An oral CD40 ligand gene therapy against lymphoma using attenuated Salmonella typhimurium

A20 is a BCL cell line derived from BALB/c mice and expressing CD40 as well as the major histocompatibility complex (MHC) class I and class II h-2d, IgG, and Fc receptor (American Type Culture Collection [ATCC], Rockville, MD).21 The level of B7-2 on the surface of A20 cells was observed to be low with B7-1 being undetectable.22 The wehi3 leukemia cell line is also derived from BALB/c mice but does not express CD40 on the surface, this was used as a negative control.22 A20 cells and wehi3 cells were cultured in 90% RPMI1640 with _L-glutamine medium supplemented with 10% fetal bovine serum (FBS) at 37°C in a humidified 5% CO₂ atmosphere. SC injection of 10⁵ A20 cells was sufficient to form lymphoma, being lethal in all injected BALB/c mice (Japan Crea, Tokyo, Japan) during the period of 26 to 35 days after injection. SC injection of 10⁵ wehi3 cells was lethal in all injected mice within 25 days. This is in concordance with previous reports.22Animal care was in accordance with institutional guidelines.

The auxotrophic ST aroA⁻ strain SL5000 was kindly donated by Dr Bruce A. D. Stocker (Stanford University School of Medicine, Stanford, CA) and used as a gene carrier.20 The full-length human CD40L gene was cloned into pcDL-SR using a CMV promoter (ATCC).23 The auxotrophic ST aroA⁻ strain SL5000 was grown in 100 mL L broth (Sigma Diagnostics, St Louis, MO) to A₆₀₀ of 0.6, chilled on ice, and harvested by centrifugation (15 minutes, 1000g at 4°C). The pellet was suspended in a final volume of 200 μL in 10% glycerol. Aliquots (40 μL) were mixed with 1 to 2 μL DNA in a chilled microcentrifuge tube and transfected to chilled cuvettes (0.2 cm electrode gap). A single pulse of 12.5 kV/cm (2.5 kV, 200 O, 25 μF) was applied and 1 mL of prewarmed SOC was immediately added as previously reported.24 The bacteria were transferred to 17 × 100 mm polypropylenetubes and shaken for 1 hour at 37°C before being plated onto LB agar with ABPC (50 ng/mL). Groups of 6 to 8 female BALB/c mice were fed with 0.5 mL phosphate-buffered saline (PBS) containing 10⁹ colony-forming units (CFU) of ST with or without CD40L gene. None of the mice exhibited any overt signs of illness during immunization.

NIH3T3 cells transfected with human CD40L gene in pcDL-SR expression vector (NIH3T3/CD40LT) and NIH3T3 cells transfected with vector alone (NIH3T3/vt) were prepared by electroporation. Transfectants were selected by growth in 400 μg/mL G418 and then subcloned. The cells were harvested at 70% confluence, washed 3 times in PBS (Sigma) to remove G418, and fixed in 1% formalin (Sigma) for 10 minutes at room temperature. After further washing with PBS, the cells were cultured with A20 cells (100 A20 cells/1 NIH3T3 cell) for 48 hours in 10% FBS plus RPMI1640.

Phenotypic changes of A20 cells (1 × 10⁶/mL) treated with NIH3T3/ D40LT (1 × 10⁴/mL) or NIH3T3/vt (1 × 10⁴/mL) for 48 hours were examined using flow cytometric analysis, as previously described.25Antibody-coated cells were enumerated by flow cytometric analysis using an EPICS V cell sorter (Coulter Electronics, Hialeah, FL). The following antibodies (Abs) were used: FITC-conjugated hamster antimouse CD80 (B7-1) Ab (hamster IgG)(Pharmigen, San Diego, CA), FITC-conjugated rat antimouse CD86 (B7-2) Ab (IgG2a, _k)(Pharmigen), and FITC-conjugated hamster antimouse CD95 (Fas) Ab (hamster IgG)(Pharmigen).

Paraffin-embedded specimens were used for hematoxylin and eosin (HE) staining and Fas ligand (FasL) (Santa Cruz Biotechnology, Santa Cruz, CA) immune-staining as previously described.26 Moreover, anti-CD4 and CD8 Abs (Pharmigen) were used for immunostaining combined with HE staining. For immunohistochemical staining, frozen tissue sections were treated with antihuman CD40L Ab (Santa Cruz).

Soluble human CD40L levels in the sera of BALB/c mice were quantified using the soluble CD40L enzyme-linked immunosorbent assay (ELISA) kit (Chemicon International Inc, Temecula, CA), which used a sandwich-based immunoassay design. The minimal detection level was 0.16 ng/mL of soluble CD40L. Two types of murine monoclonal Abs used in the ELISA system do not detect mouse soluble CD40L protein in the sera.

We first examined the expression of Fas, B7-1, and B7-2 on A20 cells cultured with formalin fixed NIH3T3/vt or NIH3T3/CD40LT cells by immunofluorescence flow cytometry, as shown in Table 1. Levels of Fas and B7-2 on the surface of A20 cells were found to be low, and B7-1 was undetectable when A20 cells were cultured with NIH3T3/vt (control). In contrast, NIH3T3/CD40LT cells upregulated the expression of Fas as well as B7-1 and B7-2 molecules.

To analyze expression of the human CD40L protein in murine tissues, samples from the small intestine, colon, liver, and spleen were analyzed by HE staining as well as immunohistochemistry. We found that in mice immunized with ST40L, the Peyer's patches were prominent (Figure 1A), and the majority of cells in the Peyer's patches could be seen to express the human CD40L protein (positive: brown or yellow color; negative: blue color) (Figure 1B). There were a few CD40L⁺ cells in spleen, but not in liver. In contrast, human CD40L was not detectable in the Peyer's patches of mice treated with ST (Figure 1C). To further confirm the secretion of human soluble CD40L by transfected murine cells into the sera, we next examined it by ELISA (Figure 2). Human soluble CD40L protein was detectable only in BALB/c mice treated with ST40L with or without administration of BCL cells, but not detectable in mice treated with ST and/or BCL cells. The level of soluble CD40L protein in the sera peaked at 1 week after oral administration and was detectable until 7 to 8 weeks.

BALB/c mice were injected SC with 10⁵ A20 cells (Figure3A), these mice were then orally administered with ST40L, ST, or PBS alone, and their survival monitored. Mice in the group treated with ST40L were found to have a significantly longer survival than those treated with ST or PBS (Kaplan-Meier method: Mantel-Cox, P < .0001). The mice in the group treated with ST alone also survived for a significantly longer period than those treated with PBS alone (P < .0001). In contrast, SC injection of an equivalent number of CD40 negative wehi3 leukemia cells showed that there was no significant difference between treatments with ST40L, ST, and PBS. When differing numbers of A20 cells (10⁵, 10⁶, or 10⁷) were injected, the survival rates of the mice were 92%, 77%, and 55%, respectively (Figure 3B). ST40L was then administered before or after tumor challenge (Figure 3C and D). When ST40L was administered 1 week before tumor cell challenge (10⁵ A20 cells, SC injection), no significant differences were observed in mice survival compared with simultaneous vaccination by SC injection of A20 cells alone. However, the efficiency of the ST40L was found to be decreased when the mice were immunized at either 3 weeks (52%, P < .01) or 2 weeks (67%, P < .05) before A20 cell (10⁵) SC injection. This effect was also seen when the mice were ST40L immunized at 3 weeks (42%, P < .01), 2 weeks (69%,P < .001), or 1 week after (70%,P < .02) A20 cell (10⁵) SC injection.

To explore the mechanisms of the protection from BCL growth, histologic analysis was performed on tumor tissue from mice treated with ST40L, ST, or PBS alone. In the mice treated with PBS alone, no cellular infiltrate expressing FasL was observed in the surrounding tissues and inside the BCL region (Figure 4A and D). In contrast, infiltrating lymphocytes expressing FasL were observed around the vessels and also scattered in the smaller tumor tissues in the mice treated with ST (Figure 4B and E). Small hard nodules (2-5 mm in diameter) were observed at the SC injection sites of the long-term survival mice that had been treated with ST40L. On histologic analysis, these small nodules were confirmed to be the result of an accumulation of lymphocytes, and not BCL cells (Figure 4C). These lymphocytes were also found to be strongly positive for FasL expression (Figure 4F). On the other hand, lymphocytes infiltrating in the nodules were stained by either CD4⁺ Ab or CD8⁺ Ab, but the ratio of CD4⁺ and CD8⁺ cells was not 1-sided (data not shown).

In the current study, we have demonstrated an outstanding protection from development of BCL in mice by the oral administration of an ST40L vaccine. SC injection of 10⁵ A20 cells was lethal in all injected BALB/c mice; in contrast, more than 90% of mice survived with simultaneous oral administration of ST40L. The efficiency was dependent on the dose of A20 cells administered and the timing of the immunization. Dilloo et al15 have demonstrated that the CD40L-dependent response generated to a preexisting malignancy is significantly increased by the coinjection of cells secreting IL-2, resulting in antileukemic activity greater than obtained with either molecule alone. French et al16 have shown that anti-CD40 Ab evokes cytotoxic T cells which eradicate BCL in mice, whereas CD40 monoclonal Ab treatment in the absence of BCL cells did not result in an expansion of T cells, suggesting that simultaneous nonspecific stimuli to the immune system may play an important role in the success of CD40L treatment. It is also of interest that the oral administration of ST alone was able to rescue some of the mice injected with BCL. Mutant strains of Salmonella have shown promise as live oral vaccines in humans, capable of stimulating both mucosal and systemic immune responses.27 In particular, the oral administration of an Aro⁻ mutant strain can evoke strong Th1 immune responses dependent on interferon γ.28 These suggest that ST itself may also be a potent inducer of anti-BCL immunity and synergies with CD40L in antitumor effects.

As we have previously reported, ligation of CD40 on multiple myeloma cells up-regulated B7-1 as well as adhesion molecules and MHC, human CD40L enhanced expression of B7-1, and B7-2 on A20 cells in vitro, which may activate tumor-specific cytotoxic T-cell proliferation through CD28.29 Indeed, ectopic expression of the B7 molecule on tumor cells has been shown to be protective in a number of animal tumor models.30-32 Furthermore, CD40L further can induce IL-12 secretion by DCs33 as well as expression of cell surface molecules required for antigen presentation leading to antitumor immunity, which might be superior to the potential of ectopic B7 expression alone. Moreover, a cross-functionality between mouse and human CD40L was demonstrated in this experiment, which is advantageous in that ectopic CD40L expression can be identified by the Abs reacting only to human CD40L. A soluble isoform of CD40L has been shown to exist in the circulation.34 This soluble molecule is a homotrimer of an 18-kd protein exhibiting full activity in B-cell proliferation and differentiation assays, and is able to rescue B cells from apoptosis and bind soluble CD40.35 Soluble CD40L was selectable in the sera of the BALB/c mice treated with ST40L with the level peaking at 1 week after oral administration and being detectable until 7 to 8 weeks after immunization; this further supports the successful transduction of the human CD40 L gene into murine cells.

When A20 cells were stimulated with NIH3T3/CD40LT, Fas expression was found to be up-regulated in vitro. CD40 ligation induces Fas expression on cells and facilitates apoptosis through the FasL/Fas pathway.36 Fas is not always expressed on malignant B cells, but CD40 ligation leads to an up-regulation of Fas antigen and enhances sensitivity to the Fas-mediated death signal in vitro.37,38 ST40L was found to trigger the infiltration of FasL expressing lymphocytes into the BCL tumors, with the degree of infiltration being associated with tumor regression, thus suggesting that ST40L might elicit a strong antitumor immunity via up-regulating Fas on A20 cells and amplifying T cells expressing FasL in tumor.

Previous reports have demonstrated that cytokine-gene therapy, such as that with GM-CSF, may be successful in tumor reduction.39However, most of the gene therapies currently available involve complicated ex vivo cell manipulation,39,40 and only the local administration or ex vivo transfection of the gene is in clinical studies.41 In contrast, the current study using an oral ST40L lymphoma vaccine is characterized by its simplicity of treatment, coupled to a significant efficiency. There are always safety concerns regarding the use of gene therapy and drug delivery systems. However, the safety of the aroA−, live oral ST vaccine has already been confirmed in both human volunteers as well as some other species of animals.42-44 Pawelek et al45,46intraperitoneally injected attenuated S typhimurium with the herpes simplex virus thimidine kinase as an antimelanoma agent, but the use of this could be limited by the potential induction of TNF α–mediated septic shock. Therefore, the oral route may be much safer than systemic administration of ST in clinical use. With respect to the use of the CD40L gene, T-lymphoblastic lymphoma, and other undesirable side effects were observed when the CD40L gene was retrovirally transfected into the bone marrow or thymic cells in CD40L^−/− mice.47 However, in this study, no tumor formation was observed in the mice administered ST40L, the type of target cell appeared to be important and therefore cells in Peyer's patches may be safe targets. From a treatment aspect, we suggested that patients in complete remission may be treated with the oral ST40L gene therapy to prevent the relapse of disease.

In conclusion, attenuated ST carrying CD40L can be used for genetic anti-BCL immunization via the oral route, which may provide a new immuno-gene therapy to B-cell malignancies.

The authors thank Tsuneo Natori, of SRL Inc (Tokyo, Japan), for the excellent specimen of mouse pathology.