Virally Induced Inflammation Triggers Fratricide of Fas-Ligand–Expressing β-Cells

Generation of H-2^b RIP-LCMV-NP transgenic mice used for this study has been described previously (19). RIP-Fas-L mice expressing Fas-L under the control of the rat insulin promoter had been generated previously (4). NOD (D^bK^dIA^g7) RIP-Fas-L mice were intercrossed with the H-2^d (D^bK^dIA^d) RIP-LCMV-NP line, resulting in double-transgenic mice. These mice were used for all studies here as littermates.

The presence of transgenic RIP-LCMV-NP or Fas-L sequences was determined by performing two independent standard PCR reactions with genomic DNA obtained from mouse tails as described (4,19).

Virus used was LCMV strain Armstrong (arm), clone 53b. LCMV was plaque purified three times on Vero cells, and stocks were prepared by a single passage on BHK-21 cells. Mice were infected with a single intraperitoneal dose of 10⁵ pfu LCMV-arm unless indicated otherwise (20).

LCMV viral titers of organ homogenates were determined by infection of Vero cells as described elsewhere (18). Briefly, organ homogenates were diluted serially and cultured with Vero cells for 5 days. Viral titers were determined from the number of counted plaques (20).

Blood samples were obtained from the retro-orbital plexus, and plasma glucose concentration was determined using Accucheck III (Roche, Indianapolis, IN). Mice with blood glucose values >300 mg/dl were considered diabetic (24).

Islets were isolated as previously described (25). Briefly, the pancreas was removed, cut into little pieces, and digested with collagenase P (Roche). Islets were purified on histopaque-1077 density gradients (Sigma, St. Louis, MO). The obtained islet-enriched fraction and the remaining portion of the pancreas containing mainly acinar cells were either immediately homogenized in tri-reagent (Molecular Research Center) for subsequent isolation of total RNA or islets were further digested with trypsin (Sigma) to obtain single β-cell suspensions (26).

Total RNA was isolated either from whole pancreas homogenates or from islet or acinar cell–enriched fractions using tri-reagent (Molecular Research Center). RNA was extracted with chloroform followed by isopropanol precipitation and washing with ethanol. Twenty micrograms of total pancreatic RNA was used for hybridization with a ³²P-UTP–labeled multitemplate set containing specific probes for Fas, Fas-L, tumor necrosis factor-α, γ-interferon (IFN-γ), and lymphotoxin-β provided by a commercial kit (Riboquant, mCK-3b and mAPO-3; Pharmingen, La Jolla, CA). The RNase protection assay was carried out according to the manufacturer’s guidelines. The resulting analytical acrylamide gel was scanned using a Storm-860 phospho-imaging system (Molecular Dynamics), and the intensity of bands corresponding to protected mRNAs was quantified using the Imagequant image analysis software (Molecular Dynamics) using L32 as a reference gene (23).

Organs were harvested at week 6 after LCMV infection unless indicated otherwise, immersed in Tissue-Tek OCT (Bayer, Elkhart, IN), and quick frozen on dry ice. Six- to 10-μm tissue sections were cut using a cryomicrotome and sialin-coated “superfrost plus” slides (Fisher Scientific, Pittsburgh, PA). Sections were fixed with 90% ethanol at −20°C, and after washing in PBS an avidin-biotin blocking step was included (Vector Laboratories, Burlingame, CA). Primary and biotinylated secondary antibodies (Vector Laboratories) were reacted with the sections for 30 min each, and color reaction was obtained by sequential incubation with avidin-peroxidase conjugate (Vector Laboratories) and diaminobenzidine-hydrogen peroxide. Primary antibodies were rat anti-mouse CD8a (ly2), rat anti-mouse CD8b (ly3), and anti–LCMV-NP Alexa 1.1.3 antibody (19,27).

LCMV-specific cytotoxic T-cell (CTL) activity in spleen was analyzed in a 5-h in vitro ⁵¹Cr-release assay (19). All samples were run in triplicate. Primary CTL activity was tested by harvesting spleens at day 7 after intraperitoneal infection with 10⁵ pfu LCMV-arm. Splenocytes were co-incubated with major histocompatibility complex–matched (Balb/C17 H-2^d) and mismatched (MC57 H-2^b) target cells that had been loaded with ⁵¹Cr. Target cells were either LCMV-infected, uninfected but coated with the immune dominant LCMV major histocompatibility complex class I peptide NP₁₁₈ (RPQASGVYM), or uninfected and uncoated. For determination of secondary CTL activity, spleens were harvested 6 weeks after LCMV infection and splenocytes were cultured for 8 days on LCMV-infected, irradiated peritoneal exudate cells before testing CTL activity in a ⁵¹Cr release assay. For precursor frequency analysis, spleen cells were serially diluted and cultured in 96-well flat bottom plates in the presence of T-cell growth factor (primarily containing interleukin-2), irradiated LCMV-infected peritoneal exudate cells and spleen feeder cells. After 8–10 days, cultures were assayed for CTL activity on LCMV-infected and uninfected target cells. Precursor frequencies were calculated as described (28).

Spleen and pancreatic draining lymph nodes were harvested at week 6 after LCMV infection. Single-cell suspensions were stimulated with 100 μ/ml interleukin-2 and 2 μg/ml Brefeldin A (Sigma) for 5–16 h at 37°C. Cells were stained for cell surface markers using monoclonal antibodies (mAbs) against CD8 and CD4, permeabilized and fixed with paraformaldehyde/saponin, and stained for intracellular cytokines using fluorescein isothiocyanate–conjugated anti-mouse tumor necrosis factor-α mAb and PE-conjugated anti-mouse IFN-γ mAb (Pharmingen). Cells were acquired and analyzed on a FACSort or FACS-Calibur flow cytometer (Becton Dickinson) using cell quest software (Becton Dickinson) (29).

Littermates from F1 mating of RIP-LCMV-NP (H-2d) × RIP-Fas-L (NOD) transgenic were infected with LCMV. As shown in Fig. 1, LCMV infection resulted in 80% diabetes in RIP-NP⁺ Fas-L⁻ controls within 2–8 weeks. In contrast, type 1 diabetes incidence was significantly reduced in littermates that expressed Fas-L in addition to NP antigen in β-cells. Thus, expression of Fas-L has a clear protective effect in the RIP-LCMV diabetes model. Importantly, systemic numbers of cytotoxic CD8 cells were equivalent when comparing Fas-L⁺ versus Fas-L⁻ NP expressors (Table 1). Comparable numbers of CTL precursors and primary killing activity was detected in spleens. In contrast, locally in the pancreatic draining lymph node, we detected a significant reduction of NP pCTL (Table 1) and, accordingly, a lesser degree of CD8 infiltration in Fas-L⁺NP⁺ protected compared with Fas-L-NP⁺ diabetic mice (Fig. 2A versus B). This reduction of NP CTLs was not only site specific but also antigen specific, because CD8 lymphocytes producing IFN-γ in response to LCMV NP but not glycoprotein peptides were reduced (Table 1). Thus, local expression of Fas-L by β-cells can exert some protective capacity by eliminating autoaggressive CD8 lymphocytes in an islet antigen-specific manner. The antigen specificity of the process argues for the fact that interaction of Fas-expressing LCMV-NP–specific CD8 lymphocytes directly with Fas-L–expressing NP⁺ β-cells leads to their demise.

To our surprise, we observed almost 50% diabetes development in Fas-L⁺, NP-nonexpressing littermates 6–10 months after LCMV infection (Fig. 1). This was not a spontaneous disease comparable to that found in the original Fas-L⁺ NOD lines, because the mice under observation were intercrosses between the original NOD and nondiabetes-prone Balb/c mice, and spontaneous disease without LCMV infection was never seen in these mice (data not shown). It is interesting that the diabetes in the Fas-L⁺ NP⁻ mice was not associated with marked CD8 (Fig. 2C) and other lymphocytic infiltration (not shown) as compared with more extensive CD8 infiltrates in NP⁺ littermates (Fig. 2A and B) or in LCMV-infected RIP-NP Balb/c mice (Fig. 2D). This observation indicated that β-cell fratricide had occurred as a consequence of virally induced inflammation of the pancreas, because LCMV infects the pancreas and the islets directly as shown by immunohistochemical stains of pancreas sections with anti-NP antibody (clone 1.1.3.; Fig. 2E) (19,27). Because earlier studies had demonstrated expression of Fas by β-cells after transfer of diabetogenic clones, we hypothesized that LCMV-induced inflammation of the pancreas would lead to Fas induction by β-cells.

Indeed, we could detect Fas expression on both RNA and protein level by RNase protection assay (RPA) and flow cytometry of isolated islets, respectively. First, Fig. 3A shows that Fas gene expression, as detected by RPA, is seen on islet cells early after LCMV infection (day 4) immediately after viral antigen is first expressed (day 2 after infection) and interferons, chemokines, and other inflammatory mediators are induced. Because Fas upregulation is induced by IFN-γ and can be blocked by neutralizing antibodies (30), we next treated LCMV-infected mice with neutralizing antibodies against IFN-γ and analyzed Fas gene expression by RPA. As expected, Fas gene expression at day 4 after LCMV infection was significantly lower in the pancreatic draining lymph node of mice that received four doses of anti–IFN-γ mAb at days 0, 1, 2, and 3 after infection than in mice that received an isotype-matched control antibody (Fig. 3B). However, no change in Fas gene expression was observed in the pancreas (Fig. 3B). Second, for detection of Fas expression on the β-cell surface, whole islets obtained by collagenase P digestion of pancreata were further treated with trypsin to obtain a single-cell suspension. Islet cells were stained for Fas surface expression by flow cytometry, and viable β-cells were identified through their high autofluorescence detectable in the FL1 channel (Fig. 4A). Increased Fas expression was first noted at day 5 after LCMV infection, and it was >10-fold higher than on β-cells isolated from uninfected mice at day 7 after infection (Fig. 4B).

We demonstrate in this study a dual effect of Fas-L expression on β-cells. On one hand, Fas-L provides immune privilege by eliminating autoaggressive CD8 lymphocytes. This finding would encourage the use of Fas-L to protect β-cells and other cell types/organs from autoimmune attacks and fits well with the known role of Fas-mediated apoptosis in CD4 populations. On the other hand, viral infection of the pancreas but not the islets can lead to induction of Fas on β-cells, which renders them susceptible to Fas/Fas-L–mediated apoptosis. This results in our system in a significant degree of clinically manifest diabetes after LCMV infection. The mechanism of virally mediated Fas upregulation in β-cells is not completely clear at this point but is likely not caused by direct viral infection of islets (Fig. 2E), because in vitro infected islets do not exhibit increased Fas levels (25). Rather, we propose an action of virally induced inflammatory mediators such as nitric oxide (NO) and/or cytokines or viral dsRNA (31). Experimentally, using neutralizing antibodies, we excluded a role for IFN-γ (Fig. 3B) or lymphotoxin-β (not shown), because no change in pancreatic/islet FAS levels was observed after their blockade. It is unlikely that virally induced NO plays a role in islet FAS upregulation as suggested by studies with islets from inducible NO synthase knockout mice (32). Our observation thus underlines an important additional caveat when using topical expression of Fas-L to provide immune protection. Viral infections of the target area might abolish any advantage by inducing Fas on selective cell types. This possibility should be taken into account before using Fas-L expression therapeutically.

This is Publication No. 550 from the Division of Developmental Immunology, La Jolla Institute for Allergy and Immunology (San Diego, CA).

This work was supported by National Institutes of Health Grants U19 AI51973, 2 PO1 AG04342-19, and R29 DK51091 to M.G.V. and Juvenile Diabetes Research Foundation Grant 3-2000-510 to U.C.

We thank Diana Frye for excellent assistance with manuscript preparation.