or differentially regulated in mice injected with the 3T3/IL-12 cells or rIL-12, or with the 3T3/ZipNeo cells as a control. Two weeks after the injection, CD4⁺ T cells were purified and cultured in vitro with 100 µg/ml of OVA, and 72-hr culture supernatants were tested for the presence of IFN-γ and IL-4. As seen in Fig. 2, vaccination with OVA/IL-12 also increased IFN-γ production and decreased IL-4 production compared with a group of unvaccinated mice (IFN-γ, from 43·5 ng/ml to 168·4 ng/ml; IL-4, from 1350 pg/ml to 495 pg/ml). Injection with free mouse rIL-12 was less effective at increasing IFN-γ production and decreasing IL-4 production compared to injection with 3T3/IL-12. Therefore, these results suggest that IL-12-secreting cells are more efficient than free rIL-12 at inducing Th1 cell-mediated immune responses.

IL-12-secreting fibroblasts enhanced the Th1 cell-mediated immune response in a dose-dependent manner

To determine if the 3T3/IL-12 cells convert the Th2-dominated response into a Th1-dominated response in a dose-dependent manner, OVA (alum)-injected mice were inoculated with 1 × 10⁷ 3T3 cells (3T3/ZipNeo + 3T3/IL-12) containing different percentages of 3T3/ZipNeo cells and 3T3/IL-12 cells. As shown in Table 1, mice inoculated with ≥ 10% 3T3/IL-12 cells showed a significant increase in the Th1 cell-mediated response (increased production of IFN-γ and anti-OVA IgG1). By contrast, animals inoculated with 2% 3T3-IL-12 cells showed only a marginal reduction.

DISCUSSION

In this work we have demonstrated that the paracrine secretion of IL-12 by the IL-12 gene-transfected fibroblasts induced in vivo Th1 responses more efficiently than free rIL-12. The efficacy was dependent on the injection dose of the IL-12 gene-inserted cells (Table 1). The precise reason for the efficiency with which 3T3/IL-12 cells increased the Th1 response, compared with free rIL-12, is not clear. One possibility is that IL-12-secreting fibroblasts may function as an effective vehicle to produce IL-12 for a prolonged period. In our studies, the 3T3/IL-12 cells were observed in mouse peritoneal cavity even at 2 weeks postinjection, and the 3T3/IL-12 cells recovered from the cavity are still capable of secreting biologically active IL-12. Responsiveness to IL-12 is known to be affected by the presence of large numbers of Ag-specific T cells that remain uncommitted with respect to differentiation and commitment to a Th1 or Th2 phenotype during Ag-driven immune responses.²⁵ Persistence of IL-12 during an Ag-driven immune response may drive, more efficiently, uncommitted Th cells towards the Th1 phenotype on subsequent exposure to Ag.

Fibroblasts were chosen for the paracrine secretion of IL-12 to induce Th1-dominated immune responses because they offer several advantages over other cell types.²⁶ Fibroblasts are readily available, can be selected and expanded rapidly in culture, and are able to produce physiologically relevant levels of cytokine after transfection. Furthermore, immunization with the 3T3/IL-12 cells did not result in any toxicities, such as high fever or weight loss. Lymphoid hyperplasia or tissue necrosis was also not observed in liver, spleen or lungs of mice receiving the 3T3/IL-12 cells (data not shown). In addition, previous reports showed that fibroblasts genetically engineered to secrete granulocyte–macrophage colony-stimulating factor (GM-CSF), IFN-γ, IL-2 or IL-12 could suppress tumour growth and induced antitumour immunity with minimal toxicity.^27–29 Clinical applications of cytokine-secreting autologous human fibroblasts are under investigation in the treatment of cancer patients.³⁰

Optimal treatment of Th2-dominated diseases with IL-12 may necessitate the paracrine secretion of appropriate cytokines, or require additional adjuvants. Administration of rIL-12 into genetically susceptible strains induced protective immunity in mice infected with Mycobacterium avium.³¹ However, no long-lasting benefit of treatment was observed when infected mice were treated with rIL-12 alone. Nabor et al. reported that IL-12 in combination with an antimony-based leishmanicidal drug, ‘Pentostam’, could induce healing of established Leishmania major lesions in BALB/c mice and converted the antiparasite T-cell response from a Th2 to a Th1 response.³² They suggested that IL-12 was effective only when the parasite burden had been reduced and that lowering antigen levels might be critical for immunotherapies directed towards reducing a deleterious Th2 response. In addition, Rempel et al. reported that in vivo administration of rIL-12 induced transient commitment to Th1-associated patterns of cytokine and antibody production.¹⁴

In summary, IL-12-secreting fibroblasts seem to serve as a safe and efficient means for delivering IL-12, and may be beneficial in the treatment of diseases caused by undesirable Th2-dominated responses, including certain parasitic infections and allergic disorders.

Glossary

Abbreviations

ELISA

enzyme-linked immunosorbent assay

IFN-γ

interferon-γ

IL

interleukin

mAb

monoclonal antibody

OVA

ovalbumin

Th1

T helper type 1