Autoimmune destruction of p-cells is greatly accelerated by transfer of CD4+ and CD8+ T-cells from diabetic NOD mice into newborn recipients (4). Adoptive ...
Accelerated p-Cell Destruction in Adoptively Transferred Autoimmune Diabetes Correlates With an Increased Expression of the Genes Coding for TNF-a and Granzyme A in the Intra-Islet Infiltrates Christoph Mueller, Werner Held, Martin A. Imboden, and Claude Carnaud
Autoimmune destruction of P-cells in nonobese diabetic (NOD) mice is greatly accelerated by adoptive cotransfer of syngeneic CD4 + and CD8 + T-cells from diabetic animals into newborn NOD mice. We followed, by in situ hybridization, the appearance of mRNA of the tumor necrosis factor (TNF)-a gene and, as a marker for activated cytotoxic T-cells, of the serine protease granzyme A gene in the cellular infiltrates generated by cell transfer at birth. Cells expressing the genes for granzyme A or TNF-a were seen in considerable numbers already on day 14, after adoptive transfer. These numbers gradually increased in the intra-islet infiltrates from day 14 through day 30 after adoptive transfer. Compared with our previous findings in NOD mice developing spontaneous insulin-dependent diabetes mellitus (IDDM) (Held W, MacDonald HR, Weissman H,, Hess MW, Mueller C: Genes encoding tumor necrosis factor alpha and granzyme A are expressed during development of autoimmune diabetes. Proc Natl Acad Sci USA 87:2239-2243, 1990), frequencies of cells with TNF-a and granzyme A mRNA were 2and 12-fold higher, respectively, in transferred IDDM (trIDDM). TNF-a mRNA positive cells were predominantly found in the CD4 + T-cell subset of the pancreasinfiltrating cells, whereas granzyme A mRNA positive cells were mainly observed in the CD4~ T-cell subset. The eflfects of the observed enhanced TNF expression upon the pathogenesis of trIDDM are as yet unknown. One may speculate, however, that a local production of TNF-a exerts direct cytotoxicity upon P-cells and promotes lymphocyte traffic to the pancreatic islets, thus resulting in an increased frequency of antigen-specific cytotoxic cells (mainly CD8 + T-cells) within the islets of Langerhans. The burst of activated granzyme A gene-expressing cells at the onset of diabetes further suggests that enhanced cell-mediated cytotoxicity may significantly contribute to the accelerated loss of P-cells. Diabetes 44: 112-117, 1995
From the Department of Pathology (CM., W.H., M.A.I.)i University of Bern, Bern, Switzerland; Ludwig Institute for Cancer Research (W.H.), Lausanne Branch, Epalinges, Switzerland; and INSERM U 25, Hopital Necker (C.C.), Paris, France. Address correspondence and reprint requests to Dr. Christoph Mueller, Department of Pathology, University of Bern, Murtenstrasse 31, CH-3010 Bern, Switzerland. The present address for W.H. is Department of Molecular and Cellular Biology, University of California, Berkeley, Berkeley, CA 94720. Received for publication 15 February 1994 and accepted in revised form 6 October 1994. IDDM, insulin-dependent diabetes mellitus; trIDDM, transferred insulin-dependent diabetes mellitus; IL, interleukin; IFN, interferon; TNF, tumor necrosis factor; EAE, experimental allergic encephalomyelitis; RIP, rat insulin promoter; LCMV-gp, lymphotropic choriomeningitis virus-glycoprotein. 112
N
onobese diabetic (NOD) mice are considered a valuable animal model for human autoimmune endocrinopathies, including insulin-dependent diabetes mellitus (IDDM). Untreated female NOD mice show the first signs of a mononuclear cell infiltration of the islets of Langerhans at 6-7 weeks of age. At ~12 weeks of age, the first animals become overtly diabetic, indicating that —90% of pancreatic p-cells are destroyed (1,2). Incidence rate and first onset of clinically overt diabetes may vary greatly among different NOD mouse colonies and may also depend on maintenance conditions (3). Autoimmune destruction of p-cells is greatly accelerated by transfer of CD4+ and CD8+ T-cells from diabetic NOD mice into newborn recipients (4). Adoptive transfer of splenocytes leads to synchronous insulitis at 2-3 weeks of age and development of clinically overt IDDM 2 weeks later, i.e., at 4-5 weeks of age in 100% of treated animals. During the later stages of insulitis, where specific p-cell loss is most significant, the frequency of CD8 T-cells in the infiltrates increases and the CD4:CD8 ratio becomes —1:1 (5). The molecular cytotoxic mechanisms involved in this autoimmune p-cell destruction are not clearly defined yet. Observations made on in vitro cultured p-cells indicated a possible cytotoxic effect of cytokines, in particular, of the pro-inflammatory cytokine interleukin (IL)-l, either alone (6) or synergistically with interferon (IFN)-7, tumor necrosis factor (TNF)-a, or TNF-p (7). The exact role of TNF-a in the pathogenesis of IDDM is still controversial. TNF-a has been found alternately beneficial or deleterious, according to circumstances, both in BB rats and in NOD mice (8-14). There are also indications for an involvement of activated cytotoxic T-cells in the destruction of P-cells in NOD mice. These include detection of perforin-containing cells in the intra-islet infiltrates by immunohistochemistry (15) and detection of cells containing mRNA of the CTL-associated serine protease granzyme A by in situ hybridization (13). Granzyme A enhances the cytolytic potential of perform and, as such, is critically involved in the DNA degradation frequently observed during cell-mediated cytotoxicity (16-18). TNF and granzyme A, therefore, highlight the two major pathways through which p-cells may possibly be destroyed. In addition, TNF-a may indirectly contribute to T-cellmediated cytotoxicity by upregulating expression of cell adhesion molecules, thus promoting the local recruitment of DIABETES, VOL. 44, JANUARY 1995
C. MUELLER AND ASSOCIATES
cytolytic T-cell precursors. We have shown in a previous study that TNF-a- and granzyme A mRNA-expressing cells increased steadily between 6 and 18 weeks of age in the insulitic infiltrates of unmanipulated female NOD mice. Here, we have attempted to find out whether the explosive disease process that follows the inoculation of diabetogenic T-cells into newborn NOD mice is accompanied by a concomitant burst of TNF-a and granzyme A gene expression in the cellular infiltrates. Our results indicate that transferred IDDM (trIDDM) in neonates is paralleled by a several-fold increase in TNF-a-expressing cells and, even more so, granzyme A-expressing cells, thus suggesting that both mediators do significantly contribute to the dramatic loss of P-cells in this model of IDDM. RESEARCH DESIGN AND METHODS
Donor and recipient animals were kept under pathogen-free conditions in the animal facilities of the Hopital Necker, Paris, France. Ten days after adoptive transfer, they were brought to the animal facilities at the University Hospital in Bern, Switzerland, where they were maintained under conventional conditions. The kinetics of disease progression and time of first onset of glucosuria in the animals used in the present study were comparable to those previously reported by Bedossa et al. (5), where the same protocol of cell transfer was used. This indicates that the transfer of the animals to conventional conditions of maintenance for a limited time did not affect the course of trIDDM. Induction of trIDDM. 20 X 106 enriched splenic T-cells (2+ were considered to be overtly diabetic. According to our experience, such values are correlated with blood glucose concentrations >2 g/1. Mice were killed in groups of five experimental animals and three untreated littermates at 14, 21, 28, and 30 days after birth. Tissue processing. Pancreases and, as control tissues, spleen and a specimen of the intestine were removed, immediately fixed in an H 2500 microwave processor (Bio-Rad, Milan, Italy) for 10 s at 48°C, and then immersed in 4% paraformaldehyde (in 1 X phosphate buffered saline) overnight for subsequent paraffin embedding. Paraffin sections of 5 |xm were either stained with hematoxylin-eosin for histological analysis of insulitis or stored at 4°C for subsequent in situ hybridization with radiolabeled RNA probes of the TNF-a and granzyme A gene, respectively. Probe synthesis. cDNA fragments of the gene coding for granzyme A (19) (provided by Dr. I.L. Weissman, Stanford University, Stanford, CA) and TNF-a (20) (provided by Genentech, South San Francisco, CA) were subcloned into pGEM-1. After linearization of the plasmids, sense and antisense RNA probes were prepared using SP6 and T7 polymerase reactions, respectively, with 35S-labeled CTP as previously described (21). In situ hybridization. Paraffin sections of the pancreas from NOD mice were hybridized with the appropriate RNA probes (2 X 106 cpm/^1 hybridization solution) as previously described (21). Hybridized slides were exposed for 20 days at 4°C. Counterstaining was done with Mayer's hematoxylin or nuclear fast red (0.05% in aluminum sulfate). At least 10 cross-sectioned islets were examined from each animal. Results are expressed as number of positive cells per 100 cross-sectioned islets to allow a direct comparison with a previous study on TNF-a and granzyme A expression in untreated NOD mice (13). Cells were considered positive when they showed twice as many silver grains as control sections hybridized with the relevant sense probe. The latter usually showed 95% between 3 and 4 weeks posttransfer. On day 28 after adoptive cell transfer, one out of five animals became overtly diabetic. Two days later, two out of five transferred animals showed glucosuria values >2+, indicating overt IDDM onset. Untreated littermates displayed no signs of mononuclear cell infiltration throughout the entire observation period (Table 1). The histopathological examination of the glucosuria-negative animals on days 28 and 30 revealed no significant difference in the extent of insulitis. Therefore, data concerning TNF-a and granzyme A gene expression in the pancreas of prediabetic animals killed at these time points were combined.
Islet-infiltrating cells expressing the genes for granzyme A and TNF-a. As shown in Fig. LA, cells with TNF-a or granzyme A transcripts are already seen in considerable numbers within the islet sections on day 14 posttransfer. The number of TNF-a positive cells remained constant or slightly increased during the period preceding clinical onset of IDDM and thereafter dropped abruptly in established diabetic mice. The process was even more dramatic with granzyme A positive cells, because the sharp drop noticed in diabetic mice was preceded by a significant (P < 0.05) burst of positive cells at days 28 and 30, just before presumed onset 113
TNF-tv AND GRANZYME A AFTER T-CELL TRANSFER
750i 750-
B
ve eelIs/ 100 a'OSSlets
• • SCO-
O O
•^ -£ 250-
Granzyme A
500-
250-
i
< .2
< .2 a: ° c o> t
TNF-a
(/>
28-30
14
GLUC0SURIA>2+
AGE (days)
18
GLUCOSURIA >2+
AGE (weeks)
FIG. 1. A: number of TNF-a- and granzyme A gene-expressing cells per 100 cross-sectioned islets of Langerhans of NOD mice transferred at birth with 20 x 10° splenic T-cells. Values were determined for each individual animal from at least 10 pancreatic islets and were averaged from all mice of the same group. T, SE. Numbers of granzyme A-expressing cells were significantly higher (P < 0.05) on days 28-30 than on day 21 and in glucosuria >2 mice (**). TNF-a-expressing cells were significantly more frequent on days 28-30 than in glucosuria >2+ littermates, but no significant difference between days 28-30 and day 21 was found for TNF-a-expressing cells (P > 0.1) (*). Numbers of granzyme A- and TNF-a-expressing cells observed in the intra-islet infiltrates on days 14 and 21 did not differ significantly {P > 0.1). B: number of TNF-a- and granzyme A gene-expressing cells in the intra-islet infiltrates per 100 cross-sections of pancreatic islets of untreated female NOD mice. Mean insulitis grades were 0.3, 1.0, 2.2, and 3.4 for glucosuria-negative female NOD mice at 6, 12, and 18 weeks of age, and glucosuria >2+ female NOD mice, respectively. Data are from Held et al. (13)
of disease. In comparison, one can clearly see lower proportions of TNF-a and granzyme A positive cells in spontaneous insulitis throughout the process leading to overt IDDM (Fig. IB). In situ hybridizations with RNA antisense probes of the TNF-a gene and the granzyme A gene of semi-serial sections from the same islet of Langerhans with numerous specific mRNA-containing cells in the intra-islet infiltrates are shown in Fig. 2A and B, respectively. No significant differences in the frequency of TNF-a- or granzyme A-expressing cells were found between female and male recipients (Table 2). Neither granzyme A nor TNF-a mRNA-containing cells were identified throughout the entire observation period in the islets of Langerhans of littermates left untreated. To obtain some information on the phenotype of TNF-a or granzyme A gene-expressing cells, pancreas-infiltrating cells from a pool of five prediabetic NOD mice at 28 days after adoptive cell transfer were isolated by repeated collagenase digestion and subsequently sorted on the basis of Thy-1.2 and CD4 surface expression. The cell suspension before sorting contained ~40% of Thy-1.2+ T-cells. Separation of the pancreas-infiltrating cells in T-cells (Thy-1.2+) and non-T-cells (Thy-1.2~) revealed a fourfold higher frequency of cells expressing the TNF-a gene (Thy-1.2+ subset 0.4% vs. 0.1% in the Thy-1.2~ subset). Further fractionation into CD4+ and CD4~ Thy-1.2+ cells revealed the preferential expression of the TNF-a gene in the CD4+ T-cell subset, whereas granzyme A-expressing cells are preferentially present in the CD4~ Thy-1.2+ T-cell subpopulation (i.e., mainly CD8+ T-cells) (Fig. 3). TNF-a- and granzyme A-expressing cells are not exclusively restricted to the intra-islet infiltrates. As shown in Table 3, at 28-30 days posttransfer, they are also found, in small numbers, in both the perivascular space and the exocrine part of the pancreas. In situ hybridization of tissue sections of the spleen, which was routinely removed to114
gether with the pancreas, revealed only low numbers of TNF-a- and granzyme A-expressing cells. Most notably, no difference in the number of TNF-a-expressing splenocytes was found between untreated and treated littermates (data not shown).
DISCUSSION In the present study, we show that the accelerated pancreatic p-cell destruction observed after adoptive transfer of splenocytes from diabetic donors can be related to an ~2and 12-fold increase of TNF-a- and granzyme A-expressing cells, respectively, compared with the autoimmune infiltrates observed in untreated female NOD mice developing spontaneous IDDM (Fig. L4 and B). The increase in the number of activated granzyme A mRNA-positive cytotoxic cells after adoptive cell transfer is most pronounced at 28 days of age, i.e., at the eve of clinical onset of IDDM. This prominent increase in the frequency of activated cytotoxic cells within pancreatic islets is in agreement with a previous immunohistochemical study on the cellular composition generated by neonatal adoptive cell transfer. A considerable increase in the number of CD8+ T-cells and activated IL-2R-expressing T-cells was indeed observed at the ultimate stages preceding overt diabetes (5). Besides its recently described involvement in DNA breakdown during cell-mediated cytotoxicity, granzyme A has been implicated in several other biological functions, including degradation of extracellular matrix proteins (24). Thus, it appears likely that secretion of this serine protease contributes to the aggressiveness of certain mononuclear cell infiltrates generated in the course of allograft rejection (21) or autoimmune manifestations (13,26). Comparing the present results with those obtained in unmanipulated NOD mice, we conclude that the immune effector functions of pancreatic DIABETES, VOL. 44, JANUARY 1995
C. MUELLER AND ASSOCIATES
TABLE 2 TNF-a and granzyme A mRNA positive cells in the intra-islet infiltrates of male and female NOD mice after adoptive cell transfer at birth Sex
Median age (days)
Granzyme A mRNA+ cells
TNF-a mRNA+ cells
Male Female
21 24.5
392 ± 125 330 ± 103
150 ± 56 152 ± 54
Data are means ± SE of granzyme A and TNF-a mRNA positive cells per 100 cross-sectioned islets of glucosuria-negative mice with trIDDM.
cytotoxicity, mediated via Fas recognition (25); differential production of cytokines, such as IL-4, IL-12, or IFN-7, which may profoundly affect the course of the immune response, or cytokine-mediated toxicity are operative in either one of these two models. Furthermore, it is possible that younger NOD mice respond differently to TNF than do more mature animals. Such a correlation between the kinetics of autoimmune disease progression and the frequencies of activated cytotoxic cells or, to a lesser extent, of TNF-a-expressing cells in affected tissues has been recently reported in two rat models of experimental allergic encephalomyelitis (EAE). In the more rapidly developing disease, observed after adoptive transfer of encephalitogenic T-cell clones, maximum frequencies of TNF-a- and perforin-expressing cells are increased 2.5- and 25-fold, respectively, compared with those found in rats that develop a less acute form of EAE following primary immunization with myelin basic protein (26). Most likely, both granzyme A and TNF-a mRNA positive cells in the intra-islet infiltrates have been activated in situ within the pancreas. Tissue sections of the spleen from transferred and control littermates showed comparable low frequencies of TNF-a and granzyme A mRNA positive cells. In addition, the level of expression of the two genes, as determined by the number of silver grains per cell, was usually lower in splenocytes than in the infiltrating cells of 1.O1
0.8-
0.6-
FIG. 2. In situ hybridization of pancreas semi-serial sections from an NOD mouse 28 days after adoptive cell transfer with a radiolabeled TNF-a antisense probe (A ), granzyme A antisense probe (J5 ), and granzyme A sense probe (C ) as a negative control. Counterstaining: hematoxylin (.4,2?), nuclear fast red (C ).
O Q.
0.4-
ac 0.2-
p-cell destruction in these two models of IDDM are not fundamentally different. The presence of activated granzyme A- or perforin-expressing cells in the intra-islet infiltrates of untreated female NOD mice indicates that in the spontaneously occurring process, activated cytotoxic cells are also present, but that unlike the situation created by adoptive transfer, their numbers remain low, possibly because regulatory populations prevent their extensive differentiation and activation. However, we cannot exclude the possibility that additional immune effector pathways, such as cell-mediated DIABETES, VOL. 44, JANUARY 1995
Thy-1.2+ CD4+
Thy-1.2 + CD4"
FIG. 3. Frequency of TNF-a- and granzyme A-expressing cells in the Thy-1.2+ CD4+ and Thy-1.2+ CD4~ cell subsets of pancreas-infiltrating cells. Cells were isolated from five pooled pancreases of glucosuria negative NOD mice, 28 days after adoptive cell transfer. A FACStar Plus was used for cell sorting, and in situ hybridization was performed with the respective radiolabeled RNA probes. 115
TNF-iv AND GRANZYME A AFTER T-CELL TRANSFER
TABLE 3 Local distribution of TNF-a and granzyme A mRNA positive cells in the pancreas of NOD mice after adoptive cell transfer at birth
RNA probe TNF-a antisense Granzyme A antisense Granzyme A sense (negative control)
Localization of mRNA positive Total cells (in % of total) positive cells Intra- Perivascular/ Exocrine tissue (=100%) islet periductular 440 898
79 76
0
0
18 16 0
0
Localization of TNF-a- and granzyme A-expressing cells in the pancreas of a total of seven prediabetic NOD mice 28-30 days posttransfer. Three pancreas sections were examined from each prediabetic animal, and the respective locations of granzyme A- and TNF-a gene-expressing cells were recorded. the same animals (data not shown). This indicates that the presence of TNF-a and granzyme A mRNA can be ascribed to neither a systemic activation of leukocytes and subsequent antigen-independent infiltration of the islets nor a passive transfer of pre-activated cells. The observed expression of TNF-a and granzyme A genes in the intra-islet infiltrates more than 14 days after adoptive transfer indicates that inflammatory cells are stimulated in the recipient animals since the mRNA for granzyme A and TNF-a present in splenic T-cells at the time of adoptive cell transfer are rapidly degraded in the absence of stimulation because of their short half-lives (19,23). Furthermore, in cytospin preparations of splenocytes or tissue sections of the spleen from diabetic NOD mice, only rarely are TNF-a- or granzyme A-expressing cells found (our unpublished observations). The role of TNF-a in the pathogenesis of IDDM remains controversial (10). Reports on both a disease-promoting and a disease-suppressing effect of TNF-a exist. TNF-a enhances the cytotoxic effect of IL-1 against pancreatic (3-cells in vitro, either alone (6) or in combination with IFN-7 (7). On the other hand, systemic administration of TNF can prevent onset of IDDM in BB rats (8) and NOD mice (9). Recently, data are accumulating on a central role for TNF-a in the recruitment of inflammatory cells to the islets of Langerhans. Expression of ICAM-1 was induced both at the gene and the protein level on isolated human islet cells by in vitro treatment with IFN-7 and/or TNF-a (27). Mice carrying a TNF-a transgene under the control of the rat insulin promotor (RIP) (11,12) showed signs of peri-insulitis and insulitis quite similar to the initial stages of inflammatory infiltration of the islets of Langerhans in NOD mice, but did not become overtly diabetic. Indications for a possible role of TNF-a in the pathogenesis of IDDM were recently obtained by Ohashi et al. (28), who crossed transgenic mice expressing the lymphotropic choriomeningitis virus-glycoprotein (LCMVgp) under the control of the RIP with mice carrying a TNF-a transgene under the control of the same promotor (11). These Fl animals became diabetic when infected with only the recombinant vaccinia virus expressing LCMV-gp, whereas it required the full LCMV-WE strain to induce IDDM in the parental LCMV-gp transgenic line of mice. Expression of TNF-a in situ may have promoted local lymphocyte traffic, thus increasing the number of LCMV-gp-specific T-cells within the islets (28). Another possible explanation of how locally produced TNF-a might affect pancreatic p-cell lysis has been recently proposed. In synergy with IFN-7, TNF-a 116
can trigger an inducible nitric oxide synthase in pancreatic P-cells, a property that TNF-a and IL-ip have in common (29). In conclusion, the accelerated autoimmune destruction of P-cells observed after neonatal adoptive transfer of splenic T-cells from diabetic donor mice is paralleled by an increased frequency of TNF-a- and granzyme A-expressing cells in the intra-islet infiltrates. One may speculate that besides its putative deleterious effects on pancreatic p-cells, the local production of TNF-a, mainly by CD4+ T-cells, promotes lymphocyte traffic to the pancreatic islets, thus resulting in an increased frequency of antigen-specific cytotoxic cells (mainly CD8+ T-cells), as evidenced by the significant burst of granzyme A positive cells within the islets of Langerhans. Enhanced secretion of granzyme A may also facilitate mononuclear cell infiltration because of its described capability to degrade extracellular matrix proteins, thereby facilitating access of cytotoxic effector cells to their target cells, i.e., (3-cells. Increased numbers of granzyme A positive cells may subsequently result in a more efficient autoimmune elimination of (3-cells. ACKNOWLEDGMENTS This work was supported by grants from the Swiss National Science Foundation to CM. and from the Institute National de la Sante et de la Recherche Medicale to C.C. We would like to thank Dr. I.L. Weissman and Genentech for providing cDNAs, Drs. H.J. Altermatt and L. Mazzucchelli for assistance in the pathological evaluation of the tissue sections, and Martine Olivi and Therese Perinat for expert technical assistance. REFERENCES 1. Gepts W: Pathological anatomy of the pancreas in juvenile diabetes mellitus. Diabetes 14:619-633, 1965 2. Ohneda A, Kobayashi T, Nikei J, Tochino Y, Kanaya H, Makino S: Insulin and glucagon in spontaneously diabetic non-obese mice. Diabetologia 27:460-463, 1984 3. Pozzilli P, Signore A, Williams AJK, Beales EP: NOD mouse colonies around the world: recent facts and figures. Immunol Today 14:193-196, 1993 4. Bendelac A, Carnaud C, Boitard C, Bach JF: Syngeneic transfer of autoimmune diabetes from diabetic NOD mice to healthy neonates. J Eocp Med 166:823-832, 1987 5. Bedossa P, Bendelac A, Bach J-F, Carnaud C: Syngeneic T-cell transfer of diabetes into NOD newborn mice: in situ studies of the autoimmune steps leading to insulin-producing cell destruction. Eur J Immunol 19:19471951, 1989 6. Mandrup Poulsen T, Bendtzen K, Dinarello CA, Nerup J: Human necrosis factor potentiates human interleukin 1-mediated rat pancreatic beta-cell cytotoxicity. J Immunol 139:4077-4082, 1989 7. Pukel C, Baquerizo H, Rabinovitch A: Destruction of rat islet cell monolayers by cytokines: synergistic interactions of interferon-gamma, tumor necrosis factor, lymphotoxin, and interleukin-1. Diabetes 37:133136, 1988 8. Satoh J, Seino H, Abo T, Tanaka S, Shintani S, Ohta S, Tamura K, Sawai T, Nobunaga T, Oteki T, Kumagai K, Toyota T: Recombinant human tumor necrosis factor alpha suppresses autoimmune diabetes in nonobese diabetic mice. J Clin Invest 84:1345-1348, 1989 9. Jacob CO, Aiso S, Michie SA, McDevitt HO, Acha Orbea H: Prevention of diabetes in nonobese diabetic mice by tumor necrosis factor (TNF): similarities between TNF-alpha and interleukin 1. Proc Natl Acad Sci USA 87:968-972, 1990 10. Jacob CO: Tumor necrosis factor alpha in autoimmunity: pretty girl or old witch? Immunol Today 13:122-125, 1992 11. Higuchi Y, Herrera P, Muniesa P, Huarte J, Belin D, Ohashi P, Aichele P, Orci L, Vassalli JD, Vassalli P: Expression of tumor necrosis factor-a in murine pancreatic p-cells results in severe and permanent insulitis without evolution towards diabetes. J Exp Med 176:1719-1731, 1992 12. Picarella DE, Kratz A, Li C-B, Ruddle NH, Flavell RA: Transgenic tumor necrosis factor (TNF)-a production in pancreatic islets leads to insulitis, not diabetes. J Immunol 150:4136-4150, 1993 13. Held W, MacDonald HR, Weissman IL, Hess MW, Mueller C: Genes DIABETES, VOL. 44, JANUARY 1995
C. MUELLER AND ASSOCIATES
14. 15.
16. 17. 18. 19. 20. 21.
encoding tumor necrosis factor alpha and granzyme A are expressed during development of autoimmune diabetes. Proc Natl Acad Sci USA 87:2239-2243, 1990 Jiang Z, Woda, BA: Cytokine gene expression in the islets of the diabetic Biobreeding/Worchester rat. J Immunol 146:2990-2994, 1991 Young LH, Peterson LB, Wicker LS, Persechini PM, Young JD: In vivo expression of perform by CD8+ lymphocytes in autoimmune disease: studies on spontaneous and adoptively transferred diabetes in nonobese diabetic mice. J Immunol 143:3994-3999, 1989 Griffiths G, Mueller C: Expression of perform and granzymes in vivo: potential diagnostic markers for actively cytotoxic cells. Immunol Today 12:415-419, 1991 Shiver JW, Su L, Henkart P: Cytotoxicity with target DNA breakdown by rat basophilic leukemia cells expressing both cytolysin and granzyme A. Cell 71:315-322, 1992 Talento A, Nguyen M, Law S, Wu JW, Poe M, Blake JT: Transfection of mouse cytotoxic T-lymphocytes with an antisense granzyme A vector reduces lytic activity. J Immunol 149:4009-4015, 1992 Gershenfeld HK, Weissman IL: Cloning of a cDNA for a T-cell specific serine protease from a cytotoxic T-lymphocyte. Science 232:854-858, 1986 Pennica D, Hayflick JS, Bringman TS, Palladino MA, Goeddel DV: Cloning and expression in Escherichia coli of the DNA for murine tumor necrosis factor. Proc Natl Acad Sci USA 82:6060-6064, 1985 Mueller C, Gershenfeld HK, Lobe CG, Okada CY, Bleackley RC, Weissman IL: A high proportion of T-lymphocytes that infiltrate H-2 incompatible heart allografts in vivo express genes encoding cytotoxic cell-specific
DIABETES, VOL. 44, JANUARY 1995
22. 23. 24. 25. 26.
27. 28.
29. 30.
serine proteases, but do not express the MEL-14-defined lymph node homing receptor. J Exp Med 167:1124-1136, 1988 Brunstedt J, Nielsen JH, Lernmark A: Methods in Diabetic Research. Vol. 1, part B. New York, Wiley, 1985, p. 245-258 Shaw G, Kamen R: A conserved AU sequence from the 3'-untranslated region of GMF-CSF mRNA mediates selective mRNA degradation. Cell 46:659-667, 1986 Kramer MD, Simon MM: Are proteinases functional molecules of Tlymphocytes? Immunol Today 8:140, 1987 Rouvier E, Luciani M-F, Golstein P: Fas involvement in Ca2+-independent T-cell-mediated cytotoxicity. J Exp Med 177:195-200, 1993 Held W, Meyermann R, Qin Y, Mueller C: Perforin and tumor necrosis factor-a in the pathogenesis of experimental allergic encephalomyelitis: comparison of autoantigen induced and transferred disease in Lewis rats. J Autoimmun 6:311-322, 1993 Campbell IL, Cutri A, Wilkinson D, Boy AW, Harrison LC: Intercellular adhesion molecule 1 is induced on isolated endocrine cells by cytokines but not by reovirus infection. Proc Natl Acad Sci USA 86:4282-4286,1989 Ohashi PS, Oehen S, Aichele P, Pircher H, Odermatt B, Herrera P, Higuchi Y, Buerki K, Hengartner H, Zinkernagel R: Induction of diabetes is influenced by the infectious virus and local expression of MHC class I and tumor necrosis factor-a. J Immunol 150:5185-5194, 1993 Farrar M, Schreiber RD: The molecular cell biology of interferon-7 and its receptor. Annu Rev Immunol 11:571-611, 1993 Miller BJ, Appel MC, O'Neil JJ, Wicker LS: Both the Lyt-2+ and L3T4+ T-cell subsets are required for the transfer of diabetes in nonobese diabetic mice. J Immunol 140:52-58, 1988
117
