Cytokine gene transduced T cells in the treatment of allergic encephalomyelitis and airway hypersensitivity

Gene Ther Mol Biol Vol 6, 101-119, 2001

Cytokine gene transduced T cells in the treatment of allergic encephalomyelitis and airway hypersensitivity

Research Article

Lizhen Chen¹, Rosemarie DeKruyff⁴, Dale Umetsu⁴, Jae-Won Oh⁴, Jeanette Thorbecke^1,3 and Gerald Hochwald^2,*

Depts. of ¹Pathology and ²Neurology, ³Kaplan Comprehensive Cancer Center, NYU School of Medicine, New York, NY 10016

⁴Division of Immunology and Transplantation Biology, Dept. of Pediatrics, Stanford University, Stanford, CA 94305

_________________________________________________________________________________________________

*Correspondence: Gerald Hochwald, Dept. of Neurology, New York University, School of Medicine, New York, NY 10016; FAX 212-2638211; e-mail: hochwg01@med.nyu.edu

Supported by The National Multiple Sclerosis Society Grants #RG-2602A5 and RG3059A1.

Key words: autoimmunity, cytokines, gene therapy, experimental allergic encephalomyelitis, Th1/Th2.

Abbreviations: encephalomyelitis, (EAE); expiratory time, (Te); inflammatory bowel disease, (IBD); Keyhole limpet hemocyanin, (KLH); latency associated protein, (LAP); myelin basic protein, (MBP); ovalbumin, (OVA); peak expiratory flow, (PEF); plasminogen activator inhibitor-1, (PAI-1); proteolipid protein, (PLP); relaxation time, (RT); spleen, (spl); Staphylococcus enterotoxin B, (SEB)

Received: 20 July 2001; accepted: 10 August 2001; electronically published: February 2004

Summary

TGF-b1 or IL-10 transduced myelin basic protein (MBP)-specific BALB/c cloned Th1 cells were injected into SJL x BALB/c F1 mice 11-15 days after immunization with proteolipid protein to induce EAE. TGF-b1/MBP T cells significantly ameliorated the EAE, while IL-10/MBP T cells were less effective. TGF-b1 transduced ovalbumin (OVA)-specific Th1 clones did not influence EAE, even when re-activated by OVA in vivo. However, TGF-b1/OVA T cells did protect against OVA-specific Th2-cell mediated airway hyper-reactivity induced by inhaled OVA. TGF-b1/KLH T cells did not prevent OVA-induced airway hyper-reactivity in mice sensitized and challenged with OVA alone, but did protect mice challenged with KLH + OVA. Thus, the antigen specificity of the Th1 cells allows site-specific delivery of therapeutic TGF-b1 to both Th1 and Th2 cell-mediated inflammatory infiltrates. EAE relapses, induced by bacterial superantigen or endotoxin within 2 weeks, but not >6 weeks, after transfer of TGF-b1 or IL-10/MBP T cells, were reduced. Relapses induced 5 weeks after immunization with PLP could be prevented by simultaneously injected TGF-b1/MBP cells. Spinal cords taken 12-50 days after TGF-b1/MBP cells contained TGF-b1 cDNA. Spinal cords from the majority of mice receiving IL-10/MBP cells contained IL-10 cDNA up to 2 weeks, but not 50 days after cell transfer. Thus, TGF-b1-transduced T cells may be useful in the therapy of autoimmune and allergic inflammatory diseases, but in the EAE model, the same approach with IL-10-transduced T cells appears less effective.

I. Introduction

Resistance to the induction of experimental autoimmune diseases, such as allergic encephalomyelitis (EAE), inflammatory bowel disease (IBD) and collagen induced arthritis (CIA), is often attributed to the presence of immune-regulatory T cells. Such T cells may either be induced prior to induction of the experimental autoimmune disorder by mucosal or systemic exposure to auto-antigens (Karpus and Swanborg, 1991; Khoury et al, 1992), or may be present spontaneously and expanded during the course of the disease. In the latter case, spontaneous recovery from an initial disease episode and subsequent resistance to reinduction of the disease is attributed to the expansion of these immune-regulatory T cells (Ellerman et al, 1988; Kumar and Sercarz, 1993). One of the mechanisms by which such T cells are thought to curb inflammatory lesions characteristic of these autoimmune diseases is by producing and inducing anti-inflammatory cytokines, among which TGF-b, IL-10 and IL-4 have been implicated as very important. Indeed, under certain conditions, neutralization of these cytokines aggravates autoimmune diseases and/or interferes with the activity of immune-regulatory T cells (Kuruvilla et al, 1991; Racke et al, 1992; Johns and Sriram, 1993; Santambrogio et al, 1993; Santos et al, 1994; Stevens et al, 1994; Crisi et al, 1995, 1996; Powrie et al, 1996; Burkhart et al, 1999; Stohlman et al, 1999).

Administration of TGF-b, IL-10 or IL-4, however, only partially protects against autoimmunity. Treatment with active TGF-b is most effective when given during the latter part of the induction phase of EAE (Santambrogio et al, 1993) or CIA (Thorbecke et al, 1992), or at the time of passive induction of EAE with myelin protein-sensitized T cells (Racke et al, 1991; Stevens et al, 1994). This cytokine can also prevent the occurrence of relapses from EAE (Racke et al, 1993; Santambrogio et al, 1993). However, TGF-b cannot cause recovery from EAE or CIA, once the disease has developed. It is of interest that TGF-b1^-/- and TGF-bRII^-/- mice exhibit generalized and fatal T lymphocyte infiltrations in various organs (Diebold et al, 1995; Gorelik and Flavell, 2000). This indicates that TGF-b1 is a cytokine with significant anti-inflammatory and immunosuppressive properties, a key regulator in the maintenance of immunological homeostasis.

Injections of IL-4 or IL-10 are even less effective in modulating autoimmune diseases. These cytokines are reported to have either no effect or to offer protection only when administered early during disease induction (Rott et al, 1994; Santambrogio et al, 1995; Cannella et al, 1996). However, IL-10 knockout (IL-10^-/-) mice are very susceptible to induction of EAE, developing a more severe and persistent form of EAE than do IL-4^-/- or wild-type mice (Bettelli et al, 1998; Samoilova et al, 1998). Moreover, IL-10 transgenic mice are resistant (Bettelli et al, 1998; Cua et al, 1999), while IL-4 transgenic and wild type mice are equally susceptible to induction of the disease. Treatment with IL-10, particularly when administered via the nasal route early during the induction of EAE decreases the severity of the disease (Xiao et al, 1998), but no such effect is observed when administration of IL-10 is delayed until after the initial induction phase or when it is given with sensitized T-cells, at the time of adoptive transfer of EAE (Rott et al, 1994; Nagelkerken et al, 1997). Similarly, neutralization of IL-10 in IL-10 transgenic mice prior to immunization with myelin proteins is needed to completely abolish the resistance of IL-10 transgenic mice to EAE (Cua et al, 1999). Thus, it seems that IL-10 may prevent the sensitization of encephalitogenic T-cells, but that it cannot reverse T-cell sensitization and EAE symptoms. It has also been shown that the local administration of the cDNA encoding viral IL-10 into knee-joints of rabbits can reduce the inflammatory lesions provoked by the intra-articular injection of ovalbumin into ovalbumin pre-sensitized animals (Lechman et al, 1999).

T cells from multiple sclerosis patients reportedly produce less TGF-b1 in culture than do T cells from normal individuals (Mokhtarian et al, 1994). If TGF-b producing T cells are important in the curtailment of autoimmunity, as also suggested by the observations on TGF-b^-/- mice (Diebold et al, 1995), treatment with auto-reactive T cells which have been engineered to produce excess TGF-b might be beneficial. To determine whether auto-reactive T cells which produce IL-10 or TGF-b1 are capable of down-regulating autoimmune disease, we have artificially increased the ability of myelin basic protein (MBP)-specific BALB/c cloned T cells to produce either IL-10 or latent TGF-b1 by transducing them with a recombinant retrovirus engineered to contain the cDNA for one of these cytokines. In previous studies (Chen et al, 1998), we showed that TGF-b1-transduced myelin basic protein (MBP)-specific T cells lose the capacity to provoke EAE in BALB/c mice, and gain instead the ability to protect against EAE, induced in (SJL x BALB/c) F1 mice by immunization with proteolipid protein (PLP). In similar studies on EAE with IL-4 transduced T hybridoma cells (Shaw et al, 1997) and IL-10 transduced T cells (Mathisen et al, 1997) protective effects were also reported. In the latter report, the transduced T cell clone also showed a high level of endogenous IL-10 production, and was not examined for production of other cytokines. It is therefore not certain whether or not the human IL-10, used for the transduction of these cells, was responsible for the protective effect against EAE.

Most of the autoimmune diseases for which a protective role for immunosuppressive cytokines, such as TGF-b and IL-10, has been described are Th1 cell mediated diseases. To determine whether typical Th2 cell induced inflammatory diseases, such as airway hyper-reactivity or asthma, are also down regulated by these cytokines, a similar approach was used in an animal model for asthma. TGF-b1-transduced OVA-specific Th1 cells were found to protect against OVA-specific Th2 cell-induced airway hyper-reactivity (Hansen et al, 2000). Thus, use is made of the migratory properties of antigen-specific activated cloned T cells to obtain enhanced local production of an immune-regulatory cytokine within inflammatory infiltrates, and thereby ameliorate the inflammation.

In the present study, a comparison was made of the relative effectiveness of IL-10 and TGF-b1 transduced MBP-specific T cells in protecting against EAE and relapses of EAE induced by bacterial superantigen or lipopolysaccharide. In addition, the requirement for antigen specificity in exerting protection was further examined for TGF-b1 transduced T cells in both the mouse model of asthma and in EAE.

II. Results

A. Characterization of transduced T cell clones

Transduced T cell clones were identified by the presence of cDNA for latent TGF-b1 or IL-10 as determined by PCR. TGF-b1 or IL-10 transduced and untransduced T cell clones were then compared with respect to their ability to produce the relevant cytokine. Acid treated and untreated serum-free tissue culture medium from antigen activated and resting TGF-b transduced and control T cell clones (10⁶cells/ml, 24 h at 37°C) were analyzed for latent and active TGF-b1 contents, respectively. All three of the latent TGF-b1 transduced clones, MBP-, OVA-, and KLH-specific T cells, exhibited 1.5-4 ng/ml of latent TGF-b in their supernatants, whether or not they had been activated by antigen (Table 1). Supernatants from untransduced T cell clones showed barely detectable amounts of TGF-b1 above the serum-free medium background (< 0.1 ng). Serum containing supernatants from untransduced T cells contained <0.05 ng of IL-10 per ml, while the supernatants from IL-10 transduced T cells contained 1.05-3.0 ng/ml.

To determine whether the production of latent TGF-b1 caused a change in the cytokine pattern produced by the MBP-specific T cell clone, a ribonuclease protection assay was performed on RNA prepared from the untransduced and the TGF-b1-transduced MBP-specific T cell clone 2-3 days after activation of the cells by antigen, using two sets of cytokine probes. The mRNA distribution for LTa, LTb, TNF-a, IFN-g, TGF-b2, TGF-b3 and MIF, expressed as a percentage of mRNA for the housekeeping gene, GAPDH, was the same for the untransduced and transduced clones. However, transduction caused an increase in TGF-b1 mRNA. The results obtained with the other set of cytokine probes showed an absence of mRNA for IL-4, IL-5, IL-6 or IL-10 before and after transduction in the TGF-b1 transduced clone (Chen et al, 1998). The IL-10 transduced clone was examined similarly. Again, no difference in the representation of mRNA for any of the other cytokines was found, even after prolonged propagation of these IL-10 transduced Th1 cells, but a marked increase in the mRNA for IL-10 was seen (data not shown).

The transduced MBP-specific T cells were also characterized with respect to their ability to proliferate in response to antigen (MBP peptide 59-76) in vitro. For both TGF-b1/MBP and IL-10/MBP cells, the dilution of carboxy fluorescein diacetate (succinimidyl ester, CFSE (Lyons and Parish, 1994)) used as label was similar to that in control (untransduced) cells over a period of 3 days in culture after exposure to MBP, and the incorporation of ³H-thymidine at the end of the 3-day culture period was also comparable to that in control cells (data not shown).

B. Comparisons of TGF-b and IL-10 transduced MBP specific T cells in vivo

In previous work we showed that TGF-b1 transduced MBP-specific T cells were able to ameliorate the course of actively induced EAE when transferred approximately at the time of first appearance of disease symptoms, i.e. 11-15 days after immunization with PLP in CFA. In order to compare the effects of IL-10/MBP and TGF-b1/MBP T cells, both transduced cells from the same original Th1 clone were activated in vitro by exposure to MBP and then injected into SJL x BALB/c mice, 11-13 days after the mice had been immunized with PLP in CFA. The results in Figure 1A show that there was an immediate effect of the TGF-b1 transduced cells, such that the severity of EAE that had already developed in these recipients did not increase any further. In contrast, both groups of mice that either received no cells or IL-10 transduced T cells showed a marked increase in EAE severity until day 15. In this experiment there was no protective effect of the IL-10 transduced cells, but in a repeat of this experiment (Figure 1B), the severity of EAE in the mice receiving no cells remained higher between days 16 and 21 than in the mice receiving IL-10 transduced T cells, although this effect was not statistically significant. It should also be noted that untransduced T cells caused a significant increase in severity of EAE symptoms between days 14 and 16, which was not seen in recipients of IL-10 transduced cells, indicating that the augmented production of IL-10 in these cells prevented them from increasing the severity of the EAE.

It was possible that the IL-10/MBP T cells did not reverse EAE because they failed to enter the CNS and/or failed to proliferate locally. We, therefore, analyzed recipients' spinal cords and lymphoid tissue to determine whether cDNA for IL-10 could be detected. The results in Figure 2 show that, indeed, IL-10 cDNA was detectable in the spinal cord of the majority of recipients killed during the first two weeks, but could no longer be detected 50 days after T cell transfer.

In contrast, the majority of mice receiving TGF-b1/MBP cells still had detectable cDNA for that cytokine in their spinal cords at 6 weeks after transfer. Thereafter, however, TGF-b1 cDNA also became undetectable. Neither cDNA was detectable in lymphoid tissue (spleen) either early or late after T cell transfer. Thus, a relatively effective accumulation of the transduced cells in the CNS occurred followed by their gradual disappearance.

C. Requirement for antigen specificity of TGF-b1 transduced T cells

We previously showed that, in order for TGF-b1 transduced T cells to have a protective effect against EAE development, they had to be specific for a myelin antigen (Chen et al, 1998). TGF-b1 transduced KLH or OVA specific Th1 cells did not have such an effect. Similarly, in the experiments on airway hyper-reactivity induced by OVA, TGF-b1/OVA cells protected but TGF-b1/KLH cells did not. In the present study, we analyzed this requirement for antigen specificity in more detail. In the experiments on EAE, in addition to exposing the T cells in vitro to the relevant antigen (OVA) a few days prior to transfer, we also injected the recipients on day 14 with 100 mg OVA ip to obtain additional activation of these cells in the mice. However, as can be seen from the results in Figure 3, there was no effect from these OVA specific T cells, whether the mice were injected with OVA or not (not shown).

Since the TGF-b1/MBP T cells enter the CNS, it is possible that the continued local stimulation in the spinal cord within local inflammatory lesions allows for activation of the latent TGF-b1 that they produce constitutively. However, in the EAE model, it is impossible to provide T cells of other specificities such as OVA or KLH with the antigen to which they respond locally within the CNS. The aspect of bystander effect was therefore further analyzed in the model of airway hyper-reactivity, where local exposure to any antigen can readily be performed by adding that antigen to the challenge inhalation. Indeed, when KLH was added to the OVA used for the challenge inhalation, TGF-b1/KLH Th1 cells could protect against airway hyper-reactivity in mice immunized to OVA (Figure 4A), although they were still less effective than TGF-b1/OVA Th1 cells (Figure 4B).

Figure 4. A protective effect of TGF-b1/KLH Th1 cells against OVA induced airway hyper-reactivity can be obtained if KLH is added to the challenge inhalation of antigen. BALB/c mice were immunized with OVA i.p. (50 mg) complexed with alum on day 1, and challenged intranasally on days 7, 8 and 9 with either 50 mg OVA alone or with 25 mg KLH + 50 mg OVA. A): l¾l TGFb1/KLH Th1 cells + OVA alone; t---t No cells + OVA; Ñ---Ñ No cells + OVA and KLH; O---O TGFb1/KLH Th1 cells + OVA and KLH, * p vs TGFb1/KLH cells + OVA alone <0.05 (n=3); B): Ñ¾Ñ No cells + KLH; l¾l No cells + OVA; t¾t TGF-b1/OVA Th1 cells + KLH; O¾O TGF-b1/OVA Th1 cells + OVA, ** p vs TGFb1/OVA cells + KLH <0.01 (n = 3).

Figure 5: Anti-TGF-b mAb (2G7, 0.5 mg/mouse, ip), injected on the same day as the T cells (day 13) reduces the inhibitory effects of TGF-b MBP cells on EAE development. l¾l No cells control (n=6); t¾t TGF-b 1/MBP Th1 cells with anti-TGF-b mAb (n=6); O¾O TGF-b 1/MBP Th1 cells alone (n=5).

Table 2

Effect of TGF-b1/MBP T Cells on EAE Relapse Incidence induced by SEB or LPS
Expt. #	TGF-b/MBP-Specific T Cells Injected	Relapse Induced With SEB or LPS	EAE Relapse Incidence*
1	Day 13 (after PLP in CFA)	SEB on Day 26	0/11
	None	SEB on Day 26	3/8

2	Day 12 (after PLP in CFA)	SEB on Day 58	5/10
	None	SEB on Day 58	11/14
2	Day 12 (after PLP in CFA)	LPS on Day 70	7/9
	None	LPS on Day 70	9/13

* Mice were considered to have relapsed when their disease incidence had increased by 0.5 or more for at least two consecutive readings within 3 days after injection of the SEB or LPS.

F. LPS induced EAE relapses

In previous studies, we have shown that TNF-a induced relapses were prevented by injection of IL-10, while SEB induced relapses were more effectively prevented by TGF-b injections (Crisi et al, 1995). Since gram negative bacterial endotoxin, LPS, induces the rapid release of TNF-a, we studied the effect of IL-10/MBP and TGF-b1/MBP T cells on the incidence and severity of EAE relapses induced by LPS. As can be seen from the results in Figure 6B and in Table 2, LPS (1 mg), injected on day 21 after immunization with PLP in CFA (or 10 days after transduced T cell transfer), caused an exacerbation of EAE severity in control (no T cells) mice. In this experiment the overall EAE severity was somewhat greater and the rate of recovery somewhat slower than in most other experiments. An additional group that received untransduced MBP specific T cells (not shown) had a slightly higher severity of EAE than the no T cell control and all of these mice died after injection of LPS. At the time of LPS injection, recipients of IL-10/MBP T cells were beginning to show a somewhat lowerEAE severity than the controls and the effect of LPS was minimal (Figure 6B and Table 2). The TGF-b1/MBP T cell recipients had significantly less disease than the other groups and failed to show a significant effect after LPS injection (Figure 6B and Table 2). An additional group of mice received both IL-10/MBP and TGF-b1/MBP T cells, but the protective effects of these combined transduced T cells were not additive (not shown). In the experiment shown in Figure 6A, LPS was injected 8 weeks after T cell transfer. As seen in Table 2 and in Figure 6A, under these conditions LPS induced similar relapses in control and TGF-b1/MBP T cells treated mice. Thus, in these mice, in which cDNA for TGF-b1 could no longer be detected in spinal cords (Figure 2), there was no protection against EAE relapses.

In an additional experiment, shown in Figure 7, TGF-b1/MBP or IL-10/MBP T cells were injected at the same time as LPS into mice that had partially recovered from PLP induced EAE. Cells (2x10⁶, iv) and LPS (1 mg, ip) were injected on day 34 after immunization with PLP in CFA. A second injection of LPS (5 mg) was given 1 week later. In the control group, each injection of LPS induced a slight increase in the EAE score, which lasted only a few days. In the mice receiving TGF-b1 transduced cells, no relapse of the EAE could be detected. In the mice receiving IL-10 transduced T cells, the EAE relapses were somewhat less marked than in the control mice. The recovery after day 45 was accelerated in both the transduced T cell-treated as compared to the control group of mice (Figure 7). On day 9-16 after cell transfer, the cDNA of the transduced cytokine was detectable in the spinal cord of a large percentage of the mice receiving TGF-b1/MBP cells and again a somewhat lower percentage of the mice receiving IL-10/MBP cells (Figure 2B). These results show that TGF-b1/MBP T cells can enter the CNS and protect against exacerbations of EAE, even when given late during the course of the disease.

III. Discussion

The present results confirm our previous findings (Chen et al, 1998) that latent TGF-b1 transduced MBP-specific Th1 cells protect against PLP-induced EAE in (SJL x BALB/c) F1 mice, even when injected shortly after the onset of disease. When left untransduced, the same Th1 cells slightly increase the severity of actively induced EAE (Chen et al, 1998), and induce adoptive EAE in BALB/c mice (Abromson-Leeman et al, 1995). It should be noted that PLP in CFA was used for the induction of EAE, so as to avoid having MBP depots present in any other sites of the body, possibly detaining MBP-specific T cells from reaching the CNS (Chen et al, 1998).

Clearly, the only difference between the transduced and the untransduced cloned T cells is the enhanced production of TGF-b1. The transduced cells remain Th1, because they produce mRNA for TNF, LTa, LTb and IFN-g, and not for IL-4 or IL-10 (Chen et al, 1998). Even though this Th1 cytokine profile is unaltered after transduction with TGF-b1, the cells lose their capacity to aggravate EAE in the recipients, and instead significantly ameliorate the development of EAE. Therefore, the functional properties of these cells in vivo have been changed by the engineered production of latent TGF-b1. In both EAE and experimental asthma, the protective effect of TGF-b1 transduced T cells is abrogated by the simultaneous injection of neutralizing anti-TGF-b, which only interacts with active TGF-b (Chen et al, 1998; Hansen et al, 2000). It should be noted that, under normal conditions, most cells including T cells only produce latent TGF-b, i.e., TGF-b from which the latency associated protein (LAP) must be removed to uncover the receptor binding region before it exerts any biological activity (Wakefield et al, 1988). In inflammatory infiltrates this most likely occurs by enzymes such as plasmin and/or acidification in macrophages (Nunes et al, 1995; Godar et al, 1999). In addition, several other proteins have been shown to be capable of removing the LAP from TGF-b, such as thrombospondulin (Ribeiro et al, 1999) and the integrin avb6 (Munger et al, 1999).

TGF-b may affect autoimmune disease through down regulation of: 1) TNF-a and LT production (Espevik et al, 1987; Stevens et al, 1994); 2) responses to IL-12 (Pardoux et al, 1997); 3) macrophage and microglia activation (Nelson et al, 1991; Vodovotz et al, 1993; Lodge and Sriram, 1996); 4) cytokine enhanced class II MHC expression (Epstein et al, 1991); and 5) migration of T cells into the CNS (Santambrogio et al, 1993; Fabry et al, 1995). TGF-b induces the synthesis of IL-10 by macrophages (Maeda et al, 1995; Kitani et al, 2000), but the present results suggest that this is unlikely to be the mechanism by which TGF-b1/MBP T cells protect against EAE, since IL-10/MBP T cells are less effective. TGF-b also stimulates its own production (Fiorelli et al, 1994) and, therefore, a few TGF-b1/MBP T cells retained in an infiltrate on the basis of their specificity for myelin protein, may cause oligodendrocytes and macrophages in their vicinity to produce more TGF-b. An additional mechanism by which TGF-b may influence autoimmunity is through the promotion of immunoregulatory CD8⁺ T cell development (Quere and Thorbecke, 1990; Rich et al, 1995; Powrie et al, 1996; Thorbecke et al, 1999).

The primary mechanism by which IL-10 protects against the development of autoimmune diseases, such as CIA, is thought to be through inhibition of the production of pro-inflammatory cytokines such as TNF-a, IL-1 and IL-6 (Walmsley et al, 1996; Kim et al, 2000) and of chemokines, such as MIP-1a and MIP-2 (Kasama et al, 1995). Moreover, IL-10 directs T cells away from harmful Th1 responses and associated IgG2a antibody formation, into the direction of Th2 (Kim et al, 2000; Stevens et al, 1988). Indeed, the resistance of IL-10 transgenic mice to induction of EAE is attributed to the inhibition of Th1 responses in such mice (Cua et al, 1999). Similar to TGF-b, IL-10 counteracts the activation of macrophages and in this respect synergizes with TGF-b (Oswald et al, 1992). In contrast to TGF-b, however, IL-10 fails to inhibit NO production by macrophages induced by an extraneous source of TNF-a (Bogdan et al, 1991; Corradin et al, 1993). Both cytokines counteract the upregulation of class II MHC and of FASL expression by IFN-g (de Waal Malefyt et al, 1991; Epstein et al, 1991; Arnold et al, 1999), and inhibit the expression of contact sensitization in sensitized mice (Epstein et al, 1991; Ferguson et al, 1994). It is, therefore, not immediately clear why TGF-b1/MBP T cells are much more effective in our model of EAE than IL-10/MBP T cells, and why the effects of these cells given simultaneously are not additive. Neither of the transduced cloned T cells has been affected in its ability to proliferate in response to MBP, and both are detectable in spinal cords after transfer, although the persistence of the IL-10/MBP cells is somewhat shorter. In the airway hyper-reactivity model, transfer of OVA-specific IL-10-transduced T cells results in pronounced inhibition of airway hyper-reactivity (Oh et al, unpublished observations). The differences in the effectiveness of OVA-transduced cells in these systems may reflect differences in the effects of IL-10 on the Th2 effector cells mediating the airway hyper-reactivity vs the Th1 cells mediating EAE, or in the effects of IL-10 on APCs in the two sites. Another possibility is that the IL-10 produced by the transduced T cells inhibits antigen presentation (van der Veen and Stohlman, 1993; Frei et al, 1994; de Vries, 1995) to themselves in vivo, resulting in a reduced proliferation of these T cells which affects their performance in the more chronic situation of the EAE model, but is less important in acute airway hyper-reactivity.

It has been reported that, unlike TGF-b, IL-10 also has immune-stimulating effects on CD8 T cells (Chen and Zlotnik, 1991; Balasa et al, 1998; Groux et al, 1999) and B cells (Briere et al, 1993). Moreover, while IL-10 inhibits pro-inflammatory cytokine production in macrophages, it does not affect endothelial cells (Sironi et al, 1993) or dendritic cells from rheumatoid synovial fluid (MacDonald et al, 1999). It is possible that the greater inhibition of NO production exerted by TGF-b1 is of importance, as NO has been linked to damage of the CNS in EAE in various studies (Lin et al, 1993; Okuda et al, 1995; Waldburger et al, 1996). It should also noted that in transgenic mice, IL-10 expressed under control of an MHC class II promoter causes enhanced susceptibility to Leishmania infection, while IL-10 expressed only in T cells does not have this effect (Groux et al, 1999). In this respect it is perhaps relevant that in the present studies, the enhanced cytokine production is only in a small population of transferred T cells. While the transduced cytokines, latent TGF-B1 and IL-10, were produced by the T cells to approximately the same levels (in ng amounts), it is not sure what the effective concentrations required in vivo might be for each of these cytokines, or how much of the latent TGF-b1 produced by the cells becomes activated at the sites where it exerts its effect. We have not been able to obtain a higher production of IL-10 in the T cells.

In view of the consideration that clinical application of transduced T-cell therapy in humans would have to be performed after initiation of disease, the possibility of affecting relapses of EAE was also investigated in these studies. The relapses studied here were induced by injection of TNF-a and IFN-g inducing agents, which may mimic clinical situations in which relapses of demyelinating disease are known to occur, such as during infections (Edwards et al, 1998; Metz et al, 1998). Both SEB and LPS induce a burst of TNF-a production and, although unlike IL-10, TGF-b cannot overcome the effects of injected TNF-a, TGF-b does inhibit TNF-a production (Espevik et al, 1987), which may be an important aspect of the inhibitory effect on these relapses. SEB, in addition, stimulates Vb3 and Vb8 T cells (Marrack and Kappler, 1990), and causes production of large amounts of T cell cytokines.

In previous studies on EAE with injected cytokines, we found that IL-10 protected against TNF-a induced relapses, while TGF-b was more effective against SEB induced relapses (Crisi et al, 1995). In the present study, TGF-b1/MBP T cells prevent both the SEB- and LPS-induced increments in EAE scores, and both IL-10/MBP and TGF-b1/MBP T cells ameliorate EAE relapses induced by injection of LPS during the interval when the transduced T cells are still detectable in the CNS. More importantly, injection of the MBP-specific T cells at the time of the induction of the EAE relapse also results in significant protection, particularly by the TGF-b1 transduced cells.

It is of interest that, even though TGF-b1 producing cells are known to be relatively abundant in mucosal linings in the lung (Magnan et al, 1997; Vignola et al, 1997), injection of TGF-b1/OVA T cells nevertheless significantly protects against the local inflammatory responses accompanying airway hyper-reactivity (Hansen et al, 2000). Apparently, a protective effect can only be obtained with these transduced T cells if they localize at the site of the inflammation. In the EAE model, this can only be obtained with T cells specific for a myelin component and activated in vitro prior to cell transfer. It has been shown that activated T cells which penetrate the blood-brain-barrier during EAE have upregulated adhesion molecules on their surfaces, such as VLA-4 and LFA-1, and that the presence of adhesion molecules on cloned T cells influences their capacity to transfer EAE to recipient mice (Kuchroo et al, 1993; Barten et al, 1995). In addition, contact of microvascular endothelial cells with activated T cells causes the enhancement o

Cytokine Production by Transduced Th1 Cells
Cells Tested --------------------------------------	TGF-b1 (ng/ml)* ---------------------------------	IL-10 (ng/ml)** --------------------------------
Untransduced MBP/Th1	0.037	< 0.03
TGF-b1 Transduced MBP/Th1	2.3	NT
IL-10 Transduced MBP/Th1	NT	3.0
TGF-b1 Transduced KLH/Th1	3.5	NT
TGF-b1 Transduced OVA/Th1	3.6	NT

Table 2

Effect of TGF-b1/MBP T Cells on EAE Relapse Incidence induced by SEB or LPS

III. Discussion