Abstract
Multiple sclerosis (MS) relapses are treated with high-dose IV glucocorticosteroids. Here we investigated mechanisms of long-circulating polyethylene glycol-coated liposomes encapsulating prednisolone (PL) in adoptive transfer experimental autoimmune encephalomyelitis. Rats received IV 10 mg/kg PL 6, 18, or 42 hr before sacrifice at disease maximum. In formalin-fixed, paraffin-embedded spinal cord we employed a nonfluorescent immunohistochemical (IHC) double labeling. We stained for tumor necrosis factor-α (TNF-α) in combination with a T-cell antigen. Compared with PBS-containing liposomes, PL at 18 hr, and more at 42 hr, significantly reduced the rate of TNF-α double-labeled T-cells. This correlated with an ameliorated disease score at day 5 after PL 42 hr. Our results help to further understand mechanisms of action of drug targeting by liposomal steroids, with possible implications for treatment of autoimmune disorders such as MS.
Keywords
M
At lower concentrations, GS effects are mainly mediated by the classical GS receptor. Only at ultrahigh tissue concentrations are alternative nongenomic mechanisms of action “activated,” which explains the superior efficacy of high and ultra-high doses in the treatment of some autoimmune disorders (Gold et al. 2001).
In these experiments we investigated in situ effects of PL treatment in EAE using an immunohistochemical (IHC) double labeling method. We show that a single IV injection of 10 mg/kg PL reduces the percentage of TNF-α-positive T-cells in the lesion as part of the downregulation of the inflammatory response. Our results help to understand therapeutic effects of liposomal GS, and ultimately may have implications for treatment of autoimmune disorders such as MS.
Female Lewis rats (Charles River; Sulzfeld, Germany) were 6-8 weeks old. All culture media and supplements were obtained from Gibco BRL (Eggenstein, Germany). Encephalitogenic T-cells for in vivo experiments were generated and maintained as previously described (Schmidt et al. 2000). Briefly, primed T-cells (3 × 105/ml) were restimulated with guinea pig myelin basic protein (MBP, 20 μg/ml) in RPMI 1640 supplemented with 1% normal rat serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM glutamine, using freshly isolated and irradiated (3000 rad) thymocytes (1.5 × 107/ml) as antigen-presenting cells. Adoptive transfer (AT)-EAE was induced by IV injection of 10-12 × 106 freshly activated MBP-specific T-cells in the tail vein. Animals were inspected daily by an observer masked to the respective treatment, using a 6 grade score: 0, healthy; 1, weight loss, limp tip of tail; 2, limp tail, mild paresis; 3, moderate paraparesis, ataxia; 4, tetraparesis; 5, moribund; 6, dead (Schmidt et al. 2000). Disease onset in all animals started at day 2 and was maximal at day 5. Bavarian state authorities approved all experiments.
Liposomes were prepared by the film-extrusion method (Metselaar et al. in press). Briefly, a lipid solution was prepared in ethanol containing dipalmitoyl phosphatidylcholine (DPPC; Lipoid GmbH, Ludwigshafen, Germany), PEG 2000-distearyl phosphatidyl-ethanolamine (PEG–DSPE), and cholesterol (Sigma Chemical; Poole, UK) in a molar ratio of 1.85:0.15:1.0. A lipid film was created by rotary evaporation. The film was hydrated with a solution of 100 mg/ml prednisolone phosphate (Bufa; Uitgeest, The Netherlands) in sterile water. The resulting lipid dispersion was sized by multiple extrusions through polycarbonate filter membranes to a diameter of 90-100 nm. Mean particle size was determined by dynamic light scattering with a Malvern 4700 system (Malvern; Malvern, UK). Phospholipid content was determined with a phosphate assay (Metselaar et al. in press) and prednisolone phosphate concentration by reversed-phase HPLC. Each 1 ml of liposomal preparation contained ∼4.5 mg prednisolone phosphate and an average of 60 μmol phospholipid.
For therapeutic studies we used prednisolone PEG liposomes (PL). The treatment regimen for AT-EAE essentially followed the protocol used in previous studies (Schmidt et al. 2000; Schmidt et al. in press). All experiments were performed in groups of five animals each and were reproduced at least once, some injection time points even three times. Under general anesthesia, 10 mg/kg body weight PL was injected into a tail vein at 6, 18, or 42 hr before sacrifice at day 5. Negative controls received PBS-containing liposomes IV, which showed no difference compared to saline injections in previous experiments. For tissue preparation, anesthetized animals were sacrificed and perfused through the left ventricle with HAES-steril 6% (Fresenius; Bad Homburg, Germany), followed by paraformaldehyde 4% in 0.1 M phosphate buffer. The spinal cord was removed, postfixed, dehydrated, and embedded in paraffin.
Five-μm cross-sections of spinal cord were deparaffinized and rehydrated. After blocking of non-specific binding with 10% BSA in 0.05 M Tris-buffered saline (0.15 M sodium, TBS) for 30 min, sections were incubated with a polyclonal rabbit anti-TNFα antibody (Serotec, via Biozol; München, Germany) at a dilution of 1:100 in TBS with 1% BSA, incubated overnight at 4C. The specificity was proved by preadsorption of the primary antibody with rat TNF-α. Except after BSA blocking, all other steps were followed by washing with TBS. The primary antibody was detected with a biotinylated goat anti-rabbit IgG antibody (Vector; Wertheim, Germany), which was preadsorbed 1:1 with normal rat serum, diluted 1:50 in TBS with 1% BSA and incubated for 45 min. An alkaline phosphatase-bound avidin-biotin complex (Dako; Hamburg, Germany) was applied for 30 min, followed by Vector red (Vector) as chromogenic substrate for 7-10 min. After blocking of all excess avidin-biotin binding sites with an AB blocking kit (Vector), the next primary antibodies were applied. T-cells were detected with a mouse monoclonal antibody to a pan-T-cell antigen (B 115-1, dilution 1:500; from HyCult Biotechnology via Sanbio, Beutelsbach, Germany), incubated for 1 hr at room temperature. Endogenous peroxidase activity was blocked with 3% H2O2 and 0.2 M sodium azide in methanol. The primary antibody was detected with a biotinylated goat anti-mouse IgG antibody preabsorbed 1:1:1 with sera from rabbit and rat at 37C for 15 min, diluted 1:200 in TBS with 1% BSA, and incubated for 45 min. A horseradish-peroxidase-bound avidin-biotin complex (Dako) was applied for 30 min, followed by 3,3′-diaminobenzidine-tetrahydrochloride-nickel (DAB-Ni, black; Vector) as chromogenic substrate for 3-4 min. For each staining we added three control sections by omitting either one or both of the primary antibodies. All sections were dehydrated and mounted in Vitro-clud (R. Langenbrinck; Emmendingen, Germany). In one lumbar (intumescentia lumbalis) spinal cord cross-section, an observer blinded to the respective treatment analyzed 10 fields of a 10 × 10 square grid at a X400 enlargement. The localization of the 10 fields followed a standardized graphic pattern that was applicable to all sections and that yielded equal areas of gray and white matter. Data for TNF-α-positive T-cells were expressed as the ratio of double-labeled T-cells and the total number of T-cells in percent. Statistical analysis of the data was performed by Student's t-test, considering p<0.05 and p<0.01 as significant p-values.
IHC double labeling for detection of TNF-α (visualized by Vector red) in combination with a T-cell marker antigen (visualized by DAB-Ni, black) in a 5-μm formalin-fixed, paraffin-embedded cross-section of spinal cord from a control group AT-EAE rat at day 5. Solid arrows indicate double labeled TNF-α-producing T-cells. Open arrow indicates a TNF-α-negative T-cell. Arrowhead shows a TNF-α-producing cell, which is not a T-cell. Bar = 10 μm.
Treatment of AT-EAE with a single IV injection of 10 mg/ kg PL at indicated time points before sacrifice at day 5, compared to PBS-liposomes (same experiment as in Table 1). Quantification of IHC double labeling for TNF-α and T-cell markers in one lumbar spinal cord cross-section, analyzed as detailed in the text. Percentage of TNF-α-positive T-cells is given as mean ± SD. Each symbol represents data from one rat (n = 5 per group); data were reproduced at least once with similar results. ∗p<0.05 for PL 18 hr vs controls; ∗∗p<0.01 for PL 42 hr vs controls.
Groups of five female Lewis rats received one IV injection of 10 mg/kg PL at 6 hr, 18 hr, or 42 hr before sacrifice at the peak of the disease course of AT-EAE on day 5. TNF-α double-labeled T-cells were detected in spinal cord by a nonfluorescent IHC double labeling technique (Figure 1). At 18 hr after the injection of 10 mg/kg PL, the rate of TNF-α-producing T-cells was clearly reduced compared to controls (Figure 2). Furthermore, PL at 42 hr significantly decreased the rate of TNF-α-producing T-cells. PL at 6 hr had no effect, which was in accord with previous findings (Schmidt et al. in press). In contrast to all other groups, the peak of the disease was significantly ameliorated after PL at 42 hr (Table 1), which correlated with the reduced percentage of TNF-α-positive T cells. Experiments were reproduced at least once with similar results.
The therapeutic goal in treatment of MS relapses is reduction of cellular inflammation as efficiently as possible to prevent ongoing tissue destruction and axon loss. The dosing of GS as mainstay of therapy in MS relapses is a continuous matter of debate. With regard to our previous findings in EAE (Schmidt et al. 2000), one of the major issues of steroids dosing is to reach ultra-high tissue levels, exerting multiple pathways of steroid action according to a new model of steroid mechanisms (Gold et al. 2001). Recently we described a novel formulation of liposomal steroids to deliver ultra-high steroid doses by drug targeting with fewer systemic doses of the free GS in treatment of EAE (Schmidt et al. in press). In these experiments we observed high steroid tissue concentrations soon after injection and accumulation of the liposomes in the inflamed CNS. In contrast to the short-lived actions of free GS, the effects of PL were clearly prolonged, and induction of T-cell apoptosis and reduction of T-cell and macrophage infiltration in the CNS peaked at 42 hr after treatment. At this time point, a high fraction of PL has been degraded after phagocytic uptake and by extracellular proteases, leading to ultra-high tissue levels of the active drug prednisolone. We observed tissue levels greater than 10−5 moles/liter for up to 18 hr after PL, which is well within the range of nongenomic steroid actions and may have been operative in the reduction of the percentage of TNF-α-positive T-cells. Two injections of PL proved superior to two injections of a free GS at a fivefold-higher dose with regard to clinical and in situ effects.
Treatment of AT-EAE with a single IV injection of 10 mg/kg PL at indicated time points before sacrifice at day 5, compared to PBS-liposomes (same experiment as in Figure 2)a
Clinical score is given as mean ± SD (n = 5 per group). Each individual time point was reproduced at least once with similar results in five independent experiments.
p<0.05 for PL 42 hr vs all other groups.
TNF-α, mainly secreted by T-cells and macrophages, is one of the most critical cytokines in the process of demyelination in the course of MS (reviewed in Steinman et al. 2002) and EAE (overview in Kassiotis and Kollias 2001). Its proinflammatory actions have long been established, but during the disease course TNF-α can also exert antiinflammatory properties (Steinman et al. 2002), which may explain the failure of neutralizing TNF-α-antibodies in therapeutic studies of the heterogenous disease MS (The Lenercept Study Group 1999), and the worsening of EAE in a TNF-α-deficient mouse model (Liu et al. 1998). Moreover, TNF-α has been shown to have neuroprotective properties by promoting oligodendrocyte progenitors and remyelination (Arnett et al. 2001). The model of a dualistic role for TNF-α is further supported by a recent study in EAE in which distinct TNF-α signaling pathways were operative either in the beneficial reduction of autoreactive T-cells or in detrimental effects during the acute phase of the disease (Kassiotis and Kollias 2001). Our AT-EAE treatment with PL is only a model to investigate rather short-lived therapeutic effects during the acute phase of CNS inflammation. However, we show how liposomal GS may interfere with the complex cytokine network, finally exerting a beneficial therapeutic effect in our model.
The aim of the present study was to further elucidate the in situ mechanisms of action of a single IV PL injection on the level of specific immune cells. Other techniques such as RT-PCR, Western blotting, and ELISA in homogenized spinal cord lack specificity at single-cell level. Second, these techniques cannot always be performed in formalin-fixed tissue of perfused animals, which is necessary to provide high-quality sections to reliably quantify mechanisms of action on immune cells in situ. The simple but efficient double labeling technique described here enables analysis of cytokine production of specific immune cells and, in contrast to fluorescent double labeling, allows morphological evaluation at the same time. In addition, the stable chromogens facilitate analysis for a long time, which is not possible with fading fluorescent labels. In situ hybridization is a more sensitive technique which, however, is clearly more cost- and labor-intensive and less feasible for quantification of experiments with large numbers of tissue specimens. Second, the latter technique lacks analysis at the protein level, which we assessed here.
Taken together, our experiments using a nonfluorescent immunohistochemical double labeling technique demonstrate that a single injection of PL down-regulates the rate of TNF-α-producing T-cells in spinal cord of EAE rats. Analyses of the production of TNF-α in situ during the demyelination vs the remyelination phase maybe useful for a better understanding of experimental treatment strategies for MS and other neuroinflammatory diseases.
Footnotes
Acknowledgements
Supported by funds from the state of Bavaria, Germany.
The invaluable technical assistance of Gabriele Köllner and Helga Brünner is gratefully acknowledged. We thank Louis van Bloois for his help with preparing the liposomes.