the Pancreatic Islet-Specific Overexpression - Europe PMC

1 downloads 0 Views 3MB Size Report
changes in the pancreas but not other organs of ... (IDDM) the pancreatic islets and, specifically, the ..... induction of accessory molecules such as MHC class.
American Journal of Pathology, Vol. 145, No. 1, July 1994 Copyright C) American Society for Investigative Pathology

Islet Inflammation and Hyperplasia Induced by the Pancreatic Islet-Specific Overexpression of Interleukin-6 in Transgenic Mice

lain L. Campbell,* Monte V. Hobbs,t Janel Dockter,* Michael B. A. Oldstone,* and Janette Allisont

neogenesis, which may indicate a role in tissue repair. (AmJ Pathol 1994, 145:157-166)

From the Departments of Neuropharmacology* and Immunology,t The Scripps Research Institute, La Jolla, California and Thymus Biology Unit,* 7Te Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia

Interleukin-6 (IL-6) is a multifunctional cytokine that is involved in the regulation of the immune response, hematopoiesis, inflammation, and cellular differentiation (see Refs. 1 and 2 for reviews). IL-6 is known to be produced by a variety of cell types including fibroblasts, endothelial cells, macrophages, and epithelial cells in response to other cytokines such as IL-1 and TNF-a or by infection with certain viruses. Elevated production of IL-6 has been reported in autoimmune (eg, rheumatoid arthritis)3 and infectious (eg, HIV infection)4 diseases as well as after traumainduced injury.5 While the exact role of IL-6 in these diverse disorders is not known, it is likely that this cytokine may contribute to both the development of inflammation as well as to repair of tissue injury. In type 1 or insulin-dependent diabetes mellitus (IDDM) the pancreatic islets and, specifically, the islet ,B cells become the target of an autoimmune response which destroys their insulin-producing capacity.6,7 Cytokines have been implicated in the development of the characteristic islet mononuclear cell infiltrate (termed insulitis) and may also mediate f cell toxicity in IDDM.8-10 IL-6 has been shown to be produced by isolated mouse islets after exposure to interferon-y (IFN-y) and tumor necrosis factor-a (TNF-a). 1 1 A correlation between the islet production of IL-6 and progression to development of autoimmune diabetes in the nonobese diabetic (NOD) mouse has been demonstrated.12'13 Finally, administration of an anti-lL6 neutralizing monoclonal antibody to cyclophosphamide-treated NOD mice was found to suppress the development of diabetes.12

Interleukin-6(IL-6) is thought to be involved in the pathogenesis of autoimmune insulin-dependent diabetes meUlitus. To examine this possibility, we developed two lines of transgenic mice (termed

RIP-IL6) which overexpressed IL-6 in the pancreatic islet (3 ceUs. RIP-IL6 mice, while showing a modest reduction in body weight, remained normoglycemic throughout their lives. Furthermore, insulin gene expression and glucose tolerance were similar to non-transgenic littermates. Histopathological examination revealed significant changes in the pancreas but not other organs of RIP-1IL6 animals, with marked alterations in the architecture of the islets, in the islet ceUs, and in surrounding tissues. In younger animals these changes included islet hyperplasia with increased mitotic figures, neo-ductularformation, fibrosis, and a scant mononuclear cell infiltration (insulitis). In addition, immunostainingfor islet hormones revealed changes in both the topography and density of 1 and a ceUs. In older RIP-IL6 mice, a moreflorid insulitis was observed which was composedpredominantly ofB220+ B lymphocytes and, to a lesser extent, Mac-i + macrophages and CD4+ and CD8+ T lymphocytes. Immunostaining for mouse IgG revealed signfifcant numbers of plasma ceUs in the pern-islet infiltrates, which suggested that IL-6 induced differentiation of the recruited B lymphocytes. Therefore, islet overexpression of IL-6 produces a complex, localized host response implicating this cytokine in not only inflammatory processes that occur in autoimmune diabetes but also celular

Supported by USPHS grants MH50456 (ILC), NS12428 (MBAO), AG09822 (MVH), AG04342 (MBAO) and the Whittier Endowment Funds. Accepted for publication March 28, 1994. Address reprint requests to lain L. Campbell, Ph.D., Department of Neuropharmacology, CVN9, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, CA 92037.

157

158

Campbell et al

AJPJuly 1994, Vol. 145, No. 1

Taken together, the above findings indicate that the localized production of IL-6 by the islet may play an important role in the development of the autoimmune disease that results in the destruction of the islet cells. Using a transgenic approach, we have now overexpressed this cytokine in islet cells and show that its production is sufficient to induce a chronic islet-associated inflammatory response, although this did not cause IDDM. Evidence is also provided that islet overexpression of IL-6 is associated with increased proliferation of islet, ductal, and fibroblast cells, implicating this cytokine in tissue repair.

Materials and Methods Transgenic Mice A full-length cDNA encoding murine IL-614 was linked to the rat insulin promoter (RIP) as follows. The BamHI-Xbal fragment of pRIP1-Tag15 containing the rat insulin 11 promoter from -695 to +8 was subcloned into the BamHI-Xbal site of pIC20H to give the vector plC-RIR 16 A PsI-Sspl fragment containing the complete murine IL-6 cDNA was excised from mlL-6 (a gift from Frank Lee, DNAX, Palo Alto, CA). It was bluntended and subcloned into the blunted Sal 1 site of plC-RIP downstream of the RIP promoter. A SmalEcoRI fragment containing the human growth hormone polyadenylation sequences was blunt-ended and cloned into the Sphl site of pIC-RIP. The RIP-1L6 construct was excised from vector sequences with HindlIl, purified, and microinjected intofertilized eggs of C57BL/6J mice. Transgenic mice were identified by slot-blot analysis of tail DNA using a 32P-labeled hGH cDNA probe.

Analysis of RNA Expression Total RNA was extracted from pancreas by guanidinium thiocyanate and purified by ultracentrifugation as described.17 Poly A+ enriched RNA was prepared from spleen and kidney as described.18 For pancreas, total RNA was extracted by guanidinium thiocyanate and partially purified by acid-phenol extraction before poly A+ RNA preparation. 17 After precipitation in ethanol, the RNA was pelleted by centrifugation, the pellet resuspended in 10 ml of lysis buffer, and poly A+ RNA prepared as described.18 Total RNAs (20 pg each) were separated by electrophoresis through 1% agarose-formaldehyde gels and transferred to nylon membrane (Biotrans, ICN) and hybridized to 32P-labeled cDNA probes: a 0.69-kb EcoRl-BgAl fragment of IL-6 cDNA19; a 0.42-kb

EcoRl-Hindlll fragment of rat insulin cDNA20; a 0.26-kb fragment of /-actin gene fragment21 generated by PCR was provided by M. Nerenberg (The Scripps Research Institute, La Jolla, CA). To examine cytokine gene expression, a RNase protection assay utilizing the ML-1 1 probe set22 was performed as described previously22'23 using 5 pg of poly Al RNA per reaction.

IL-6 Production from Isolated Islet Cultures Islets were isolated by collagenase digestion of whole pancreas from overnight-fasted mice as described previously.12 Groups of 25 islets were placed in 0.5 ml of RPMI 1640 medium containing 10% FCS and maintained at 37 C in an atmosphere of 5% CO2 in air. Culture supernatants were collected at 48 hours and analyzed for IL-6 production. IL-6 production was measured by bioassay with IL-6-dependent B9 cells.24 B9 cells were seeded in 96-well plates in growth medium and cultured at 104 cells/well (final volume, 100 p1) in the presence of serial dilutions of culture supernatants or a murine recombinant IL-6 standard (specific activity: 1 x 107 U/pg; Genzyme, Cambridge, MA). After 3 days the number of viable cells was determined by 3-(4,5-diethylthiazol-2 pl)2,5-diphenyltetrazolium (MTT, Sigma, St. Louis, MO) assay25 The minimum detection limit of this assay was 0.5 U/ml.

Histology and Immunocytochemistry For routine histology, pancreas was fixed in Bouin's fixative and embedded in paraffin. Sections (5 p) were processed and stained with hematoxylin and eosin. Adjacent sections were also stained for insulin and IL-6 using an immunoperoxidase technique. Guinea pig anti-insulin antibody (Miles-Yeda, Rehovot, Israel) was detected with a peroxidase-conjugated antiguinea pig Ig (Dako, Santa Barbara, CA) as described previously.26 For IL-6, a rabbit anti-murine IL-6 antibody (kindly provided by Dr. Richard Nordin, NCI, Bethesda, MD) was detected with an anti-rabbit avidin-biotinylated horseradish peroxidase (ABC) kit according to the manufacturers instructions (Vector, Burlingame, CA). In order to identify plasma cells, sections were reacted with a peroxidase-conjugated anti-mouse IgG antibody (Sigma). For cell phenotype analysis, pancreas was removed and immediately frozen in liquid nitrogen. Cryostat cut sections (6 p) were fixed briefly (10 minutes) in ice-cold acetone

Islet Pathology in RIP-1L6 Transgenic Mice

159

AJPJuly 1994, Vol. 145, No. 1

before immunoperoxidase staining with rat monoclonal antibodies (all from PharMingen, San Diego, CA) to B220 (B lymphocyte), Mac-1 (macrophage), CD4 (T lymphocyte), and CD8 (T lymphocyte and NK cells). Primary antibody was detected with a biotinconjugated anti-rat antibody (Southern Biotechnology Associates, Birmingham, AL) followed by peroxidase-conjugated streptavidin (BoehringerMannheim, Indianapolis, IN). For all immunocytochemical reactions, color development employed 3'3' diaminobenzidine (Sigma) as a substrate.

Glucose Tolerance Testing Overnight fasted mice were weighed and bled for fasting blood glucose determination. Each mouse was then injected intraperitoneally with glucose (2 g/kg in PBS) and subsequently bled at 50 minutes for glucose determination. Blood glucose was measured immediately with B-M glucose test strips (BoehringerMannheim) and read in an Accu-Check III (Boehringer-Mannheim) glucose monitoring system.

Results Generation of Transgenic Mice and Expression of Transgene-Encoded IL-6 in the Pancreas Screening of progeny mice by slot-blot analysis of tail DNA revealed five animals positive for the RIP-IL6

8

c

,

transgene. All five founder animals were subsequently bred with C57BL/6 mice and offspring were screened at the RNA level for transgene-encoded IL-6 expression in the pancreas and other organs. Offspring from two founders (designated 71 and 131, respectively) were found to express detectable levels of IL-6 in the pancreas but not in the spleen or the kidney (Figure 1). Because of problems with the breeding of the inbred transgenic mice, offspring from the 71 and 131 founders were subsequently bred with C57BL/6 x SJL (FI) hybrid mice to establish continuous lines of RIP-1L6 mice for further study. Significant production of bioactive IL-6 was detectable in supernatants of cultured islets derived from both RIPIL6 lines which was neutralized by a rabbit antimurine IL-6 antibody (Figure 2). The level of IL-6 production from the islets correlated with the level of pancreatic IL-6 mRNA (Figure 1) and indicated higher expression of transgene-encoded IL-6 in the 71 line compared with the 131 line. With the exception of low levels of IL-1 , mRNA, transcripts for other cytokines, including TNF-a/1, IL-2, IL-3, IL-4, IL-5, IFN-y, and IL-l a, were not detectable in pancreas from the RIPIL6 mice (Figure 1). Immunostaining for IL-6 revealed a majority of islet cells positive for IL-6 and a similar overall distribution of cells staining for insulin (Figure 2). Interestingly, the density of cells staining positive for IL-6 was not as great as that for insulin, suggesting that not all f cells expressed detectable levels of the transgene.

i2

=TNF-j3 ITNF-a -

TNF-, TNF-a =

IL-4

- IL-la

-IFN-y -IL-2

IL-6-

-L-1

11-1 P-

- IL-3

L-132

132-

Pancreas

Spleen

Kidney

Figure 1. Cytokine gene expression in different organsfrom 6-month-old RIP-IL6 and non-transgenic (littermate) mice. For each lane, 5 yg ofpoly A' RNA was analyzed by RNase protection assay (see Materials and Methods) using the ML- 1122 probe set. In addition to the cytokines shown, probe L32 serves as an internal control for RNA loading. Autoradiographic exposure times were 16 hours for pancreas and kidney and 4 hours for spleen. The lane designated ref probe corresponds to the 32plabeled ML-11 probe (5000 cpm loaded). Note the high levels of IL-6 mRNA expressed by pancreas, but not other organs, in the transgenic mice (lineages 71 and 131, respectively).

160

Campbell et al

AJP July 1994, Vol. 145, No. 1

Urn

RPALS

Control

Figure 2. Production and localization of IL-6 in pancreatic islets of RIP-IL6 transgenic mice. Production of IL-6 from isolated islets of Langerhans (top) was assessed using the IL-6dependent B9 cell line. Islets were isolated (2month-old transgenic or non-transgenic mice) and culturedfor 48 hours, at which time supernatants were removed for IL-6 bioassay (see Materials and Methods). Immunolocalization (bottom) of insulin and IL-6 in adjacent pancreas tikssue sections showv localization of IL-6 expressing cells to the islets. Pancreas u'as removed from a 3-month-old RIP-IL6 mouse of the 131 lineage or an age-matched nontransgenic littermate. Original magnifications,

IL4G Production (untibt/sh) +

entf-ILO Ab

5*1

-

15*3

71

+

177036

131

+

112t18

x100.

Islet Inflammation and Hyperplasia in Transgenic Mice Expressing IL-6 in Islet,B Cells Histological examination revealed a number of changes in the pancreas of RIP-1L6 animals with marked alterations in the morphology of the islets and surrounding tissues (Figure 3). At 2 to 3 months of age, the islets were considerably enlarged and often displayed irregular boundaries (Figure 3B). Consistent with their hyperplastic appearance, occasional mitotic figures were also observed in the islets (Figure 3C). Additional features which were suggestive of tissue neogenesis included neo-ductular formation and peri-islet fibrosis. Interestingly, occasional duct cells were found to be positive for insulin by immunostaining (Figure 3D). Islet-associated inflammation (insulitis) was also evident with minor accumulations of mononuclear leukocytes present in peri- and intraislet spaces. The insulitis was progressive with far more extensive infiltrates observed in older RIP-1L6 mice (8-12 months), particularly from the 71 line (Figure 3E). Occasional mitotic figures were seen in cells present in these infiltrates (Figure 3F). Examination of other organs, including spleen, liver, and kidney did not reveal significant changes in RIP-1L6 mice compared with non-transgenic littermates (not shown). Immunocytochemical staining for the islet hormones insulin, glucagon, and somatostatin further highlighted the disrupted architecture of the islets in the RIP-1L6 mice. This was particularly evident for insulin (Figure 2 and Figure 4, A and B), where a reduction in the number and overall intensity of staining was evident. Cells immunostained for glucagon (Figure 4, C and D) but not somatostatin (Figure 4, E and F) also showed an altered distribution and decreased intensity of staining in the transgenic mice.

Ot3 n

1-

insuiin

Infiltrating Cells Are Predominantly B Lymphocytes in RIP-IL6 Mice The islet-associated mononuclear cell infiltrate in older mice (6-8 months) of lineage 71 was analyzed by immunocytochemical staining (Figure 5). The vast majority of cells present in the infiltrates were B220+ B lymphocytes, while significant numbers of Mac-1 + macrophages and CD4+ T lymphocytes were also present. CD8+ cells were the least represented population. With the exception of rare Mac-1 + cells, islets in non-transgenic littermates did not stain for any of the tested markers (not shown). To examine whether the B lymphocytes were differentiating into plasma cells, sections of pancreas were immunostained for mouse IgG. This analysis revealed significant numbers of cells with dense cytoplasmic staining (indicative of IgG positive plasma cells) which were distributed in clusters in the infiltrates surrounding the islets (Figure 6, A and B). Individual plasma cells were also often observed scattered throughout the exocrine tissue (Figure 6A).

Normal Insulin Gene Expression and Glucose Handling in RIP-IL6 Mice To determine whether overexpression of IL-6 by islet ,B cells was associated with changes in insulin gene product expression, an analysis of pancreatic insulin RNA and protein levels was undertaken. As shown in Figure 7, insulin RNA levels were not significantly different between the RIP-1L6 mice and their nontransgenic littermates. Consistent with this finding, total pancreatic insulin protein contents was also unchanged (12.5 ± 2.2 pg/g versus 14.1 + 1.1 pg/g in RIP-1L6 71 versus normal, respectively). Clinically, except for a modest reduction in body mass in the 71

Islet Pathology in RIP-1L6 Transgenic Mice 161 AJPJuly 1994, Vol. 145, No. 1

Figure 3. Islet inflammation and hyperplasia in RIP-IL6 mice. A: Normal appearance of a pancreatic islet in a non-transgenic littermate. B: In RIP-IL6 mice, islets were enlarged, often with an irregular shape and were surrounded hyfibrous tissue and numerous ducts. Both peri- and intraislet infiltration of mononuclear cells utas evident. This specimen was prepared from a 2-month-old mouse of the 71 lineage. A and B, original magnification, x 100. C: Mitotic figures (arrows) were common in the islets of RIP-IL6 mice. Original magnification, x 600. Specimen as for B. D: Immunostaining for inssulin showed scattered insulin-positive cells (arrow) present uwithin the duct walls. Original magnification, X 400. E: More exte(nsive mononuclear cell infiltration uas evident in older RIP-IL6 mice. This specimen from a 6-month-old mouse of the 71 lineage. Original magnification, x 400. F: Mitotic figures (arrow) uere numerous, indicating proliferation of the cells present in the infiltrates. Original magnification, x 600. All panels stained or counterstained with hematoxylin and eosin.

lineage, no significant disease features have been observed over a 1-year period in the RIP-1L6 mice. In particular, random blood glucose levels of transgenic mice were not significantly different from those of nontransgenic littermates. These findings, together with the observation that transgenic mice of the 71 lineage showed normal glucose tolerance (Figure 8), indicated normal glucose handling in the RIP-IL6 mice.

Discussion It has previously been postulated that IL-6 has an important role in the coordination and progression, rather than the initiation, of the anti-f3 cell immunity that leads to the destruction of the insulin-producing cells in IDDM.12,13 In the present study, our findings support this concept by showing that the localized

162

Campbell et al

AJPJuly 1994, Vol. 145, No. I

Figure 4. Immunostaining for insulin (A, B), glucagon (C, D), and somatostatin (E, F). Specimens were prepared from a non-transgenic control mouse (A, C, E) or a 2-month-old RIP-IL6 mouse from the 71 lineage (B, D, F). Note decreased and altered pattern of staining, particularly in the case for insulin, in the RJP-IL6 animal. Original magnifications, x 400.

i

+*e. V

*.