Journal of Autoimmunity (1992) 5 (Supplement A), 133-143
S t u d i e s o n t h e R o l e o f T u m o r N e c r o s i s F a c t o r in Murine and Human Autoimmunity
Chaim
O. Jacob
Department of Inflammation Biology and Immunology, Syntex Research, Palo Hlto, C A 94303, U S A
We have analyzed the roles of t u m o r necrosis factor alpha (TNF-a) in h u m a n systemic lupus e r y t h e m a t o s u s (SLE) and routine models of lupus as well as in type I diabetes in NOD mice. These studies suggest an i m p o r t ant role for TNF-a in the pathogenesis o f a u t o i m m u n e disease. R a t h e r than being involved m a i n l y in the effector a r m of the i n f l a m m a t o r y process of a u t o i m m u n e o r g a n destruction, our data suggest a p r i m a r y involvement in some of the basic m e c h a n i s m s of the a u t o i m m u n e process. Evidence has been presented that emphasizes the possibility of the involvement of this cytokine in the genetic predisposition to SLE. The data m a y i m p l y that the effect of TNF on the i m m u n e system m a y be m o r e relevant to the pathogenesis of the a u t o i m m u n e disease than direct local effects at some target organs. Based on the data presented, one should be cautious in extrapolating the effects of this cytokine in various in vitro systems to the in vivo situation.
M H C class I I - a s s o c i a t e d v a r i a t i o n in t u m o r n e c r o s i s f a c t o r p r o d u c t i o n T h e production of t u m o r necrosis factor ( T N F ) by peritoneal macrophages activated with lipopolysaccharide (LPS) and interferon gamma ( I F N - y ) , obtained from various mouse strains revealed a significant variation in the production of cytokine among the different strains. This variability seemed not to be random but rather M H C associated. For example, Iad mice (BALB/c, N Z B and DBA/2] produce similar intermediate levels of T N F - a while Ia q mice (SWR, DBA/1 and B10.Q) produce very high levels of TNF-ct. : ~: T o extend these studies to humans, the production of T N E by activated peripheral blood lymphocytes or enriched monocyte populations of H L A - A , -B, -C, - D R and - D Q serotyped normal donors was assayed. As in mice, major differences in the level of T N F - a inducibility were observed between different individuals. T N F - a 133 0896-8411/92/0A0133+ 11 $03.00/0
© 1992AcademicPress Limited
134 C.O.Jacob production was found associated with the M H C class II genotype. DR2 and D Q w l positive individuals were found to produce significantly lower levels of T N F than donors not possessing these specificities,. On the other hand normal donors positive for DR3 or DR4 were almost always high producers of the cytokine [1].
TNF in systemic lupus erythematosus (SLE) Because the T N F genes are located within the M H C [2-4], the possibility of the involvement of a 'disregulated' T N F gene in autoimmune disease development is an attractive hypothesis [2]. We have recently presented evidence suggesting that a ( T N F - ~ ) allele contributed by the N Z W mouse strain may be involved in the pathogenesis of lupus nephritis in the (NZB x N Z W ) F 1 mouse [5]. N Z W mice have no autoimmune disease, while the (NZB x N Z W ) F 1 develop severe glomerulonephritis. T h e N Z W parental strain was found to produce very low levels of T N F - m T h e r a p y with recombinant T N F - ~ induced a very signifcant diminution in the development of lupus nephritis in the (NZB x N Z W ) F 1 mouse. We suggest that the N Z W parent may contribute to the F1 disease a T N F - ¢ gene capable of only low levels of T N F production. T h u s , recombinant T N F - ~ may possibly act as replacement therapy suggesting that such treatment may be supplementing the low endogenous T N F production. Similar results were obtained in human SLE: DR2, D Q w l - p o s i t i v e S L E patients show low levels of T N F - ~ production [ 1] and this genotype is also associated with an increased incidence of lupus nephritis [ 1, 6]. However, a substantial portion of S L E patients are DR3 associated and these patients do not have low T N F - ~ production, similar to M R L - l p r / l p r and BXSB mice which also show moderate to high levels of T N F - a production [7]. Taken together these studies suggest that S L E is not a single condition but rather can be subdivided from the genetic point of view into two or more subsets (Figure 1). These data also emphasize that the MHC-associated genetic predisposition may actually proceed through different mechanisms involving different genes in different subsets of the disease. While T N F - ~ significantly delays the development of nephritis in (NZB x N Z W ) F 1 and to a lesser extent also in the M R L - l p r / l p r model, it has no protective effect in BXSB mice [7]. T h e BXSB model is believed to have a totally different genetic basis [8] and this could account for T N F - ~ having no effect on disease development in this strain of mice. An inverse correlation has been found between the level of T N F - a inducibility in peritoneal macrophages in vitro, and the effect of T N F - a administration in vivo. Based on these results, it is possible to predict that T N F - ~ treatment might be beneficial only in situations in which the T N F - ~ inducibility is low. It is also of interest that a 14-week treatment protocol was more effective in delaying the development of nephritis than continuation of T N F - ~ therapy for 27 weeks [7]. We have no explanation for these results, but have found that a n t i - T N F - ~ antibody had not developed in these mice. T h e protective effect of T N F - a in the (NZB x N Z W ) F 1 has been confirmed independently by G o r d o n et al. [9]. On the other hand in two separate publications Kelley and associates argue that rather than a beneficial effect, T N F - ~ may have a deleterious effect in murine lupus. First, Boswell et al. [10] have found increased
T u m o r necrosis factor in a u t o i m m u n i t y
135
Genetic Subsets of SLE
DR 2 (non Caucasian) DQwl (Caucasian) Low TNF Nephritis C4A null (normal genes) High anti-DNA Ab
DR 3 Ro,La Ab High TNF SICCA Syndrome C4A null (gene deletion)
- ve assoc. Nephritis - ve assoc, anti-DNA High TNF
Figure 1. Possible genetic subsets of systemiclupus erythematosus(from Jacob et al. [43]). m R N A expression of interleukin-1 (IL-1) and T N F - a locally in the kidneys of M R L - l p r / l p r mice, suggesting that intra-renal production of these cytokines is responsible for a cascade of events leading to lupus nephritis. In the second paper Brennan et al. [ 11 ] report enhanced T N F - a and IL-1 ]3 expression in the kidneys of (NZB x N Z W ) F 1 mice. Moreover, very low doses of T N F - a (0.2 ~tg/injection) given from 4 to 8 months of age accelerated renal disease and mortality rate. While our basic concept of the potential role of T N F in autoimmune lupus nephritis is clearly different from the above, the data from these studies are not necessarily contradictory. First, we have tested T N F - a protein production by stimulated peritoneal macrophages, while these studies assayed T N F - u m R N A levels locally in the kidneys. In this regard it should be emphasized that despite the fact that the clinical syndrome may be dominated by the antigen-antibody complex induced nephritis, it is very unlikely that the kidney is the initial target organ where the recognition, binding and presentation of the 'autoantigen' is taking place. Moreover, a significant discrepancy between the amount of T N F - ~ m R N A and the amount of T N F - a protein translated has been shown. Thus, in certain instances, T N F - ~ m R N A may be expressed in considerable abundance, while the protein is not expressed or is expressed at a very low level [12]. Second, different doses and ages of mice are used in these studies. We have used doses approximately 10 times higher than Brennan et al. [ 11]. Brennan et al. show an accelerating effect only with very low doses of T N F - ~ and only if treatment is started after 4 months of age and continued until 8 months of age. It is perhaps possible that this deleterious effect seen in 4 to 8 months treatment is related to our observation
136 C.O.Jacob that 3 to 9.5 months treatment is less effective than 3 to 6.5 months treatment [7]. O u r studies show that the earlier T N F - u is given the m o r e effective is its protective effect [7]. It is reasonable to assume that T N F - a is involved in a protective m o d e only in the early induction phases of the a u t o i m m u n e process, rather than in the late effector arm of the disease process. In any case, further investigation is required to understand the basis of differential dose response to TNF-ct. T N F in t y p e 1 d i a b e t e s m e l l i t u s It has been shown that I L - 1 is toxic to pancreatic ]3-cells in vitro [13] and that T N F significantly enhances this toxicity [14]. T h u s , release of I L - 1 and T N F by cells infiltrating the pancreatic islets have been suggested as the effector mechanism of tissue distraction in a u t o i m m u n e diabetes [ 15]. In accordance with this hypothesis, Bendtzen et al. [16] suggested that low T N F production by DR2-positive individuals may explain the DR2-associated resistance to insulin-dependent diabetes mellitus. Surprisingly, and in apparent contradiction with a role for T N F at the effector stage of the disease, we [17] and Satoh et al. [18] independently have shown that systemic administration of T N F - a has a protective effect in the non-obese diabetic ( N O D ) mice. We assessed the in vivo effects of T N F - a in the N O D mice on three levels: insulitis development, development of overt diabetes, and adoptive transfer of diabetes by splenic lymphocytes from diabetic female mice to irradiated young non-diabetic N O D males. First, treatment with T N F - ~ for 4 weeks was sufficient to significantly reduce the incidence of insulitis [ 17]. Second, administration of T N F - a on a long-term basis caused a major reduction in the observed incidence of diabetes [17]. T N F - c t was given at a dose of 3 ~tg/injection ip three times/week for a period of 3 months, starting at the age 8-9 weeks. Mice were followed up to approximately 12 months of age. Even 7 months after stopping the treatment only 26% of mice developed diabetes while about 70% of age matched controls had diabetes during this period of time (P < 0.05). Similar results have been obtained by Satoh et al. [ 18] treating N O D mice with TNF-~t (3 x 103 U) ip twice a week between 4 and 25 weeks of age. Third, T N F - a was effective in preventing the adoptive transfer of diabetes by lymphocytes from diabetic mice to non-diabetic mice [17]. Furthermore, more recently T N F has also been shown to be protective in the other model of autoimmune diabetes, namely the BB rat [19]. In addition in humans no correlation between T N F production and M H C class I I genotype has been found among type 1 diabetic patients [20]. Since we and Satoh et al. have studied in vivo effects while the other groups tested in vitro effects, it is possible that locally T N F and IL-1 may be toxic to pancreatic ]3-cells whereas systemically they may have global immunoregulatory effects, and therefore the two mechanisms may not be mutually exclusive. Since T N F - a has a very short half life w h e n administered in vivo and T N F receptors are believed to be widely distributed on most cell types, it is reasonable to assume that most of the effects we have observed by injecting T N F - ~ intraperitoneally would not be local at the level of the pancreatic ]3-cell. Monoclonal antibodies (mAb) on the other hand, have long half lives ( m a n y days in most cases), therefore we can assume that antiT N F antibodies will block endogenous T N F both systemically and also locally. Using a recently produced anti-murine T N F m A b provided to us by D r Bob
T u m o r necrosis factor in autoimmunity
137
Schreiber, we found that administration of this antibody to N O D females for a period of 8 weeks, increased the incidence of insulitis significantly compared with age-matched female N O D mice receiving non-relevant antibody or PBS [21 ]. Since mAb to T N F might be expected to block T N F locally as well as systemically, these results argue in favor of T N F having a net protective effect in disease pathogenesis. TNF gene polymorphism
Polymorphic markers are essential tools for studying the potential disease linkage to the T N F genes. At the R F L P level a two-allele polymorphism has been identified [5]. However, analysis of this polymorphism in multiple strains of mice failed to show correlation with T N F - a production and linkage between murine lupus and the T N F - a gene. On the other hand, using the P C R methodology to clone and sequence 5' regulatory regions of the T N F - a gene in various mice strains, an AC dinucleotide repeat was identified, which is polymorphic in length among different strains of mice and thus represents a new potential genetic marker. Variation in the length of this so-called microsatellite can be directly demonstrated by amplifying the D N A immediately flanking the repeat region and then resolving the amplified D N A on polyacrylamide D N A sequencing gels [22, 23]. Using the (AC)n marker, ten T N F - ~ alleles could be demonstrated [24]. Perhaps even more importantly these data imply that the N Z W mice may have a unique T N F - a allele. This supports our notion that the N Z W mouse may make a unique genetic contribution to the F1 disease by contributing a T N F - a gene capable of only low level of T N F - ~ production. These data are in agreement and confirm similar studies by Jongeneel et al. [25]. S e a r c h i n g f o r m e c h a n i s m s o f T N F effect in a u t o i m m u n e d i s e a s e
In an attempt to determine the mechanism(s) by which T N F - ~ suppresses autoimmunity, activated cell sorting analysis was conducted on T N F - a treated and PB S treated age-matched (NZB × N Z W ) F 1 mice at 2-3-week intervals during and after treatment. T N F - a treatment appeared to have no discernable effects on the frequency or phenotype of B, T or monocyte cell population in cells from the peritoneal cavity, spleen, lymph node and thymus. T h e r e was, however, a slight increase in the n u m b e r ofmacrophages and a decrease in L y l + B cells in the peritoneum, especially during the first few weeks of treatment (Table 1). In contrast to other immune intervention methods applied to the (NZB × N Z W ) F 1 model system (anti-L3T4 antibody; anti-Ia antibody; T L I ; cyclosporin treatment), T N F - a administration seems to have the least overall effect. In previous studies, increased Ly-1 B cell frequencies in N Z B mice [26] were interpreted as reflecting polyclonal expansion and activation of an autoantibody-producing Ly-1 B cell population. Recently, however, it was shown that these increases in Ly-1 B cell frequencies actually represent expansion of individual L y - 1 B clones [27], which do not appear to be responsible for the increased autoantibody production observed in these strains. T h u s , it is not possible, at this point, to evaluate the significance of the reduction in the frequencies o f L y - 1 B cells in T N F - a treated mice. We have also found that treatment of lupus prone mice with T N F - a does not reduce autoantibody production even though it retards lupus nephritis. For
138
C . O . Jacob f~
" ~ " ~ ,.~
m~ i
0
o 0
+
~q
x +
~em oo=~
t~
+ h'3 t¢'5 t'~
.
~ ~.'~ ~
x ~
0
zzr
o
o
..~'~ ~ ~
0
~i;dN 0
m)r.
°
0 ~ l , e3 o
T u m o r necrosis factor in a u t o i m m u n i t y
139
example, T N F - a treatment did not change the incidence or the levels of IgG antisingle or double-stranded D N A antibody in comparison to age-matched PBStreated controls [7]. This finding may be of significance. First, these data support previous genetic studies with respect to inheritance of autoantibody production and nephritis development as not being the result of a single gene but rather, result from faulty regulation of independently segregating genes [28-31 ]. In this regard, genetic studies have shown that the development of kidney disease in (NZB x N Z W ) F 1 mice map within the M H C of the N Z W mice [32, 33] while the contribution of M H C genes to the inheritance of autoantibody production is probably only secondary [31, 34]. Furthermore, these data are consistent with findings that T N F - a does not suppress humoral immunity in normal or lupus mice [35]. These observations imply that T N F - a may affect autoimmune disease not through effects on humoral immunity, but via an as yet unidentified mechanism. Recently, it was suggested that long-term administration of T N F may lead to a suppression of cell-mediated immunity [35], possibly by a systemic macrophage activation, which might inhibit the development of the autoimmune response before the effector stage. Class-II MHC regulation by TNF revised
We have noted a significant decrease in M H C class I I I a expression on macrophages in the peritoneum of T N F - ~ treated (NZB x N Z W ) F 1 mice. Ia expression on peritoneal macrophages was reduced 5-20-fold in two representative experiments [7]. Although some variation in the level of Ia expression even within the same group of animals was noted, it should be emphasized that we have never seen higher expression of I a in T N F - ~ treated mice than in age and sex matched PBS controls [7]. T h e observation that class II M H C expression is not upregulated but actually downregulated by T N F - u treatment may at least resolve the apparent paradoxical in vivo effects of T N F - ~ and I F N - y in (NZB x N Z W ) F 1 mice. I F N - y has been established as the prototype lymphokine that induces enhanced expression of M H C class II antigens. In addition, inappropriate expression of M H C class II molecules by I F N - y has been suggested to play a role in the initiation of autoimmune processes [36]. In recent experiments we have shown that I F N - y can enhance the development of lupus nephritis in (NZB x N Z W ) F 1 mice [37]. T h u s , the apparent opposite effects of T N F - a and I F N - y in regulation of class II M H C molecules in this in vivo system may imply that regulation of Ia expression is the mechanism by which T N F - a or I F N - y affect the development of lupus nephritis in the (NZB x N Z W ) F 1 mice. T N F - a has been shown both to synergize with I F N - y and also to antagonize I F N - y regarding regulation of class II M H C expression [38-41] (Table 2). In a separate publication we propose that depending on the stage of differentiation and maturation of the target cell, T N F - a might synergize or antagonize I F N - y induced regulation of M H C class II antigens [42]. T h u s , in immature cells such as H L - 6 0 or T H P - 1 , T N F - a enhances I F N - y induction of M H C class II antigens, while in differentiated cells such as skin fibroblasts or activated macrophages, T N F - a downregulates the I F N - y induced class II expression. In bone marrow cells induced to differentiate in vitro, T N F - a decreased the I F N - y induced M H C class II expression in a maturation-dependent fashion.
140
C.O.Jacob
¢q ~c,~
~
~ "~
~--~
.~
o ¢J
~g
(o
,3
-=g -== ~,
I
O0
b U~g.~
o
o
oE=
E o m0o ~
on~
8 o
"z • D~NN
N ~
T u m o r necrosis factor in a u t o i m m u n i t y
141
It is reasonable to assume that antigen presentation capability is a property of mature, differentiated and activated cells rather than primitive or immature cells. Therefore the downregulation of M H C class I I expression by T N F - ~ shown on mature macrophages might be m u c h more relevant biologically than the upregulatory effect seen on undifferentiated cells.
Acknowledgements I wish to thank m y collaborators who have contributed to the data cited in this article, notably Zdenka Fronek, Sadakazu Also, John Hansen, Gail Lewis, May Koo, Frances Hwang, Yoshifumi Watanabe, Hans Acha-Orbea, Alan Stall, Sara Michie and Bob Schreiber. A significant portion of these studies have been performed in D r H u g h M c D e v i t t ' s laboratory. I wish to express m y gratitude to D r H u g h McDevitt for advice, help and interest throughout these studies.
References 1. Jacob, C. O., Z. Fronek, G. D. Lewis, M. Koo, J. A. Hansen, and H. O. McDevitt. 1990. Heritable major histocompatibility complex class II-associated differences in production of tumor necrosis factor-ct: relevance to genetic predisposition to systemic lupus erythematosus. Proc. Natl. Acad. Sci. USA 87:1233-1237 2. Muller, U., C. V. Jongeneel, S. A. Nedospasov, K. F. Lindahl, and M. Steinmetz. 1987. Tumor necrosis factor and lymphotoxin genes map close to H-2D in mouse major histocompatibility complex. Nature 325:265-268 3. Dunham, I., C. A. Sargent, J. Trowsdale, and D. R. Campbell. 1987. Molecular mapping of the human major histocompatibility complex by pulse field gel electrophoresis. Proc. Natl. Acad. Sci. USA 84:7237-7241 4. Ragoussis, J., K. Bloemer, E. H. Weiss, and A. Ziegler. 1988. Localization of the genes for tumor necrosis factor and lymphotoxin between the HLA class I and class I I I regions by field inversion gel electrophoresis. Immunogenetics 27:66 5. Jacob, C. O. and H. O. McDevitt. 1988. Tumor necrosis factor-c~in murine autoimmune 'lupus' nephritis. Nature 331:356-358 6. Fronek, Z., L. A. Timmerman, C. A. Alper, B. H. Hahn, K. Kalunian, B. M. Peterlin, and H. O. McDevitt. 1990. Major histocompatibility complex genes and susceptibility to systemic lupus erythematosus. Arthritis Rheum. 33:1542 7. Jacob, C. 0., F. Hwang, G. D. Lewis, and A. M. Stall. 1991. Tumor necrosis factor-c~ in murine systemic lupus erythematosus disease models. Implications for genetic predisposition and immune regulation. Cytokine 3(6): 1-11 8. Murphy, E. D. and J. B. Roths. 1979. A Y-chromosome associated factor in strain BXSB producing accelerated autoimmunity and lymphoproliferation. Arthritis Rheum. 22: 1188-1194 9. Gordon, C., G. E. Ranges, J. S. Greenspan, and D. Wofsy. 1989. Chronic therapy with recombinant tumor necrosis factor-a in autoimmune NZB/NZW F1 mice. Clin. Immunol. Immunopathol. 52:421 10. Boswell, J. M., M. A. Yui, D. W. Burt, and V. E. Kelley. 1988. Increased tumor necrosis factor and IL-113gene expression in the kidneys of mice with lupus nephritis. J. Immunol. 141: 3050-3054 11. Brennan, D. C., M. A. Yui, R. P. Wuthrich, and V. E. Kelley. 1989. Tumor necrosis factor and IL-1 in New Zealand black/white mice. J. Immunol. 143" 3470 12. Beutler, B., N. Krochin, I. W. Milsark, C. Luedke, and A. Cerami. 1986. Control of cachectin (tumor necrosis factor) synthesis: mechanisms of endotoxin resistance. Science 232:977
142
C.O.Jacob
13. Bendtzen, K., T. Mandrup-Poulsen, J. Nerup, J. H. Nielsen, C. A. Dinarello, and M. Svenson. 1986. Cytotoxicity of human pI7 interleukin-1 for pancreatic islets of Langerhans. Science 232:1545-1547 14. Mandrup-Poulsen, T., K. Bendtzen, C. A. Dinarello, and J. Nerup. 1987. Human T N F potentiates human IL-1 mediated pancreatic [3-cell cytotoxicity. J. Immunol. 139: 4077-4082 15. Bendtzen, K. 1989. Immune hormones (cytokines): Pathogenic role in autoimmune rheumatic and endocrine diseases. Autoimmunity 2:177-189 16. Bendtzen, K., N. Morling, A. Formsgaard, M. Svenson, B. K. Jacobsen, N. Odum, and A. Svejgaard. 1988. Association between H L A - D R 2 and production of tumor necrosis factor alfa and interlukin 1 by mononuclear cells activated by lipopolysacharide. Stand. J. Immunol. 28:599-606 17. Jacob, C. O., S. Also. S. A. Michie, H. O. McDevitt, and H. Acha-Orbea. 1990. Prevention of diabetes in nonobese diabetic mice by tumor necrosis factor (TNF): Similarities between TNF-ct and IL-1. Proc. Natl. Acad. Sci. USA 87:968-972 18. Satoh, J., H. Seino, T. Abo, S. Tanaka, S. Shintani, S. Ohta, K. Tamura, T. Sawai, T. Nobunaga, T. Otchi, K. Kumagai, and T. Toyota. 1989. Recombinant human T N F - a suppresses autoimmune diabetes in N O D mice. J. Clin. Invest. 84:1345-1348 19. Satoh, J., H. Seino, S. Shintani, and S. Tanaka. 1990. Inhibition of type 1 diabetes in BB rats with recombinant human tumor necrosis factor-alpha. J. Immunol. 145: 1395-1399 20. Santamaria, P., R. G., Gehrz, M. K. Bryan, and J. J. Barbosa. 1989. Involvement of class I I M H C molecules in the L P S-induction of I L - 1 / T N F secretions by human monocytes. J. Immunol. 143:913-922 21. Jacob, C. O., S. Also, R. D. Schreiber, and H. O. McDevitt. Submitted 22. Weber, J. L. and P. E. May. 1989. Abundant class of human D N A polymorphisms which can be typed using the polymerase chain reaction. Am. J. Hum. Genet. 44:388-396 23. Litt, M. and J. A. Luty. 1989. A hypervariable microsatellite revealed by in vitro amplication of a dinucleotide repeat within the cardiac muscle actin gene. Am. J. Hum. Genet. 44:397-401 24. Jacob, C.O. a n d F . H w a n g . 1992. Definition ofmicrosatellite size variants for T N F - a and HSP70 in autoimmune and nonautoimmune mouse strains. Immunogenetics in press 25. Jongeneel, C. V., H. Acha-Orbea, and T. Blankenstein. A polymorphic microsatellite in the tumor necrosis factor a promoter identifies an allele unique to the N Z W mouse strain. J. Exp. Med. 171:2141-2146 26. Hayakawa, K., R. R. Hardy, M. Honda, L. A. Herzenberg, A. D. Steinberg, and L. A. Herzenberg. 1984. Ly-1 B cells: Functionally distinct lymphocytes that secrete I g M autoantibodies. Proc. Natl. Acad. Sci. USA 81:2494-2498 27. Stall, A. M., M. C. Farinas, D. M. Tarlinton, P. A. Lalor, L. A. Herzenberg, S. Strober, and L. A. Herzenberg. 1988. Ly-1 B cell clones similar to human chronic lymphocytic leukemias routinely develop in older mice and young autoimmune (New Zealand Black-related) animals. Proc. Natl. Acad. Sci USA 85:7312-7316 28. Theofilopoulos, A . N . a n d F . J. Dixon. 1985. M u r i n e m o d e l s o f s y s t e m i c l u p u s erythematosus. Adv. Immunol, 37:269 29. Knight, J. G. and D. D. Adams. 1978. Three genes for lupus nephritis in NZB × N Z W mice. J. Exp. Med. 147:1635 30. Raveche, E. S., A. D. Steinberg, K. W. Klassen, and J. H. Tijo. 1978. Genetic studies in N Z W mice. J. Exp. Med. 147:1487 31. Hirose, S., R. Nagasawa, and I. Sekikawa. 1983. Enhancing effect of H-2 linked N Z W genes on the autoimmune traits of (NZB × N Z W ) F 1 mice. J. Exp. Med. 158:228 32. Kotzin, B. L. and E. D. Palmer. 1987. T h e contribution of N Z W genes to lupus-like disease in (NZB × N Z W ) F 1 mice. J. Exp. Med. 165:1237 33. Babcock, S. K., V. B. Appel, M. Schiff, E. Palmer, and B. L. Kotzin. 1989. Genetic analysis of the imperfect association of H-2 haplotype with lupus-like autoimmune disease. Proc. Natl. Acad. Sci. USA 86:7552-7555
T u m o r necrosis factor in a u t o i m m u n i t y
143
34. Bell, D. A. and P. J. Madison. 1980. Serologic subsets in systemic lupus erythematosus. An example of autoantibodies in relationship to clinical features of disease and H L A antigens. Arthritis Rheum. 23:1268 35. Gordon, C. and D. Wofsy. 1990. Effects of recombinant murine tumor necrosis factor-c~ on immune function.ft. Immunol. 144:1753 36. Bottazzo, G. F., R. Pujol-Borrell, and T. Hanafusa. 1983. Role of aberrant H L A - D R expression and antigen presentation in induction of endocrine autoimmunity. Lancet 2: 1115-1119 37. Jacob, C. O., P. H. Van Der Meide, and H. O. McDevitt. 1987. In vivo treatment of (NZB x NZW)F1 lupus-like nephritis with monoclonal antibody to y-interferon, ft. Exp. Med. 166:798-802 38. Chang, R. J. and S. H. Lee. 1986. Effects of interferon-y and tumor necrosis factor-a on the expression of an Ia antigen on a murine macrophage cell line. ft. Immunol. 137: 2853-2856 39. Hoffman, M. and J. B. Weinberg. 1987. T u m o r necrosis factor-c~ induces increased hydrogen peroxide production and Fc receptor expression, but not increased Ia antigen expression by peritoneal macrophages, ft. Leukocyte Biol. 42:704-707 40. Pfizenmaier, K., P. Scheurich, C. Schluter, and M. Kronke. 1987. T u m o r necrosis factor enhances HLA-A, B, C and H L A D R gene expression in human tumor cells, ft. Immunol. 138:975-980 41. Leeuwenberg, J. F. M., J. van Damme, T. Meager, T. M. A. A. Jeunhomme, and W. W. Buurman. 1988. Effects of tumor necrosis factor on the interferon-y-inducedmajor histocompatibility complex class II antigen expression by human endothelial cells. Eur. ft. Immunol. 18:1469 42. Watanabe, Y. and C. O. Jacob. 1991. Regulation of M H C class II antigen expression: Opposing effects of TNF-c~ on I F N - ? induced H L A - D R and Ia expression depends on the maturation and differentiation stage of the cell. ft. Immunol. 146:899-905 43. Jacob, C. O., G. D. Lewis, and H. O. McDevitt. 1991. M H C class II associated variation in the production of tumor necrosis factor in mice and humans: relevance to the pathogenesis of autoimmune diseases. Immunol. Res. 10:156-168 44. Cannella, B., and C. S. Raine. 1989. Cytokines up-regulate Ia expression in organotypic cultures of central nervous system tissue, ft. Neuroimmunol. 24:239 45. Arenzana-Seisdedos, F., S. C. Mogensen, F. Vuillier, W. Fiers, and-J.SL.-Vireiizier. 1988 Autocrine secretion of tumor necrosis factor under the influence of interferon-v amplifies H L A - D R gene induction in human monocytes. Proc. Natl. Acad. Sci. U S A . 85:6087 46. Wedgwood, J., L. Hatam, and V. Bonagura. 1988. Effect of interferon-gamma and tumor necrosis factor on the expression of class I and class II major histocompatibility complex molecules by cultured human umbilical vein endothelial cells. Cell Immunol. 111:1-9
S t u d i e s o n t h e R o l e o f T u m o r N e c r o s i s F a c t o r in Murine and Human Autoimmunity
Chaim
O. Jacob
Department of Inflammation Biology and Immunology, Syntex Research, Palo Hlto, C A 94303, U S A
We have analyzed the roles of t u m o r necrosis factor alpha (TNF-a) in h u m a n systemic lupus e r y t h e m a t o s u s (SLE) and routine models of lupus as well as in type I diabetes in NOD mice. These studies suggest an i m p o r t ant role for TNF-a in the pathogenesis o f a u t o i m m u n e disease. R a t h e r than being involved m a i n l y in the effector a r m of the i n f l a m m a t o r y process of a u t o i m m u n e o r g a n destruction, our data suggest a p r i m a r y involvement in some of the basic m e c h a n i s m s of the a u t o i m m u n e process. Evidence has been presented that emphasizes the possibility of the involvement of this cytokine in the genetic predisposition to SLE. The data m a y i m p l y that the effect of TNF on the i m m u n e system m a y be m o r e relevant to the pathogenesis of the a u t o i m m u n e disease than direct local effects at some target organs. Based on the data presented, one should be cautious in extrapolating the effects of this cytokine in various in vitro systems to the in vivo situation.
M H C class I I - a s s o c i a t e d v a r i a t i o n in t u m o r n e c r o s i s f a c t o r p r o d u c t i o n T h e production of t u m o r necrosis factor ( T N F ) by peritoneal macrophages activated with lipopolysaccharide (LPS) and interferon gamma ( I F N - y ) , obtained from various mouse strains revealed a significant variation in the production of cytokine among the different strains. This variability seemed not to be random but rather M H C associated. For example, Iad mice (BALB/c, N Z B and DBA/2] produce similar intermediate levels of T N F - a while Ia q mice (SWR, DBA/1 and B10.Q) produce very high levels of TNF-ct. : ~: T o extend these studies to humans, the production of T N E by activated peripheral blood lymphocytes or enriched monocyte populations of H L A - A , -B, -C, - D R and - D Q serotyped normal donors was assayed. As in mice, major differences in the level of T N F - a inducibility were observed between different individuals. T N F - a 133 0896-8411/92/0A0133+ 11 $03.00/0
© 1992AcademicPress Limited
134 C.O.Jacob production was found associated with the M H C class II genotype. DR2 and D Q w l positive individuals were found to produce significantly lower levels of T N F than donors not possessing these specificities,. On the other hand normal donors positive for DR3 or DR4 were almost always high producers of the cytokine [1].
TNF in systemic lupus erythematosus (SLE) Because the T N F genes are located within the M H C [2-4], the possibility of the involvement of a 'disregulated' T N F gene in autoimmune disease development is an attractive hypothesis [2]. We have recently presented evidence suggesting that a ( T N F - ~ ) allele contributed by the N Z W mouse strain may be involved in the pathogenesis of lupus nephritis in the (NZB x N Z W ) F 1 mouse [5]. N Z W mice have no autoimmune disease, while the (NZB x N Z W ) F 1 develop severe glomerulonephritis. T h e N Z W parental strain was found to produce very low levels of T N F - m T h e r a p y with recombinant T N F - ~ induced a very signifcant diminution in the development of lupus nephritis in the (NZB x N Z W ) F 1 mouse. We suggest that the N Z W parent may contribute to the F1 disease a T N F - ¢ gene capable of only low levels of T N F production. T h u s , recombinant T N F - ~ may possibly act as replacement therapy suggesting that such treatment may be supplementing the low endogenous T N F production. Similar results were obtained in human SLE: DR2, D Q w l - p o s i t i v e S L E patients show low levels of T N F - ~ production [ 1] and this genotype is also associated with an increased incidence of lupus nephritis [ 1, 6]. However, a substantial portion of S L E patients are DR3 associated and these patients do not have low T N F - ~ production, similar to M R L - l p r / l p r and BXSB mice which also show moderate to high levels of T N F - a production [7]. Taken together these studies suggest that S L E is not a single condition but rather can be subdivided from the genetic point of view into two or more subsets (Figure 1). These data also emphasize that the MHC-associated genetic predisposition may actually proceed through different mechanisms involving different genes in different subsets of the disease. While T N F - ~ significantly delays the development of nephritis in (NZB x N Z W ) F 1 and to a lesser extent also in the M R L - l p r / l p r model, it has no protective effect in BXSB mice [7]. T h e BXSB model is believed to have a totally different genetic basis [8] and this could account for T N F - ~ having no effect on disease development in this strain of mice. An inverse correlation has been found between the level of T N F - a inducibility in peritoneal macrophages in vitro, and the effect of T N F - a administration in vivo. Based on these results, it is possible to predict that T N F - ~ treatment might be beneficial only in situations in which the T N F - ~ inducibility is low. It is also of interest that a 14-week treatment protocol was more effective in delaying the development of nephritis than continuation of T N F - ~ therapy for 27 weeks [7]. We have no explanation for these results, but have found that a n t i - T N F - ~ antibody had not developed in these mice. T h e protective effect of T N F - a in the (NZB x N Z W ) F 1 has been confirmed independently by G o r d o n et al. [9]. On the other hand in two separate publications Kelley and associates argue that rather than a beneficial effect, T N F - ~ may have a deleterious effect in murine lupus. First, Boswell et al. [10] have found increased
T u m o r necrosis factor in a u t o i m m u n i t y
135
Genetic Subsets of SLE
DR 2 (non Caucasian) DQwl (Caucasian) Low TNF Nephritis C4A null (normal genes) High anti-DNA Ab
DR 3 Ro,La Ab High TNF SICCA Syndrome C4A null (gene deletion)
- ve assoc. Nephritis - ve assoc, anti-DNA High TNF
Figure 1. Possible genetic subsets of systemiclupus erythematosus(from Jacob et al. [43]). m R N A expression of interleukin-1 (IL-1) and T N F - a locally in the kidneys of M R L - l p r / l p r mice, suggesting that intra-renal production of these cytokines is responsible for a cascade of events leading to lupus nephritis. In the second paper Brennan et al. [ 11 ] report enhanced T N F - a and IL-1 ]3 expression in the kidneys of (NZB x N Z W ) F 1 mice. Moreover, very low doses of T N F - a (0.2 ~tg/injection) given from 4 to 8 months of age accelerated renal disease and mortality rate. While our basic concept of the potential role of T N F in autoimmune lupus nephritis is clearly different from the above, the data from these studies are not necessarily contradictory. First, we have tested T N F - a protein production by stimulated peritoneal macrophages, while these studies assayed T N F - u m R N A levels locally in the kidneys. In this regard it should be emphasized that despite the fact that the clinical syndrome may be dominated by the antigen-antibody complex induced nephritis, it is very unlikely that the kidney is the initial target organ where the recognition, binding and presentation of the 'autoantigen' is taking place. Moreover, a significant discrepancy between the amount of T N F - ~ m R N A and the amount of T N F - a protein translated has been shown. Thus, in certain instances, T N F - ~ m R N A may be expressed in considerable abundance, while the protein is not expressed or is expressed at a very low level [12]. Second, different doses and ages of mice are used in these studies. We have used doses approximately 10 times higher than Brennan et al. [ 11]. Brennan et al. show an accelerating effect only with very low doses of T N F - ~ and only if treatment is started after 4 months of age and continued until 8 months of age. It is perhaps possible that this deleterious effect seen in 4 to 8 months treatment is related to our observation
136 C.O.Jacob that 3 to 9.5 months treatment is less effective than 3 to 6.5 months treatment [7]. O u r studies show that the earlier T N F - u is given the m o r e effective is its protective effect [7]. It is reasonable to assume that T N F - a is involved in a protective m o d e only in the early induction phases of the a u t o i m m u n e process, rather than in the late effector arm of the disease process. In any case, further investigation is required to understand the basis of differential dose response to TNF-ct. T N F in t y p e 1 d i a b e t e s m e l l i t u s It has been shown that I L - 1 is toxic to pancreatic ]3-cells in vitro [13] and that T N F significantly enhances this toxicity [14]. T h u s , release of I L - 1 and T N F by cells infiltrating the pancreatic islets have been suggested as the effector mechanism of tissue distraction in a u t o i m m u n e diabetes [ 15]. In accordance with this hypothesis, Bendtzen et al. [16] suggested that low T N F production by DR2-positive individuals may explain the DR2-associated resistance to insulin-dependent diabetes mellitus. Surprisingly, and in apparent contradiction with a role for T N F at the effector stage of the disease, we [17] and Satoh et al. [18] independently have shown that systemic administration of T N F - a has a protective effect in the non-obese diabetic ( N O D ) mice. We assessed the in vivo effects of T N F - a in the N O D mice on three levels: insulitis development, development of overt diabetes, and adoptive transfer of diabetes by splenic lymphocytes from diabetic female mice to irradiated young non-diabetic N O D males. First, treatment with T N F - ~ for 4 weeks was sufficient to significantly reduce the incidence of insulitis [ 17]. Second, administration of T N F - a on a long-term basis caused a major reduction in the observed incidence of diabetes [17]. T N F - c t was given at a dose of 3 ~tg/injection ip three times/week for a period of 3 months, starting at the age 8-9 weeks. Mice were followed up to approximately 12 months of age. Even 7 months after stopping the treatment only 26% of mice developed diabetes while about 70% of age matched controls had diabetes during this period of time (P < 0.05). Similar results have been obtained by Satoh et al. [ 18] treating N O D mice with TNF-~t (3 x 103 U) ip twice a week between 4 and 25 weeks of age. Third, T N F - a was effective in preventing the adoptive transfer of diabetes by lymphocytes from diabetic mice to non-diabetic mice [17]. Furthermore, more recently T N F has also been shown to be protective in the other model of autoimmune diabetes, namely the BB rat [19]. In addition in humans no correlation between T N F production and M H C class I I genotype has been found among type 1 diabetic patients [20]. Since we and Satoh et al. have studied in vivo effects while the other groups tested in vitro effects, it is possible that locally T N F and IL-1 may be toxic to pancreatic ]3-cells whereas systemically they may have global immunoregulatory effects, and therefore the two mechanisms may not be mutually exclusive. Since T N F - a has a very short half life w h e n administered in vivo and T N F receptors are believed to be widely distributed on most cell types, it is reasonable to assume that most of the effects we have observed by injecting T N F - ~ intraperitoneally would not be local at the level of the pancreatic ]3-cell. Monoclonal antibodies (mAb) on the other hand, have long half lives ( m a n y days in most cases), therefore we can assume that antiT N F antibodies will block endogenous T N F both systemically and also locally. Using a recently produced anti-murine T N F m A b provided to us by D r Bob
T u m o r necrosis factor in autoimmunity
137
Schreiber, we found that administration of this antibody to N O D females for a period of 8 weeks, increased the incidence of insulitis significantly compared with age-matched female N O D mice receiving non-relevant antibody or PBS [21 ]. Since mAb to T N F might be expected to block T N F locally as well as systemically, these results argue in favor of T N F having a net protective effect in disease pathogenesis. TNF gene polymorphism
Polymorphic markers are essential tools for studying the potential disease linkage to the T N F genes. At the R F L P level a two-allele polymorphism has been identified [5]. However, analysis of this polymorphism in multiple strains of mice failed to show correlation with T N F - a production and linkage between murine lupus and the T N F - a gene. On the other hand, using the P C R methodology to clone and sequence 5' regulatory regions of the T N F - a gene in various mice strains, an AC dinucleotide repeat was identified, which is polymorphic in length among different strains of mice and thus represents a new potential genetic marker. Variation in the length of this so-called microsatellite can be directly demonstrated by amplifying the D N A immediately flanking the repeat region and then resolving the amplified D N A on polyacrylamide D N A sequencing gels [22, 23]. Using the (AC)n marker, ten T N F - ~ alleles could be demonstrated [24]. Perhaps even more importantly these data imply that the N Z W mice may have a unique T N F - a allele. This supports our notion that the N Z W mouse may make a unique genetic contribution to the F1 disease by contributing a T N F - a gene capable of only low level of T N F - ~ production. These data are in agreement and confirm similar studies by Jongeneel et al. [25]. S e a r c h i n g f o r m e c h a n i s m s o f T N F effect in a u t o i m m u n e d i s e a s e
In an attempt to determine the mechanism(s) by which T N F - ~ suppresses autoimmunity, activated cell sorting analysis was conducted on T N F - a treated and PB S treated age-matched (NZB × N Z W ) F 1 mice at 2-3-week intervals during and after treatment. T N F - a treatment appeared to have no discernable effects on the frequency or phenotype of B, T or monocyte cell population in cells from the peritoneal cavity, spleen, lymph node and thymus. T h e r e was, however, a slight increase in the n u m b e r ofmacrophages and a decrease in L y l + B cells in the peritoneum, especially during the first few weeks of treatment (Table 1). In contrast to other immune intervention methods applied to the (NZB × N Z W ) F 1 model system (anti-L3T4 antibody; anti-Ia antibody; T L I ; cyclosporin treatment), T N F - a administration seems to have the least overall effect. In previous studies, increased Ly-1 B cell frequencies in N Z B mice [26] were interpreted as reflecting polyclonal expansion and activation of an autoantibody-producing Ly-1 B cell population. Recently, however, it was shown that these increases in Ly-1 B cell frequencies actually represent expansion of individual L y - 1 B clones [27], which do not appear to be responsible for the increased autoantibody production observed in these strains. T h u s , it is not possible, at this point, to evaluate the significance of the reduction in the frequencies o f L y - 1 B cells in T N F - a treated mice. We have also found that treatment of lupus prone mice with T N F - a does not reduce autoantibody production even though it retards lupus nephritis. For
138
C . O . Jacob f~
" ~ " ~ ,.~
m~ i
0
o 0
+
~q
x +
~em oo=~
t~
+ h'3 t¢'5 t'~
.
~ ~.'~ ~
x ~
0
zzr
o
o
..~'~ ~ ~
0
~i;dN 0
m)r.
°
0 ~ l , e3 o
T u m o r necrosis factor in a u t o i m m u n i t y
139
example, T N F - a treatment did not change the incidence or the levels of IgG antisingle or double-stranded D N A antibody in comparison to age-matched PBStreated controls [7]. This finding may be of significance. First, these data support previous genetic studies with respect to inheritance of autoantibody production and nephritis development as not being the result of a single gene but rather, result from faulty regulation of independently segregating genes [28-31 ]. In this regard, genetic studies have shown that the development of kidney disease in (NZB x N Z W ) F 1 mice map within the M H C of the N Z W mice [32, 33] while the contribution of M H C genes to the inheritance of autoantibody production is probably only secondary [31, 34]. Furthermore, these data are consistent with findings that T N F - a does not suppress humoral immunity in normal or lupus mice [35]. These observations imply that T N F - a may affect autoimmune disease not through effects on humoral immunity, but via an as yet unidentified mechanism. Recently, it was suggested that long-term administration of T N F may lead to a suppression of cell-mediated immunity [35], possibly by a systemic macrophage activation, which might inhibit the development of the autoimmune response before the effector stage. Class-II MHC regulation by TNF revised
We have noted a significant decrease in M H C class I I I a expression on macrophages in the peritoneum of T N F - ~ treated (NZB x N Z W ) F 1 mice. Ia expression on peritoneal macrophages was reduced 5-20-fold in two representative experiments [7]. Although some variation in the level of Ia expression even within the same group of animals was noted, it should be emphasized that we have never seen higher expression of I a in T N F - ~ treated mice than in age and sex matched PBS controls [7]. T h e observation that class II M H C expression is not upregulated but actually downregulated by T N F - u treatment may at least resolve the apparent paradoxical in vivo effects of T N F - ~ and I F N - y in (NZB x N Z W ) F 1 mice. I F N - y has been established as the prototype lymphokine that induces enhanced expression of M H C class II antigens. In addition, inappropriate expression of M H C class II molecules by I F N - y has been suggested to play a role in the initiation of autoimmune processes [36]. In recent experiments we have shown that I F N - y can enhance the development of lupus nephritis in (NZB x N Z W ) F 1 mice [37]. T h u s , the apparent opposite effects of T N F - a and I F N - y in regulation of class II M H C molecules in this in vivo system may imply that regulation of Ia expression is the mechanism by which T N F - a or I F N - y affect the development of lupus nephritis in the (NZB x N Z W ) F 1 mice. T N F - a has been shown both to synergize with I F N - y and also to antagonize I F N - y regarding regulation of class II M H C expression [38-41] (Table 2). In a separate publication we propose that depending on the stage of differentiation and maturation of the target cell, T N F - a might synergize or antagonize I F N - y induced regulation of M H C class II antigens [42]. T h u s , in immature cells such as H L - 6 0 or T H P - 1 , T N F - a enhances I F N - y induction of M H C class II antigens, while in differentiated cells such as skin fibroblasts or activated macrophages, T N F - a downregulates the I F N - y induced class II expression. In bone marrow cells induced to differentiate in vitro, T N F - a decreased the I F N - y induced M H C class II expression in a maturation-dependent fashion.
140
C.O.Jacob
¢q ~c,~
~
~ "~
~--~
.~
o ¢J
~g
(o
,3
-=g -== ~,
I
O0
b U~g.~
o
o
oE=
E o m0o ~
on~
8 o
"z • D~NN
N ~
T u m o r necrosis factor in a u t o i m m u n i t y
141
It is reasonable to assume that antigen presentation capability is a property of mature, differentiated and activated cells rather than primitive or immature cells. Therefore the downregulation of M H C class I I expression by T N F - ~ shown on mature macrophages might be m u c h more relevant biologically than the upregulatory effect seen on undifferentiated cells.
Acknowledgements I wish to thank m y collaborators who have contributed to the data cited in this article, notably Zdenka Fronek, Sadakazu Also, John Hansen, Gail Lewis, May Koo, Frances Hwang, Yoshifumi Watanabe, Hans Acha-Orbea, Alan Stall, Sara Michie and Bob Schreiber. A significant portion of these studies have been performed in D r H u g h M c D e v i t t ' s laboratory. I wish to express m y gratitude to D r H u g h McDevitt for advice, help and interest throughout these studies.
References 1. Jacob, C. O., Z. Fronek, G. D. Lewis, M. Koo, J. A. Hansen, and H. O. McDevitt. 1990. Heritable major histocompatibility complex class II-associated differences in production of tumor necrosis factor-ct: relevance to genetic predisposition to systemic lupus erythematosus. Proc. Natl. Acad. Sci. USA 87:1233-1237 2. Muller, U., C. V. Jongeneel, S. A. Nedospasov, K. F. Lindahl, and M. Steinmetz. 1987. Tumor necrosis factor and lymphotoxin genes map close to H-2D in mouse major histocompatibility complex. Nature 325:265-268 3. Dunham, I., C. A. Sargent, J. Trowsdale, and D. R. Campbell. 1987. Molecular mapping of the human major histocompatibility complex by pulse field gel electrophoresis. Proc. Natl. Acad. Sci. USA 84:7237-7241 4. Ragoussis, J., K. Bloemer, E. H. Weiss, and A. Ziegler. 1988. Localization of the genes for tumor necrosis factor and lymphotoxin between the HLA class I and class I I I regions by field inversion gel electrophoresis. Immunogenetics 27:66 5. Jacob, C. O. and H. O. McDevitt. 1988. Tumor necrosis factor-c~in murine autoimmune 'lupus' nephritis. Nature 331:356-358 6. Fronek, Z., L. A. Timmerman, C. A. Alper, B. H. Hahn, K. Kalunian, B. M. Peterlin, and H. O. McDevitt. 1990. Major histocompatibility complex genes and susceptibility to systemic lupus erythematosus. Arthritis Rheum. 33:1542 7. Jacob, C. 0., F. Hwang, G. D. Lewis, and A. M. Stall. 1991. Tumor necrosis factor-c~ in murine systemic lupus erythematosus disease models. Implications for genetic predisposition and immune regulation. Cytokine 3(6): 1-11 8. Murphy, E. D. and J. B. Roths. 1979. A Y-chromosome associated factor in strain BXSB producing accelerated autoimmunity and lymphoproliferation. Arthritis Rheum. 22: 1188-1194 9. Gordon, C., G. E. Ranges, J. S. Greenspan, and D. Wofsy. 1989. Chronic therapy with recombinant tumor necrosis factor-a in autoimmune NZB/NZW F1 mice. Clin. Immunol. Immunopathol. 52:421 10. Boswell, J. M., M. A. Yui, D. W. Burt, and V. E. Kelley. 1988. Increased tumor necrosis factor and IL-113gene expression in the kidneys of mice with lupus nephritis. J. Immunol. 141: 3050-3054 11. Brennan, D. C., M. A. Yui, R. P. Wuthrich, and V. E. Kelley. 1989. Tumor necrosis factor and IL-1 in New Zealand black/white mice. J. Immunol. 143" 3470 12. Beutler, B., N. Krochin, I. W. Milsark, C. Luedke, and A. Cerami. 1986. Control of cachectin (tumor necrosis factor) synthesis: mechanisms of endotoxin resistance. Science 232:977
142
C.O.Jacob
13. Bendtzen, K., T. Mandrup-Poulsen, J. Nerup, J. H. Nielsen, C. A. Dinarello, and M. Svenson. 1986. Cytotoxicity of human pI7 interleukin-1 for pancreatic islets of Langerhans. Science 232:1545-1547 14. Mandrup-Poulsen, T., K. Bendtzen, C. A. Dinarello, and J. Nerup. 1987. Human T N F potentiates human IL-1 mediated pancreatic [3-cell cytotoxicity. J. Immunol. 139: 4077-4082 15. Bendtzen, K. 1989. Immune hormones (cytokines): Pathogenic role in autoimmune rheumatic and endocrine diseases. Autoimmunity 2:177-189 16. Bendtzen, K., N. Morling, A. Formsgaard, M. Svenson, B. K. Jacobsen, N. Odum, and A. Svejgaard. 1988. Association between H L A - D R 2 and production of tumor necrosis factor alfa and interlukin 1 by mononuclear cells activated by lipopolysacharide. Stand. J. Immunol. 28:599-606 17. Jacob, C. O., S. Also. S. A. Michie, H. O. McDevitt, and H. Acha-Orbea. 1990. Prevention of diabetes in nonobese diabetic mice by tumor necrosis factor (TNF): Similarities between TNF-ct and IL-1. Proc. Natl. Acad. Sci. USA 87:968-972 18. Satoh, J., H. Seino, T. Abo, S. Tanaka, S. Shintani, S. Ohta, K. Tamura, T. Sawai, T. Nobunaga, T. Otchi, K. Kumagai, and T. Toyota. 1989. Recombinant human T N F - a suppresses autoimmune diabetes in N O D mice. J. Clin. Invest. 84:1345-1348 19. Satoh, J., H. Seino, S. Shintani, and S. Tanaka. 1990. Inhibition of type 1 diabetes in BB rats with recombinant human tumor necrosis factor-alpha. J. Immunol. 145: 1395-1399 20. Santamaria, P., R. G., Gehrz, M. K. Bryan, and J. J. Barbosa. 1989. Involvement of class I I M H C molecules in the L P S-induction of I L - 1 / T N F secretions by human monocytes. J. Immunol. 143:913-922 21. Jacob, C. O., S. Also, R. D. Schreiber, and H. O. McDevitt. Submitted 22. Weber, J. L. and P. E. May. 1989. Abundant class of human D N A polymorphisms which can be typed using the polymerase chain reaction. Am. J. Hum. Genet. 44:388-396 23. Litt, M. and J. A. Luty. 1989. A hypervariable microsatellite revealed by in vitro amplication of a dinucleotide repeat within the cardiac muscle actin gene. Am. J. Hum. Genet. 44:397-401 24. Jacob, C.O. a n d F . H w a n g . 1992. Definition ofmicrosatellite size variants for T N F - a and HSP70 in autoimmune and nonautoimmune mouse strains. Immunogenetics in press 25. Jongeneel, C. V., H. Acha-Orbea, and T. Blankenstein. A polymorphic microsatellite in the tumor necrosis factor a promoter identifies an allele unique to the N Z W mouse strain. J. Exp. Med. 171:2141-2146 26. Hayakawa, K., R. R. Hardy, M. Honda, L. A. Herzenberg, A. D. Steinberg, and L. A. Herzenberg. 1984. Ly-1 B cells: Functionally distinct lymphocytes that secrete I g M autoantibodies. Proc. Natl. Acad. Sci. USA 81:2494-2498 27. Stall, A. M., M. C. Farinas, D. M. Tarlinton, P. A. Lalor, L. A. Herzenberg, S. Strober, and L. A. Herzenberg. 1988. Ly-1 B cell clones similar to human chronic lymphocytic leukemias routinely develop in older mice and young autoimmune (New Zealand Black-related) animals. Proc. Natl. Acad. Sci USA 85:7312-7316 28. Theofilopoulos, A . N . a n d F . J. Dixon. 1985. M u r i n e m o d e l s o f s y s t e m i c l u p u s erythematosus. Adv. Immunol, 37:269 29. Knight, J. G. and D. D. Adams. 1978. Three genes for lupus nephritis in NZB × N Z W mice. J. Exp. Med. 147:1635 30. Raveche, E. S., A. D. Steinberg, K. W. Klassen, and J. H. Tijo. 1978. Genetic studies in N Z W mice. J. Exp. Med. 147:1487 31. Hirose, S., R. Nagasawa, and I. Sekikawa. 1983. Enhancing effect of H-2 linked N Z W genes on the autoimmune traits of (NZB × N Z W ) F 1 mice. J. Exp. Med. 158:228 32. Kotzin, B. L. and E. D. Palmer. 1987. T h e contribution of N Z W genes to lupus-like disease in (NZB × N Z W ) F 1 mice. J. Exp. Med. 165:1237 33. Babcock, S. K., V. B. Appel, M. Schiff, E. Palmer, and B. L. Kotzin. 1989. Genetic analysis of the imperfect association of H-2 haplotype with lupus-like autoimmune disease. Proc. Natl. Acad. Sci. USA 86:7552-7555
T u m o r necrosis factor in a u t o i m m u n i t y
143
34. Bell, D. A. and P. J. Madison. 1980. Serologic subsets in systemic lupus erythematosus. An example of autoantibodies in relationship to clinical features of disease and H L A antigens. Arthritis Rheum. 23:1268 35. Gordon, C. and D. Wofsy. 1990. Effects of recombinant murine tumor necrosis factor-c~ on immune function.ft. Immunol. 144:1753 36. Bottazzo, G. F., R. Pujol-Borrell, and T. Hanafusa. 1983. Role of aberrant H L A - D R expression and antigen presentation in induction of endocrine autoimmunity. Lancet 2: 1115-1119 37. Jacob, C. O., P. H. Van Der Meide, and H. O. McDevitt. 1987. In vivo treatment of (NZB x NZW)F1 lupus-like nephritis with monoclonal antibody to y-interferon, ft. Exp. Med. 166:798-802 38. Chang, R. J. and S. H. Lee. 1986. Effects of interferon-y and tumor necrosis factor-a on the expression of an Ia antigen on a murine macrophage cell line. ft. Immunol. 137: 2853-2856 39. Hoffman, M. and J. B. Weinberg. 1987. T u m o r necrosis factor-c~ induces increased hydrogen peroxide production and Fc receptor expression, but not increased Ia antigen expression by peritoneal macrophages, ft. Leukocyte Biol. 42:704-707 40. Pfizenmaier, K., P. Scheurich, C. Schluter, and M. Kronke. 1987. T u m o r necrosis factor enhances HLA-A, B, C and H L A D R gene expression in human tumor cells, ft. Immunol. 138:975-980 41. Leeuwenberg, J. F. M., J. van Damme, T. Meager, T. M. A. A. Jeunhomme, and W. W. Buurman. 1988. Effects of tumor necrosis factor on the interferon-y-inducedmajor histocompatibility complex class II antigen expression by human endothelial cells. Eur. ft. Immunol. 18:1469 42. Watanabe, Y. and C. O. Jacob. 1991. Regulation of M H C class II antigen expression: Opposing effects of TNF-c~ on I F N - ? induced H L A - D R and Ia expression depends on the maturation and differentiation stage of the cell. ft. Immunol. 146:899-905 43. Jacob, C. O., G. D. Lewis, and H. O. McDevitt. 1991. M H C class II associated variation in the production of tumor necrosis factor in mice and humans: relevance to the pathogenesis of autoimmune diseases. Immunol. Res. 10:156-168 44. Cannella, B., and C. S. Raine. 1989. Cytokines up-regulate Ia expression in organotypic cultures of central nervous system tissue, ft. Neuroimmunol. 24:239 45. Arenzana-Seisdedos, F., S. C. Mogensen, F. Vuillier, W. Fiers, and-J.SL.-Vireiizier. 1988 Autocrine secretion of tumor necrosis factor under the influence of interferon-v amplifies H L A - D R gene induction in human monocytes. Proc. Natl. Acad. Sci. U S A . 85:6087 46. Wedgwood, J., L. Hatam, and V. Bonagura. 1988. Effect of interferon-gamma and tumor necrosis factor on the expression of class I and class II major histocompatibility complex molecules by cultured human umbilical vein endothelial cells. Cell Immunol. 111:1-9
