Pentoxifylline Prevents Autoimmune Mediated Inflammation in Low Dose Streptozotocin Induced Diabetes

Xanthine derivative, pentoxifylline (PTX), has been recently shown to exert a protective effects in certain animal models of autoimmunity, including diabetes in NOD mice. In the present study, the immunomodulatory potential of PTX was investigated in autoimmune diabetes induced by multiple low doses of streptozotocin (MLD-SZ) in genetically susceptible CBA/H mice (tested with 40 mg SZ/kg b.w. for 5 days) and DA rats (tested with 20 mg/kg b.w. for 5 days). In both species, 2 – 3 weeks following the MLD-SZ treatment, sustained hyperglycemia developed, as an outcome of inflammatory reaction with endothelial cell activation and accumulation of mononuclear cells. Although there was no evidence of typical insulitis in early disease development (day 10), in both rats and mice, macrophages, CD4+ and CD8+ cells were present in the islets of Langerhans as diffuse mononuclear infiltrates with the expression of IFN-γ and inducible NO synthase (iNOS). Administration of PTX (200 mg/kg/day for 10 days) in combination with MLD-SZ reduced insulitis and the production of mediators tested, and prevented the development of hyperglycemia. These results suggest that beneficial effects of PTX involve down-regulation of local proinflammatory cytokine-mediated NO synthase pathway. They also demonstrate that in addition to ameliorating spontaneous autoimmunity in NOD mice, PTX may be effective in downregulating an inflammatory autoimmune process triggered in susceptible host by an external agents, such as streptozotocin.

INTRODUCTION Pentoxifylline (PTX) [3,7-dimethyl-l-(5-oxohexyl) xanthine] is the drug that is widely used for the treatment of vascular disorders. However, there are new evidences for the potential therapeutic properties of this compound. Preclinical studies in laboratory rodents have shown that PTX and related compounds may favorably influence the course of experimental inflammatory and autoimmune disorders, such as autoimmune encephalomyelitis (Rott et al., 1993;Nataf et al., 1993), meningitis (Saez-Lorens et al., 1990) and autoimmune neuritis (Constantinescu et al., 1996). Insulin-dependent diabetes mellitus is characterized by a failure of self-tolerance leading to Thl cell mediated autoimmune attack leading to destruction of insulin-producing cells located in the islets of Langerhans. While high doses of -cell toxin streptozotocin (SZ) induce diabetes through a direct toxic effect on 13-cells, multiple low doses in susceptible strains of mice and rats initiate an autoimmune destructive process similar to that observed in human disease (Like and Rossini, 1976;Luki et al., 1991a;Kolb and Kroncke, 1993). Therefore, MLD-SZ induced diabetes allows for the study on agents potentially designed to modulate -cell specific autodestructive process. However, it is not known whether there are any differences or similarities between spontaneously and experimentally induced diabetic animal models in the mode of action of different immunomodulatory drugs. Recently Liang et al (1998) have demonstrated that PTX inhibits secretion of IL-12 by macrophages and IFN-7 by Thl cells. It appeared that this effect may be responsible for the amelioration of insulitis in NOD mice (Liang et al., 1998). Our data therefore strengthen the notion that similar effector mechanisms are operative in spontaneous and toxin induced autoimmune diabetes. By using the model of diabetes induced by MLD-SZ we demonstrated that PTX, similarly to spontaneous models of diabetes (Liang et al., 1998), suppresses development of hyperglycemia mainly by downregulating the development of inflammatory lesions in the pancreata.

Effects of PTX on Hyperglycemia
In order to evaluate the ability of PTX to interfere with diabetogenic process induced by an external agents, streptozotocin, we used both murine and rat model of autoimmune diabetes induction. Although the dose regimens of 13-cell toxin SZ for mice and rats were not identical, in both rodent species, as already reported (Lukid et al., 1991a;Luki et al., 1991b), repeated injections with 5 subtoxic daily doses of SZ induced delayed hyperglycemia 10 to 20 days after completion of the treatment (Figure 1). In vivo treatment of animals with 10 consecutive constant doses (200 mg/kg/day) of PTX significantly reduces hyperglycemia in both experimental species. The protective effect of the drug is long-lasting, at least for 8 weeks after the end of the treatment when experiment was terminated ( Figure 1).

Cellular Changes
The interference of PTX with a pathological process leading to islet dysfunction have further been studied by immunological and immunohistochemical analysis at the level of target tissue. Histological analysis and comparison with untreated control was done by day 10 after the induction of the disease and completion of the PTX treatment, and by day 56 of the monitoring of hyperglycemia. In comparison to normal architecture of the islets found in animals without any treatment, on day 10 in control MLD-SZ-induced animals histopathological changes are evident. Analysis in some of the islets of susceptible CBA/H mice at this time revealed heavy mononuclear infiltrates ( Figure 2). In susceptible DA rats small infiltrates were rarely seen (grade 1) and infiltrating cells were scattered throughout the islets (Figure 3). In both species mononuclear infiltration was accompanied with initial necrotic changes and slight edema in the connective spaces around the islets and blood vessels (Figures 2 and 3). However, the majority of the islets were still well preserved and free from infiltration, thus suggesting asynchronous process. Immunohistochemical analysis in rats confirmed our earlier findings (Luki et al., 1991 a) that mononuclear cells participating in the insulitis process were CD4 + and CD8 + lymphocytes, as well as blood borne ED1 + macrophages. Some mononuclear cells within the islets also were positively stained for MHC ClassII. At the later stage of the disease (day 56) most of the islets lost clear margins and their characteristic structure. In both species progression of insulitis led to more severe necrotic changes throughout the affected islets (grade 2 prevailed), such as vacuolization related to focal necrosis and picnosis of the islet cells (data not shown). Treatment with PTX concomi-  (Figures 2 and 3). At this time, inflammatory cells with the characteristic phenotype were virtually absent, or rare and spread only to the periphery of the islets ( Figure 3). Vascular dilatation with hypertrophy of the endothelium was hardly observed. In the later period, in comparison to PTX nontreated diabetic animals, hypocellularity and atrophy of the islets were less prominent and most of the islets still remained intact. In general, although PTX prophylaxis could not completely suppress insulitis development and 13-cell damage, it drastically reduced it severity, which was sufficient for the normoglycemic status of the animals ( Figure 1).

Molecular Changes
To investigate the effects of PTX on the molecular alterations accompanying islet destruction, we ana-lyzed immunohystochemically the presence of proinflammator mediators (IFN-/ and iNOS) in the islets. In control MLD-SZ-induced diabetic animals mononuclear cell infiltration in pancreata was accompanied by prominent expression of IFN-7 and iNOS ( Figure 4). In addition to intraislet cells, marked iNOS expression was observed on the endothelial cells of the hypertrophic blood vessels. By contrast, in the PTX-treated group, in some of the islets, rare IFN-7 + and iNOS + cells were evenly dispersed throughout, but most of the remaining islets were completely negative (Figure 4), similarly to healthy nontreated animals. Concordantly with diabetic status, in MLD-SZ-treated animals impaired insulin secretion was found, as revealed by insulin staining of pancreata, while treatment with PTX resulted in unal-

DISCUSSION
In an attempt to elucidate the value of new drug regimens in human disorders, a number of animal models has been used with common molecular mechanisms that are influenced by the compund tested. For IDDM, widely accepted models similar to human disease are spontaneously developing diabetes in NOD mice and BB rats. By using these models it has been recently shown that PTX, and related compounds, rolipram (Liang et al., 1998)   islets, it drastically reduced its severity. It has been recently shown that PTX inhibits ICAM-1 expression in monocytes (Neuner et al., 1997) and the adhesion of T lymphocytes to ICAM-1 and VCAM-1 (Gonzales-Amaro et al., 1998). On the other hand, it has been shown in MLD-SZ-induced diabetic mice that adhesion of lymphocytes to islet endothelium depends on the expression of VCAM-1 and ICAM-1 in the pancreas (Ludwig et al., 1999). Therefore, it seems that beneficial effect of PTX on MLD-SZ-induced pathway of immune attack of -cells may reside, at least in part, in its capacity to interfere with the recruitment of immune and inflammatory cells from circulation. Indeed, we have shown that PTX protects both mice and rats from development of destructive intrainsulitis.
MLD-SZ-induced autoimmune diabetes appears to be a T cell-dependent disease (Elliot et al., 1997). Since T cell reactivity is regulated by APCs, in animal models of IDDM it has been postulated that the functional state of APCs is responsible for the progression of Thl-dependent destructive insulitis (Rothe and Kolb, 1998). It appears that by stimulating Th cells (Cockfield et al., 1989), MLD-SZ caused up-regulation of IFN-y and MHC Class II expression that are required to propagate the autoimmune process leading to IDDM. IFN-7 production can be reduced by treatment with PTX ( Figure 4 and Liang et al., 1998). In addition, the spatial distribution of MHC Class II / cells in control MLD-SZ-induced animals expressed scattered pattern, while in PTX-treated animals Class II / cells were found only in peri-insulitis ( Figure 3). Thus, according to our results, it seems that PTX may influence the disease by down-regulating APCs.
In MLD-SZ induced autoimmune diabetes several mechanisms have been implicated in the pathological processes leading to -cell dysfunction and death, including proinflammatory cytokine mediated induction of iNOS expression and NO production (Lukid et al., 1991b;Lukid et al., 1998). Both exogenous NO derived from nonendocrine islet cells (macrophages or endothelial cells) and NO generated by the 13-cells itself may contribute ,to tissue destruction (Corbett et al., 1992;Kaneto et al., 1995;Corbett and McDaniel, 1995). Since PTX has been found to be able to inhibit the release of inflammatory cytokines ( Van-Furth et al., 1994;Rieneck et al., 1995), and favor Th2 response (Liblau et al., 1995), the drug may indirectly interfere with the induction of iNOS. It fits well with the finding that PTX reduces deleterious effects of TNF-ot on human islets (Ariasdiaz et al., 1995).
Recently we showed that PTX may have opposite effects on iNOS expression in different cell types: inhibitory in macrophages and enhancing in astrocytes (Trajkovi et al., 1997). Although the cellular sources of iNOS in our experimental model may be both endocrine and nonendocrine cells, in the present study we showed that local expression of iNOS by both intraislet cells and endothelial cells were reduced after in vivo treatment with PTX. From this, it can be concluded that another mechanism by which PTX could exert its effect is by reducing NO-mediated destruction of pancreatic -cells Whatever the cellular source of NO is, interference with local NO produc-tion by PTX could be relevant for autoimmune process affecting the pancreas. However, the precise mechanism of action on NO synthesis at the level of NO-producing cells in the pancreas is at present not known, but warrants further study.

Animals
Genetically susceptible inbred male Dark Agouti (DA) rats, and CBA/H mice were obtained from our own breeding colony (Institute for Biological Research, Belgrade, Yugoslavia), determined to be free from common pathogens. Animals were used when 10 to 16 weeks old, and kept in groups of 5 to 6 per cage.

Induction of Diabetes and Drug Treatment
Diabetes was induced in two rodent species with multiple subtoxic doses of streptozotocin (SZ, S-0130, Sigma, St Louis, MO), (20 mg/kg b. w./day in rats, and 40 mg/kg b. w./day in mice), given i.p. for 5 consecutive days. Pentoxyfilline (PTX, Panfarma, Belgrade, Yugoslavia) was given i.m. at a dose of 200 mg/kg/day, from day 0 through day 9 in relation to the induction of diabetes. Control, nontreated animals received injections of PBS. Plasma glucose was determined by a glucose-oxidase method using a glucometer (Glucotronic C; Macherey-Nagel, Duren, Germany) once a week for up to 8 weeks. Clinical diabetes was defined by hyperglycemia in nonfasted animals (blood glucose > 11 mmol/1).

Histological and Immunohistochemical Analyses of Pancreas
For histology, pancreata were prepared by embedding in paraffin after fixing in formalin. To assess the incidence of inflammatory changes, degree of islet cell destruction and changes of connective tissue, histologic sections (7 gm in thickness) were stained with hematoxylin and eosin. Histological analysis was per-formed in a blind fashion by two observers. The degree of inflammatory changes was graded according to the following arbitrary scale: 0 intact islet with no cellular infiltrates, 1 few infiltrated intraislet cells, but intact islet architecture, and 2-heavy mononuclear infiltration with or without loss of islet architecture. At least 5 animals per each experimental group, and minimum 20 islets per animal were scored individually.

Statistical Analysis
Data were expressed as means + s.e.m. Student's test was used for evaluation statistical significance of differences between groups. P-values were considered significant at P< 0.05.