Immunological and Pathological Consequences of Mutations in both Fas and Fas Ligand

CELLULAR IMMUNOLOGY ARTICLE NO.

186, 8 –17 (1998)

CI981290

Immunological and Pathological Consequences of Mutations in both Fas and Fas Ligand1 Jory P. Weintraub,* Virginia Godfrey,† P. Anne Wolthusen,* Robert L. Cheek,* Robert A. Eisenberg,‡ and Philip L. Cohen*,2 *Departments of Medicine and Microbiology/Immunology and †Department of Pathology, University of North Carolina, Chapel Hill, North Carolina 27599-7280; and ‡Department of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104-6100 Received November 3, 1997; accepted March 24, 1998

phoid tissues in the body. Defects in apoptosis can lead to autoimmunity, cancer, and developmental and neurological abnormalities (1– 4). A variety of stimuli can initiate the apoptotic process, including irradiation, growth factor deprivation, and ligation of the appropriate death-signaling receptors (5). Two key components involved in regulating apoptosis of mammalian cells are the interacting proteins Fas (CD95) and Fas ligand (FasL)3 (6, 7). Fas is a 35-kDa, type I transmembrane protein belonging to the TNFreceptor family and is expressed in a wide variety of tissues, including thymus, spleen, lymph nodes, liver, lung, kidney, ovary, and heart (8). FasL is a 40-kDa, type II transmembrane protein which belongs to the TNF family and is expressed primarily on activated T lymphocytes and NK cells (9, 10), although it has also been shown recently on B cells (11), in the testes (12), and in the anterior chamber of the eye (13). FasL can also be shed from the cell surface and exist functionally as a soluble trimer (14, 15). Interaction of Fas and FasL (or cross-linking of Fas by monoclonal antibody) can result in rapid apoptosis of susceptible Fas-bearing cells (16). Anti-Fas treatment of mice results in massive liver damage due to apoptosis and death within 24 h (17). That Fas signaling is crucial to the maintenance of immunological tolerance is evidenced by mice homozygous for mutations in the genes for Fas or FasL (lpr (18) or gld (19) mutations, respectively). These mice spontaneously produce T-dependent, lupus-like autoantibodies and develop severe lymphadenopathy, primarily due to the expansion of a unique set of T cells which are CD3low, Thy1low, B2201, CD42, and CD82 (double negative (DN) T cells) (20). These DN T cells are nonmalignant and show little function under ordinary experimental conditions (21, 22). In mice 5

The lpr mutation in mice results in premature termination of transcription of the gene encoding the apoptosis-signaling receptor Fas. As a result, lpr mice develop severe lymphoproliferation and lupus-like autoantibodies. Growing evidence suggests that the lpr mutation is ‘‘leaky’’ and that low levels of Fas are expressed in lpr mice. To determine if Fas expressed in lpr mice is contributing to apoptosis we generated a novel strain of mice (B6/lprgld) which is homozygous for both the lpr mutation and the gld mutation which encodes nonfunctional Fas ligand (FasL) protein. If low levels of Fas in B6/lpr mice contribute to apoptosis and lessen the severity of disease, the B6/lprgld mice, which also lack functional FasL, would be expected to develop a more severe form of disease than B6/lpr mice. Our results revealed no significant increase in either lymphoproliferation or autoimmunity in B6/ lprgld mice compared to B6/lpr or B6/gld mice. Additionally, no increase in surface expression of Fas was detected by flow cytometry on B6/lprgld lymphocytes compared to B6/lpr lymphocytes. However, histological examination of B6/lprgld liver revealed a marked increase in lymphocytic infiltration, compared to livers of B6/lpr and B6/gld mice. Our results suggest that, while low levels of Fas in lpr mice may not be contributing to amelioration of lymphoproliferation or autoimmunity, they may be partially protecting the liver from abnormalities which arise in the absence of Fasmediated apoptosis. © 1998 Academic Press

INTRODUCTION Apoptosis, or programmed cell death, is crucial to the maintenance of immunological tolerance and to the normal development and function of many nonlym1

This work was supported by USPHS Grants AR42573, AR33887, and AR07416. 2 To whom correspondence and reprint requests should be addressed at Division of Rheumatology and Immunology, CB No. 7280, University of North Carolina at Chapel Hill, Chapel Hill, NC 275997280. 0008-8749/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.

3 Abbreviations used: B6, C57BL/6; B6/gld, C57BL/6-gld/gld; B6/ lpr, C57BL/6-lpr/lpr; B6/lprgld, C57BL/6-lpr/lpr-gld/gld; FasL, Fas ligand; ETn, early transposable element; TNP, trinitrophenyl; LN, lymph node; DN, double negative; ANOVA, analysis of variance; RF, rheumatoid factor; H&E, hematoxylin and eosin.

8

CONSEQUENCES OF MUTATIONS IN BOTH Fas AND FasL

months of age and older, they make up over 80% of the cells in the lymph nodes and cause lymph node weight to increase to greater than 20 times normal. Phenotypically, the diseases caused by the lpr or gld mutations are virtually indistinguishable; however, both mutations are characterized by significant variability in severity of disease among individuals, even within a single litter of mice (20). The gld mutation is a point mutation which results in a single amino acid change, from phenylalanine to leucine, in the region of FasL which interacts with the extracellular domain of Fas (19). Although FasL is expressed at comparable levels on the surface of activated T cells from both gld and normal mice (23), gld FasL is unable to interact with a Fas.Fc fusion protein or induce apoptosis in Fas-bearing cells in vitro (24). The lpr mutation is caused by insertion of a retrotransposon (the early transposable element, or ETn) in the second intron of the fas gene, resulting in aberrant transcription of this gene (25). Accumulating evidence suggests that the lpr mutation is ‘‘leaky.’’ Messenger RNA studies have revealed reduced, but detectable, levels of wild-type Fas message (present at 2 to 10% of normal levels) (26, 27). More recently, a number of laboratories including our own have demonstrated Fas protein expression in lpr mice, although primarily in thymocytes (28, 29) or in apoptotic peripheral lymphocytes which have been subjected to irradiation or heat shock followed by in vitro culture (J. Booker et al., submitted for publication). It is still unclear whether live peripheral lymphocytes from lpr mice express functional Fas protein in vivo or when freshly isolated. To address this, we generated a novel strain of mouse, B6/lprgld, by intercrossing C57BL/6-lpr/lpr (B6/lpr) mice with C57BL/6-gld/gld (B6/gld) mice. B6/ lprgld mice are homozygous for both the lpr and gld mutations, and any effect of minimal Fas expression would be negated by a lack of functional FasL. Our results revealed no increase in the severity of lymphoproliferation or autoimmunity in B6/lprgld mice, indicating that any leakiness of the lpr mutation does not serve to ameliorate these aspects of lpr disease. Histological examination of the livers of B6/lprgld mice, however, revealed a significant increase in lymphocytic infiltration, suggesting that the low levels of Fas expressed in lpr mice may be sufficient to provide protection from hepatic abnormalities associated with defective Fas signaling.

9

lpr,gld/gld (B6/lprgld) strain was established in our breeding facility by crossing B6/lpr and B6/gld mice to obtain F1 heterozygotes, crossing F1 mice to obtain F2 mice, and screening F2 mice by polymerase chain reaction (PCR) to identify B6/lprgld breeders which were homozygous for both the lpr and gld mutations. All studies were performed on age-matched male animals which were 5 to 8 months old, an age at which significant lymphoproliferation and autoimmunity are known to exist in B6/lpr and B6/gld mice. PCR Genotyping Genomic DNA was obtained from mouse tails using the QIAamp Tissue Kit from QIAGEN (Chatsworth, CA). The lpr mutation was identified using a standard threeprimer, 50-ml PCR reaction with 1 mM MgCl2, as described previously (30). Thirty amplification cycles were performed at temperatures of 94°C (separation), 55°C (annealing), and 72°C (extension). The three primers were specific for sites flanking and within the ETn which is inserted into intron II of the fas gene in lpr mice (25). The primers yielded a 240-bp product from the wild-type fas gene and/or a 445-bp product from the lpr fas gene, allowing determination of wild-type, lpr, or heterozygous genotype with a single PCR. (Primer sequences were as follows: FAS39: GCA GAG ATG CTA AGC AGC AGC CGG; FAS59: CAA GCC GTG CCC TAG GAA ACA CAG; ETN39: GTG GAG CTC CAA TGC AGC GTT CCT.) The gld mutation was detected using four-primer, allele-specific PCR (31) in a 50-ml reaction volume with 2 mM MgCl2. Thirty-six amplification cycles were performed at temperatures of 92°C (separation), 60°C (annealing), and 72°C (extension). The two primer sets yielded a 350-bp product from the wild-type fas ligand gene and/or a 220-bp product from the gld fas ligand gene, allowing determination of wild-type, gld, or heterozygous genotype with a single PCR reaction. (Primer sequences were as follows: 39FASL: CTC TTG GCC ATT TAA CAT CAG ACA GTT CTT; 59FASLWT2: CTC TGA TCA ATT TTG AGG AAT CTA AGA CGT; 39GLD: CTA TAT GAG GAA CTC TAA GTA TCC TGA GGA; 59GLD1: TTT CTT TTA AAG CTT ATA CAA GCC GAA ACG.) All reactions used AmpliTaq DNA Polymerase, dNTP cocktail, 103 PCR Buffer II, and MgCl2 from Perkin–Elmer (Norwalk, CT). PCR products were separated on a 1.8% agarose gel and visualized by ethidium bromide staining. Antibodies

MATERIALS AND METHODS Mice C57BL/6 (B6), C57BL/6-lpr/lpr (B6/lpr), and C57BL/ 6-gld/gld (B6/gld) strains were originally obtained from The Jackson Laboratory (Bar Harbor, ME) and maintained in our breeding facility at the University of North Carolina School of Medicine. The C57BL/6-lpr/

Anti-IgM-FITC (II/41, rat IgG2a), anti-CD3-PE (1452C11, hamster IgG), anti-CD4-FITC (RM4-5, rat IgG2a), anti-CD8-FITC (53-6.7, rat IgG2a), antiFas-PE (Jo2, hamster IgG), and anti-TNP-PE (UC84B3, hamster IgG) were obtained from PharMingen Inc. (San Diego, CA). Anti-Fcg III/II receptor antibody (2.4G2, rat IgG2b) was obtained from American Type Culture Collection (Rockville, MD).

10

WEINTRAUB ET AL.

Preparation of Cells Mice were bled from the retroorbital sinus and sacrificed by cervical dislocation. Spleens and pooled lymph nodes (cervical, axillary, inguinal, mesenteric, periaortic, and adrenal) were removed, weighed, and transferred to cold, complete medium [RPMI 1640 supplemented with 10% fetal calf serum (Hyclone Laboratories, Logan, UT), 100 U/ml penicillin, 100 mg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate, nonessential amino acids (UNC Cancer Center Tissue Culture Facility), and 5 3 1025 M 2-ME (Sigma Chemical Co., St. Louis, MO)]. Single-cell suspensions were obtained by pressing spleens or lymph nodes through sterile nylon mesh with a 3-cc syringe plunger. RBCs were lysed in NH4Cl for 5 min on ice, and cells were washed two times in cold medium and counted. Immunofluorescence and Flow Cytometry Cells were stained in 96-well microtiter plates at 1 3 106 cells/well. All samples were pretreated with 25 ml of overgrown TC supernatant of anti-Fcg III/II receptor antibody for 15 min at 4°C to block nonspecific binding. Seventy-five microliters of fluorochrome-conjugated antibody (diluted in PBS/0.1% NaN3) was then added to each well (except for unstained, control samples), and cells were incubated for an additional 45 min at 4°C. Cells were then washed two times in cold PBS/ 0.1% NaN3 and fixed in an equal volume of PBS/0.1% NaN3 and 2% paraformaldehyde. Multicolor flow cytometric analyses were performed on a FACScan (Becton–Dickinson & Co., Mountain View, CA) flow cytometer with Cyclops data acquisition software (Cytomation Inc., Fort Collins, CO). In all cases, at least 2 3 104 events were recorded, with gating on the lymphocyte population as determined by forward-scatter and sidescatter profiles. Quantification of Autoantibody Levels by ELISA Mice were bled at the time of sacrifice. Serum samples were assayed in duplicate for IgG anti-chromatin, IgM RF anti-IgG1a, and IgM RF anti-IgG2bb antibodies, as described previously (32, 33). Results are reported in equivalent dilution factors (EDF) of standardized reference MRL/lpr sera, as defined by the formula EDF 5 (dilution of standard reference sera which gives the equivalent OD of the test serum) 3 106. Histology Tissue samples were fixed in 10% neutral buffered formalin and paraffin-embedded or snap-frozen in OCT medium (Miles Laboratories, Naperville, IL). Five-micrometer sections were cut and stained with hematoxylin and eosin. All histologic samples were examined without reference to the genotype of the animal. Samples were scored on a scale from 0 to 4, with 0 being

FIG. 1. Genotyping of B6/lprgld mice confirmed homozygosity for both the lpr and gld mutations. PCR was performed on genomic DNA purified from mouse tail segments, as described under Materials and Methods. In each panel, samples from each of the 14 B6/lprgld used in these studies are contained within the bracketed region, and the first lane (labeled ‘‘het’’) contains DNA from an F1 mouse which is heterozygous for both the lpr and gld mutations. (a) Shown are results from the lpr PCR reaction, in which the wild-type gene yields a 240-bp product and the lpr gene yields a 445-bp product. (b) Shown are results from the gld PCR reaction, in which the wild-type gene yields a 350-bp product and the gld gene yields a 220-bp product.

normal and 4 representing most severe lymphocytic infiltration. Statistical Analysis Results are presented as means 6 SEM. Statistical significance between two groups was determined by Student’s t test. When three or more groups were compared simultaneously, statistical significance was determined by one-way analysis of variance (ANOVA). In all cases, significance was defined as P # 0.05. For each figure, the method of statistical analysis used is indicated in the legend. RESULTS PCR Genotyping of B6/lprgld Mice Figure 1 shows the results of PCR genotyping experiments for the 14 B6/lprgld mice used in these studies, confirming that they were homozygous for both the lpr and gld mutations. Fig. 1a shows the products of the lpr PCR, in which amplification of the wild-type fas gene yields a 240-bp product and amplification of the fas gene possessing the lpr mutation yields a 445-bp product, due to the insertion of the Etn (25). Figure 1b shows the products of the gld PCR, in which amplification of the wild-type fas ligand gene yields a 350-bp product and amplification of the fas ligand gene possessing the gld mutation yields a 220-bp product. In each case, lane 1 (labeled ‘‘het’’) contains a standard

CONSEQUENCES OF MUTATIONS IN BOTH Fas AND FasL

11

consisting of template DNA from an F1 mouse, which is heterozygous at both the fas and fas ligand loci. Measurement of Surface Expression of Fas Recent reports have demonstrated the expression of low, but detectable, levels of Fas protein on the surface of thymocytes from mice with the lpr mutation (28, 29). In contrast, no significant Fas expression has been observed on live, freshly isolated peripheral lymphocytes from lpr mice, suggesting that these cells either do not express Fas or, perhaps due to the overexpression of FasL in the periphery of lpr mice, that any such Fas-bearing cells are rapidly deleted. To address this issue, we stained freshly isolated spleen cells from B6/lpr mice and B6/lprgld mice with anti-Fas antibody or an isotype control. If the leakiness of the lpr mutation in B6/lprgld mice results in Fas expression on peripheral lymphocytes, these cells should persist and be detectable due to the lack of functional FasL. However, as Fig. 2 demonstrates, the anti-Fas and isotype control profiles are virtually indistinguishable for both B6/lpr (Fig. 2a) and B6/lprgld (Fig. 2b) spleen cells, suggesting that neither of these two strains express significant levels of Fas on their peripheral lymphocytes. As a control, freshly isolated spleen cells from a normal B6 mouse were stained with anti-Fas or an isotype control (Fig. 2c) and shown to demonstrate significant expression of Fas on their surfaces. Quantification of Lymphadenopathy The accumulating lymphocytes in lpr and gld mice, which consist mainly of the unusual DN T cells, are concentrated primarily in the spleen and lymph nodes (20). To compare severity of lymphoproliferation between the B6/lprgld mice and single mutants (B6/lpr or B6/gld), we weighed their spleens and pooled lymph nodes, as well as those of normal, B6 mice. Figure 3 shows these values, expressed as a percentage of the total body weight of each animal, to compensate for variations among individuals. No statistically significant difference was found in spleen or lymph node weights between the single and double mutants, suggesting that lymphoproliferation is no more severe when the lpr and gld mutations are combined. The spleens and lymph nodes of normal mice were significantly smaller than those of the single or double mutants, especially in the case of the lymph nodes, where DN T cell accumulation is known to be more profound. The large variation within each mutant strain (as demonstrated by the error bars, each of which represents the standard error of the mean within that group) is consistent with previous findings that the severity of the disease due to the lpr or gld mutations can vary greatly from animal to animal (20). To further assess the degree of lymphadenopathy in these animals, flow cytometric studies were performed on lymphocytes isolated from the spleens and lymph

FIG. 2. Flow cytometric measurement of Fas expression on the surface of freshly isolated splenic lymphocytes from (a) B6/lpr, (b) B6/lprgld, and (c) B6 mice revealed no significant expression of Fas in either B6/lpr or B6/lprgld mice. Each histogram shows an overlay of anti-Fas and isotype control (anti-TNP) profiles. Representative data from one mouse are shown in each case (n 5 6).

nodes to quantify the percentages of various lymphocyte subsets. Figure 4a shows the results of the studies performed on splenic lymphocytes from B6, B6/lpr, B6/

12

WEINTRAUB ET AL.

Examination of Other Tissues in Which Fas Is Normally Expressed In addition to primary and secondary lymphoid tissues, Fas is normally expressed in many other tissues, such as liver, lung, kidneys, brain, and ovary (8). While the role of Fas in these tissues has not been studied as widely as in lymphoid tissues (and no overt pathology is typically observed in these tissues in lpr and gld mice), we wondered if subtle roles for Fas in development and maintenance of homeostasis might be revealed when the lpr and gld mutations were combined. Figure 5 shows the weights of the livers (Fig. 5a), lungs (Fig. 5b), kidneys (Fig. 5c), and thymi (Fig. 5d) of double-mutant B6/lprgld mice and, as controls, agematched single-mutant B6/lpr and B6/gld mice. In each case, organ weight is expressed as a percentage of

FIG. 3. No significant differences were observed in spleen or lymph node weights among B6/lpr (hatched bars), B6/gld (solid bars), and B6/lprgld (stippled bars) mice. B6 spleens and lymph nodes (white bars) were significantly smaller than those from single- or double-mutant animals. In each case, organ weight was expressed as a percentage of total body weight to compensate for variations among individuals. Error bars represent standard error of the mean (SEM). Statistical significance was determined by ANOVA.

gld, and B6/lprgld mice. As with the organ weights, no significant differences were observed between the single- and double-mutant strains with respect to percentages of B cells, CD41 T cells, CD81 T cells, and, importantly, DN T cells (identified as CD31/CD42/CD82), again supporting the conclusion that a combination of both mutations does not affect severity of the disease compared to that seen in single-mutant animals. Figure 4b shows the results of similar studies performed on lymphocytes from the lymph nodes of B6/lpr, B6/gld, and B6/lprgld mice. (Due to the limited number of lymphocytes from lymph nodes of healthy, normal B6 mice, this strain was not included in the lymph node studies.) Again, no significant differences were detected between single- and double-mutant mice with respect to the percentages of various lymphocyte subsets in the lymph nodes of these animals. The extremely high percentage of DN T cells (;60 –70%) in the lymph nodes of the mutant animals compared to the percentages of DN T cells seen in the spleens (;15– 20%) supports previous observations that lymphoproliferation due to the lpr and gld mutations affects the lymph nodes more severely than the spleen (34). Consistent with the immunofluorescence phenotyping results, automated complete blood counts (CBCs) revealed no significant differences among single- and double-mutant mice with respect to the proportions of cells of the lymphoid and myeloid lineages (data not shown).

FIG. 4. Flow cytometric analysis revealed no significant differences among B6/lpr (hatched bars), B6/gld (solid bars), and B6/lprgld (stippled bars) mice with respect to percentages of lymphocyte subsets in the spleen (a) or lymph nodes (b). In the B6 spleen (white bars), there was a significant increase in the percentage of CD81 T cells and a corresponding significant decrease in the percentage of DN T cells (denoted by asterisk; P , 0.05). Error bars represent SEM. Statistical significance was determined by ANOVA and Student’s t test.

CONSEQUENCES OF MUTATIONS IN BOTH Fas AND FasL

13

FIG. 5. No significant differences were observed among B6/lpr, B6/gld, and B6/lprgld mice with respect to weight of (a) liver, (b) lung, or (c) kidney. Thymus weight (d) was significantly higher in B6/lpr mice than in B6/gld mice; however, the B6/lprgld thymi were intermediate and were not significantly larger or smaller than those of either of the single-mutant strains. In each case, organ weight was expressed as a percentage of total body weight to compensate for variations among individuals. Error bars represent standard error of the mean (SEM). Statistical significance was determined by ANOVA and Student’s t test.

total body weight to compensate for variations among individuals. As determined by ANOVA, there were no significant differences among the three strains with respect to liver, lung, or kidney weights. ANOVA did reveal a significant difference in thymus weights among the three strains (P 5 0.02). However, when a Student’s t test was performed to compare two groups a significant difference could be found only between the B6/lpr and B6/gld strains (P 5 0.02; this observation could possibly be a manifestation of the leaky lpr mutation or perhaps a reflection of differing rates of the thymic involution which normally occurs in older mice), while the B6/lprgld thymi were not significantly larger or smaller than either of the single-mutant strains. Comparison of Autoimmunity by Autoantibody ELISA Having shown that a combination of the two mutations did not affect lymphoproliferation or organ weight in Fas-expressing tissues, we next wanted to

determine if autoimmunity was more severe in the double mutants. All mice were bled at the time of sacrifice and ELISAs were performed to quantify IgG anti-chromatin (Fig. 6a), IgM rheumatoid factor (RF) anti-IgG1a (Fig. 6b), and IgM RF anti-IgG2bb (Fig. 6c) antibodies. We observed no significant differences between single- and double-mutant mice with respect to all autoantibodies assayed, suggesting that, as with lymphoproliferation, autoimmunity was no more severe when the lpr and gld mutations were combined than when they occurred individually. As additional controls, sera from normal B6 mice were tested and, as expected, showed significantly lower levels of all three autoantibodies. Additionally, when sera from B6/lpr, B6/gld, and B6/lprgld mice were examined by immunoprecipitation, no antibodies were detected against a variety of autoantigens (Su, nRNP/Sm, threonyl t-RNA synthetase, glycyl t-RNA synthetase, and histidyl t-RNA

14

WEINTRAUB ET AL.

FIG. 7. Histological evaluation of H&E-stained sections revealed a significant increase in lymphocytic infiltration of the livers of B6/lprgld mice compared to those of B6 and B6/lpr mice. The mean histological score of the B6/gld livers was intermediate to those for B6/lpr and B6/lprgld livers; however, it was not significantly larger or smaller than either. All liver sections were graded blindly on a scale of 0 – 4, with 0 representing normal tissue and 4 representing most severe lymphocytic infiltration. Mean histological scores for each group are designated by a horizontal line.

synthetase; data not shown). Finally, when mice were examined for proteinuria no significant differences were detected among the single and double mutants (data not shown). Collectively, these results suggest that autoimmunity in the double-mutant B6/lprgld mice is no more severe than that in the single-mutant B6/lpr and B6/gld mice. Histological Examination of Liver

FIG. 6. Autoantibody ELISA revealed no significant differences among B6/lpr, B6/gld, and B6/lprgld mice with respect to titers of (a) anti-chromatin, (b) IgM rheumatoid factor (RF) anti-IgG1a, and (c) IgM RF anti-IgG2bb antibodies. In each case, B6 mice had significantly lower titers than the single- or double-mutant mice. Results are reported in equivalent dilution factors (EDF) of standardized reference MRL/lpr sera, as described under Materials and Methods. Error bars represent SEM. Statistical significance was determined by ANOVA and Student’s t test.

Histologic examination of tissues from B6/lpr, B6/ gld, and B6/lprgld mice revealed the expected enlargement and effacement of lymph nodes by atypical lymphoid cells (20). Nodular, perivascular infiltrates composed of lymphocytes, macrophages, and plasma cells were found in the portal triads of the livers, and occasionally in the lungs, of animals in all groups. However, infiltrates in the livers of B6/lprgld mice were usually multifocal and far more extensive than those found in livers of mice with only the lpr or gld mutation. Figure 7 shows the results of blinded scoring of histological sections on a scale from 0 (normal tissue) to 4 (most severe lymphocytic infiltration), with each point representing the score of an individual animal. The mean histological scores of the B6 (0.10 6 0.10), B6/lpr (0.80 6 0.25), B6/gld (1.14 6 0.55), and B6/ lprgld (2.36 6 0.58) strains are designated in Fig. 7 by horizontal lines. Infiltrating leukocytes in the B6/ lprgld livers often formed follicular arrangements containing one or more mitotic figures (Fig. 8). Perivascular and peribronchial infiltrates in the lungs were less prominent and lacked the follicular formations seen in the liver. Despite the increase in lymphocytic infiltration in the livers of lprgld mice, we observed neither hepato-

CONSEQUENCES OF MUTATIONS IN BOTH Fas AND FasL

FIG. 8. H&E-stained section of liver from a B6/lprgld mouse revealed multifocal lesions consisting of nodular, perivascular infiltrates (lymphocytes, macrophages, and plasma cells). Mitotic figures could frequently be seen within the follicular arrangements of infiltrating leukocytes.

cellular damage nor damage to bile ducts. In addition, serum levels of aspartate serine transaminase (AST), g-glutamyl transpeptidase (GGT), and total bilirubin were normal in both the single- and double-mutant mice (data not shown), suggesting that liver function was not impaired. It therefore appears that the principal effect of combining the lpr and gld mutations is an increase in lymphocytic infiltration in the liver. DISCUSSION We have generated a novel congenic strain of autoimmune mouse, B6/lprgld, which is homozygous for both the lpr mutation (a leaky mutation in the fas gene caused by a retrotransposon insertion in intron II (25)) and the gld mutation (a point mutation in the extracellular portion of the fas ligand gene (19)). Examination of lymphoproliferation and autoimmunity in these animals allowed us to determine the consequences of the absence of the Fas/FasL system in mice. We were unable to detect significant differences in either lymphoproliferation or autoimmunity among the B6/lpr, B6/gld, and B6/lprgld mice, suggesting that residual Fas/FasL interactions in B6/lpr and B6/gld mice did not contribute to amelioration of disease. The findings give a clearer picture of the significance of Fas/FasL under normal conditions. It is apparent from this study that the complete absence of this receptor–ligand pair is not only compatible with life, but allows normal development and only an age-related autoimmune/lymphoproliferative syndrome in mice with otherwise normal genetic backgrounds, a syndrome nearly indistinguishable from the well-described lpr and gld single mutations (20). The significant increase in lymphocytic infiltration in the liver, however, suggests that the low levels of Fas in B6/lpr mice may afford these animals some protection from liver abnormalities which occur in mice completely lacking in Fas-mediated apoptosis. It is

15

known that the liver is exquisitely sensitive to Fasmediated apoptosis. In vivo, anti-Fas treatment results in massive hepatocyte apoptosis and death of the animal within 24 h, well before any immunological effects can be observed (17). Additionally, liver damage was reported in a wasting-disease model (lpr spleen cell transfer into normal mice) (35) which is thought to be caused by the documented overexpression of FasL on lpr lymphocytes (36, 37). In humans, there is evidence for Fas/FasL involvement in hepatitis B virus-related cirrhosis (38) and chronic hepatitis C infection (39). Therefore, it may be that low levels of Fas in B6/lpr mice simply do not exceed the threshold necessary to prevent lymphoproliferation in the spleen and lymph nodes, while in the liver infiltrating lymphocytes are more susceptible to the effects of extremely low levels of Fas. This would explain the relative absence of infiltrating lymphocytes in B6/lpr mice and their presence in the double-mutant B6/lprgld mice. However, while the low levels of Fas seen in B6/lpr mice may be protecting these animals from lymphocytic infiltration in the liver, the hepatocytes themselves are resistant to anti-Fas antibody treatment in vivo, as demonstrated by Ogasawara et al. (17). Consistent with this, our data from the B6/lprgld mice revealed neither impairment of liver function (as detected by altered liver enzyme levels), nor hepatocyte damage, but we did observe lymphocytic infiltrates in the livers of these animals which are completely lacking in Fas-mediated apoptosis. Interestingly, a statistically significant difference could only be found between the B6/lpr and B6/lprgld livers; while the mean pathology score for the B6/gld livers was intermediate between those for the B6/lpr and B6/lprgld livers, it was not significantly higher than the former or significantly lower than the later. There are several possible interpretations of these results. One is that the Fas expressed in lpr mice is nonfunctional. Our results from the lymphoproliferation and autoantibody studies (no significant differences between B6/lpr and B6/gld) would support this notion, while the liver data (a higher mean histological score for B6/gld livers compared to B6/lpr livers), as well as chimera data from our laboratory (J. Booker et al., manuscript submitted for publication) argue to the contrary. An alternate interpretation of the results of the histological evaluation of B6/lpr, B6/gld, and B6/ lprgld livers (lower mean histological scores for B6/lpr and B6/gld mice than for B6/lprgld mice) is that a low level of functional Fas/FasL interaction is taking place in gld mice. It is also possible that autoimmunity and lymphoproliferation are parameters which are not sufficiently sensitive to detect subtle contributions of Fas/ FasL-mediated apoptosis in these mice. Adachi et al. (40) recently reported mice with a targeted deletion of the Fas gene (Fas knockout mice). In these animals, they observed increased lymphocytic infiltration in the liver and lungs, accompanied by liver

16

WEINTRAUB ET AL.

and lymphoid hyperplasia (41). In addition, lymphoproliferation and autoantibody formation were more severe than in MRL/lpr mice and occurred more rapidly (40). Hyperplasia was not observed in other tissues which normally express Fas, and serum levels of AST, ALT, LDH, and bilirubin were not altered. The Fas KO mice and our B6/lprgld mice both suggest a critical role for Fas in the liver and indicate that even very low levels of Fas can have profound effects in this organ. However, the Fas KO mice showed interstitial infiltrates in the liver as well as perivascular cells, while the B6/lprgld mice did not. Additionally, mitotic figures were more prevalent in the follicular arrangements of the B6/lprgld livers than in those from the Fas KO mice. The Fas KO mice also showed more severe lesions in the lungs. Finally, the Fas KO mice developed more rapid and severe lymphoproliferation and autoimmunity than MRL/lpr controls, while this was not observed when comparing B6/lprgld mice to B6/lpr and B6/gld controls. The Fas KO mice had a chimeric genetic background (containing genes from both 129Sv and C57BL/6 strains) while the lpr mice used for comparison were on an MRL background. The large difference in background, as acknowledged by the authors, may have confounded the comparison. In contrast, the B6/lprgld mice, as well as the B6/lpr and B6/gld controls, were all on an identical genetic background. The only difference, therefore, was the presence of the lpr and/or gld mutations. As background genes are known to significantly affect the levels of lymphoproliferation and autoimmunity observed in animals with lpr or gld mutations, this may explain the discrepancy seen in the Fas KO versus B6/lprgld mice. Wang et al. (42) recently reported the generation of a similar lprgld double-mutant strain on an MRL background. These animals were used to demonstrate that FasL cell-mediated cytotoxicity could be inhibited by bystander Fas expressed on gld T cells, but not by T cells from lpr/gld T cells lacking Fas. While the authors did not describe phenotypic differences between the MRL-lpr/gld mice and the single-mutant parental strains (MRL-lpr and MRL-gld), it would be interesting to examine these animals for differences similar to those seen between our single- and double-mutant mice on a B6 background. In conclusion, our results support a critical role for Fas, even at extremely low levels, in liver homeostasis, while suggesting that the leaky lpr mutation does not contribute to amelioration of lymphoproliferation or autoimmunity in lpr mice. Future studies will take advantage of the novel nature of the B6/lprgld strain to examine unique aspects of lpr mice, such as the wasting disease observed when lpr lymphocytes are transferred to normal mice. The combination of the lpr and gld mutations in one animal should elucidate many long-standing questions associated with lpr mice.

ACKNOWLEDGMENTS We thank Dr. Minoru Satoh for assistance with the immunoprecipitation assays and Dr. Karamarie Fecho for assistance with the automated complete blood counts. We thank Dr. Michael Maldonado and Sarah Baik for providing the protocol and primer sequences for the gld PCR. We thank Dr. Elizabeth Reap for critical reading of the manuscript.

REFERENCES 1. Nagata, S., Genes Cells 1, 873, 1996. 2. Kuchino, Y., and Kitanaka, C., Hum. Cell 9, 223, 1996. 3. Rudin, C. M., and Thompson, C. B., Annu. Rev. Med. 48, 267, 1997. 4. Jacobson, M. D., Weil, M., and Raff, M. C., Cell 88, 347, 1997. 5. Abastado, J. P., Res. Immunol. 147, 443, 1996. 6. Nagata, S., Cell 88, 355, 1997. 7. Nagata, S., and Golstein, P., Science 267, 1449, 1995. 8. Watanabe-Fukunaga, R., Brannan, C. I., Itoh, N., Yonehara, S., Copeland, N. G., Jenkins, N. A., and Nagata, S., J. Immunol. 148, 1274, 1992. 9. Suda, T., Takahashi, T., Golstein, P., and Nagata, S., Cell 75, 1169, 1993. 10. Suda, T., and Nagata, S., J. Exp. Med. 179, 873, 1994. 11. Hahne, M., Renno, T., Schroeter, M., Irmler, M., French, L., Bornard, T., MacDonald, H. R., and Tschopp, J., Eur. J. Immunol. 26, 721, 1996. 12. Bellgrau, D., Gold, D., Selawry, H., Moore, J., Franzusoff, A., and Duke, R. C., Nature 377, 630, 1995. 13. Griffith, T. S., Brunner, T., Fletcher, S. M., Green, D. R., and Ferguson, T. A., Science 270, 1189, 1995. 14. Tanaka, M., Suda, T., Takahashi, T., and Nagata, S., EMBO J. 14, 1129, 1995. 15. Dhein, J., Walczak, H., Baumler, C., Debatin, K. M., and Krammer, P. H., Nature 373, 438, 1995. 16. Nagata, S., Philos. Trans. R. Soc. London Ser. B: Biol. Sci. 345, 281, 1994. 17. Ogasawara, J., Watanabe-Fukunaga, R., Adachi, M., Matsuzawa, A., Kasugai, T., Kitamura, Y., Itoh, N., Suda, T., and Nagata, S., Nature 364, 806, 1993. 18. Watanabe-Fukunaga, R., Brannan, C. I., Copeland, N. G., Jenkins, N. A., and Nagata, S., Nature 356, 314, 1992. 19. Takahashi, T., Tanaka, M., Brannan, C. I., Jenkins, N. A., Copeland, N. G., Suda, T., and Nagata, S., Cell 76, 969, 1994. 20. Cohen, P. L., and Eisenberg, R. A., Annu. Rev. Immunol. 9, 243, 1991. 21. Davignon, J. L., Budd, R. C., Ceredig, R., Piguet, P. F., MacDonald, H. R., Cerottini, J. C., Vassalli, P., and Izui, S., J. Immunol. 135, 2423, 1985. 22. Sobel, E. S., Kakkanaiah, V. N., Rapoport, R. G., Eisenberg, R. A., and Cohen, P. L., Clin. Immunol. Immunopathol. 74, 177, 1995. 23. Hahne, M., Peitsch, M. C., Irmler, M., Schroter, M., Lowin, B., Rousseau, M., Bron, C., Renno, T., French, L., and Tschopp, J., Int. Immunol. 7, 1381, 1995. 24. Ramsdell, F., Seaman, M. S., Miller, R. E., Tough, T. W., Alderson, M. R., and Lynch, D. H., Eur. J. Immunol. 24, 928, 1994. 25. Adachi, M., Watanabe-Fukunaga, R., and Nagata, S., Proc. Natl. Acad. Sci. USA 90, 1756, 1993. 26. Wu, J., Zhou, T., He, J., and Mountz, J. D., J. Exp. Med. 178, 461, 1993.

CONSEQUENCES OF MUTATIONS IN BOTH Fas AND FasL 27. 28. 29. 30. 31. 32. 33. 34. 35.

Chu, J. L., Drappa, J., Parnassa, A., and Elkon, K. B., J. Exp. Med. 178, 723, 1993. Mariani, S. M., Matiba, B., Armandola, E. A., and Krammer, P. H., Eur. J. Immunol. 24, 3119, 1994. Cui, H., Ju, S. T., and Sherr, D. H., Cell. Immunol. 174, 35, 1996. Maldonado, M. A., Eisenberg, R. A., Roper, E., Cohen, P. L., and Kotzin, B. L., J. Exp. Med. 181, 641, 1995. Sommer, S. S., Groszbach, A. R., and Bottema, C. D., BioTechniques 12, 82, 1992. Morris, S. C., Cheek, R. L., Cohen, P. L., and Eisenberg, R. A., J. Immunol. 144, 916, 1990. Sobel, E. S., Katagiri, T., Katagiri, K., Morris, S. C., Cohen, P. L., and Eisenberg, R. A., J. Exp. Med. 173, 1441, 1991. Nagata, S., and Suda, T., Immunol. Today 16, 39, 1995. Theofilopoulos, A. N., Balderas, R. S., Gozes, Y., Aguado, M. T., Hang, L. M., Morrow, P. R., and Dixon, F. J., J. Exp. Med. 162, 1, 1985.

36. 37.

38.

39.

40.

41. 42.

17

Watanabe, D., Suda, T., Hashimoto, H., and Nagata, S., EMBO J. 14, 12, 1995. Chu, J. L., Ramos, P., Rosendorff, A., Nikolic-Zugic, J., Lacy, E., Matsuzawa, A., and Elkon, K. B., J. Exp. Med. 181, 393, 1995. Galle, P. R., Hofmann, W. J., Walczak, H., Schaller, H., Otto, G., Stremmel, W., Krammer, P. H., and Runkel, L., J. Exp. Med. 182, 1223, 1995. Hiramatsu, N., Hayashi, N., Katayama, K., Mochizuki, K., Kawanishi, Y., Kasahara, A., Fusamoto, H., and Kamada, T., Hepatology 19, 1354, 1994. Adachi, M., Suematsu, S., Suda, T., Watanabe, D., Fukuyama, H., Ogasawara, J., Tanaka, T., Yoshida, N., and Nagata, S., Proc. Natl. Acad. Sci. USA 93, 2131, 1996. Adachi, M., Suematsu, S., Kondo, T., Ogasawara, J., Tanaka, T., Yoshida, N., and Nagata, S., Nature Genet. 11, 294, 1995. Wang, J. K. M., Zhu, B., Ju, S. T., Tschopp, J., and MarshakRothstein, A., Cell. Immunol. 179, 153, 1997.