The mechanism by which cord factor (CF), a toxic glycolipid from mycobacteria, induces cachexia was studied in BALB/c mice. Body weight was markedly ...
Vol. 56, No. 12
INFECTION AND IMMUNITY, Dec. 1988, p. 3067-3071
0019-9567/88/123067-05$02.00/0 Copyright © 1988, American Society for Microbiology
Tumor Necrosis Factor (Cachectin) Mediates Induction of Cachexia by Cord Factor from Mycobacteria CtLIO L. SILVA'* AND LUCIA H. FACCIOLI2 Department of Parasitology, Microbiology and Immunology' and Department of Pharmacology,2 School of Medicine of Ribeirdo Preto, University of Sdo Paulo, 14049 Ribeirdo Preto, Sdo Paulo, Brazil Received 20 May 1988/Accepted 23 August 1988
The mechanism by which cord factor (CF), a toxic glycolipid from mycobacteria, induces cachexia was studied in BALB/c mice. Body weight was markedly reduced 48 h after CF administration; the animals became severely wasted and exhibited hypertriglyceridemia, hypoglycemia, and high levels of tumor necrosis factor (TNF) in plasma. After CF administration, a transferable factor which caused cachexia and hypertriglyceridemia in recipient mice was detected in the blood. Dexamethasone partially inhibited the cachexia-inducing action of CF. Conditioned medium from adherent peritoneal cell cultures incubated with CF produced the same wasting symptoms when inoculated intravenously into mice. These studies also demonstrated that adherent peritoneal cells produced a humoral factor in response to CF which was related to CF-induced cachexia. Antiserum to recombinant TNF-a prevented the cachectin action in passive-transfer experiments. Our findings indicate that cachectin (TNF) plays a role as a central mediator of the wasting induced by CF.
A generalized infectious process leads to a broad and complex array of metabolic responses within the host, some of which have a direct causal relationship with the interactions of body cells with invading microorganisms or their toxic products or with specific host defensive mechanisms. Cord factor (6,6'-trehalose dimycolate; CF) is one of the immunostimulant and toxic glycolipid constituents of the cell walls of several strains of bacteria, including mycobacteria (15), nocardia (21), and corynebacteria (10, 22). The capacity of mycobacteria to enhance the nonspecific resistance of mice to tumors (1) and to bacterial infections (16) is related to the presence of CF. Bloch (6) demonstrated that CF from Mycobacterium tuberculosis has a peculiar characteristic toxicity for mice: a few repeated intraperitoneal (i.p.) injections of small amounts of glycolipid dissolved in paraffin oil killed a majority of animals following the rather precipitous weight loss characterizing cachexia, which was evident even after the first injection. Cachectin, or tumor necrosis factor (TNF), a cytokine released by mononuclear phagocytes in response to lipopolysaccharide (5) or other stimuli simulating host invasion (9), elicits a complex repertoire of metabolic reactions during inflammation (2). This molecule was originally identified as a cachexia-producing factor in experimental animals (13, 17). In adipose tissue cachectin causes complete suppression of the enzyme lipoprotein lipase, thereby preventing the uptake of exogenous triglyceride by fat cells and causing the lipemia frequently associated with infections (13). This effect reflects one of the physiological bases for cachexia. The range of stimuli that induce the production of cachectin is not known in detail. Infectious agents and their products have been identified as cachectin inducers (3). In this study, we examined the participation of cachectin in the induction of cachexia by CF in mice.
MATERIALS AND METHODS CF preparation. A 100-g portion of heat-killed, dried Mycobacterium bovis BCG (Moreau strain, provided by Oliveira Lima, Ataulfo de Paiva Foundation, Rio de Janeiro, *
Corresponding author.
Brazil) was repeatedly soaked in a mixture of chloroformmethanol (1:1, vol/vol), and the residual material extracted (20.5 g) was fractionated as previously reported for trehalose cdimycolate purification (5). The purified glycolipid (0.650 g) had an [t]25 of + 47 (C = 0.5 CHCl3), a melting point of 58°C, 11% sugar content as determined by the phenolsulfuric acid method (7), and an infrared spectrum similar to that described for CF isolated from M. tuberculosis (12). After alkaline hydrolysis (10), trehalose and mycolic acid were identified in the aqueous and ether phases, respectively. Physical and chemical analyses of the isolated mycolic acid (22) showed that the carbon chain length was centered in C84, in agreement with previous results (14). Acetylated CF (AC-CF) was prepared exactly as described previously (19). Animals. Male BALB/c mice, 4 to 6 weeks old and weighing 18.0 to 20.0 g, were used in all of the experiments. Induction of cachexia and preparation of plasma from animals. The material to be tested for toxicity was dissolved in mineral oil (MO) (Nujol; Plough Inc.). Before use, the clear solution was heated to 56°C and subjected to ultrasonic vibration for 5 min at 100 W, using a probe 9 mm in diameter. CF toxicity was evaluated by injecting mice i.p. with 10 ,ug of the glycolipid in 0.1 ml of MO and recording weight loss. Control groups were given corresponding amounts of AC-CF in MO or of MO alone. Mice injected with CF were exsanguinated 48 h after CF injection, and plasma was obtained. Control plasma was obtained from noninjected mice and from mice injected with AC-CF in MO or with MO alone in a similar manner. All plasma samples were stored at -20°C until used. Determination of plasma triglyceride and glucose. To assay for plasma triglyceride and glucose, we used triglyceride and glucose kits (Labtest Diagnostica), respectively. Plasma transfer. Undiluted plasma from treated and untreated mice was transferred to normal recipient mice in four 0.1-ml doses injected at 6-h intervals. The weight loss and plasma triglyceride levels of recipient mice were then recorded. Plasma to be used in passive-transfer experiments was preincubated with an excess of a polyclonal rabbit anti-recombinant murine TNF-a- antiserum (kindly provided 3067
3068
SILVA AND FACCIOLI
by Charles A. Dinarelo, Tufts University School of Medicine, Boston, Mass.) containing approximately 2 x 106 neutralizing units per milliliter. Glucocorticoid treatment of animals. To demonstrate the effect of glucocorticoid on hypertriglyceridemia and CFinduced cachexia, the animals were treated with one dose of dexamethasone (0.5 mg/kg of body weight) before CF injection and with four additional doses administered at 8-h intervals. Controls included animals that were not treated with dexamethasone and animals that were treated with dexamethasone but not injected with CF. Preparation of APC supernatant. Adherent peritoneal cell (APC) supernatant containing cachexia-inducing factor was prepared from peritoneal cells from mice that had been injected i.p. with sterile Brewer thioglycolate medium (3 ml; Difco Laboratories, Detroit, Mich.) 5 days before harvest. The cells (1.5 x 106 in 1.5 ml of RPMI 1640 medium) were incubated in 35-mm tissue culture dishes (Descarplast, Sao Paulo, Brazil) for 3 h at 37°C in a humidified atmosphere of 5% C02-95% air, after which nonadherent cells were removed by three washings with medium. Cells adhering to the plate were primarily macrophages. These cells were further incubated in serum-free RPMI 1640 medium (containing 100 U of penicillin, 100 pLg of streptomycin, and 2 ,ug of indomethacin per ml) in the presence or absence of a 10-,ug/ml aqueous suspension of CF (18) or AC-CF for 24 h at 37°C in a 5% CO2 environment. The culture medium was removed after incubation and centrifuged at 1,000 x g for 10 min at 4°C; 5 mg of charcoal power was added to the supernatant to remove residual CF, and the medium was again centrifuged at 10,000 x g for 10 min. The supernatants were filtered through a 0.2-,um-pore-size filter and frozen at -20°C. The cachexia-inducing activity of the APC supernatants was evaluated by inoculating mice intravenously at 6-h intervals with four 0.1-ml doses and by recording weight loss and plasma triglyceride levels. APC supernatants preincubated with rabbit anti-recombinant murine TNF-ox antiserum were used in passive-transfer experiments. Measurement of TNF levels. Killed L929 mouse tumor cells were used to measure TNF levels in plasma and in the macrophage supernatants on the basis of a standard assay (18). Briefly, L929 cells in RPMI 1640 medium containing 5% fetal calf serum were seeded at 3 x 104 cells per well in 96-well microdilution plates (Linbro; Flow Laboratories, McLean, Va.) and incubated overnight at 37°C in an atmosphere of 5% CO2 in air. Serial 1:2 dilutions of plasma or APC supernatants were made in the above-described medium containing 1.0 jig of actinomycin D (Sigma Chemical Co., St. Louis, Mo.) per ml, and 100-,u volumes of each dilution were added to the wells. On the next day, cell survival was assessed by fixing and staining the cells with crystal violet (0.2% in 20% methanol), and 0.1 ml of 1% sodium dodecyl sulfate was added to each well to solubilize the stained cells. The absorbance of each well was read at 490 nm with a model BT-100 Microelisa Autoreader (BioTek). Percentage of cytotoxicity was calculated as 1 - (A490 of sample/A490 of control) x 100. One TNF unit was defined as the amount of TNF giving 50% cell survival. For characterization of the cytotoxic activity in plasma or in the supernatants of CF-treated APC, the samples were incubated with an excess of rabbit anti-recombinant murine TNF-a antiserum or with control rabbit serum. After 2 h at 37°C, residual cytotoxicity was determined by adding the test samples to L929 cells.
INFECT. IMMUN.
Oh
I-
3: 0 0
24 48 HOURS POST INJECTION FIG. 1. Mean body weight (± standard deviation) of groups of six mice at 24 and 48 h after a single i.p. injection of 10 ,ug of CF in 0.1 ml of MO (A), 10 ,ug of AC-CF in 0.1 ml of MO (-), and 0.1 ml of MO alone (0) and in noninjected mice (*). Results are from a typical experiment. RESULTS
The CF glycolipid was emulsified on MO, and 10 jig of this suspension was injected i.p. into a group of six mice whose body weights were observed for 24 and 48 h (Fig. 1). At 48 h, the animals had lost 3.0 g of their initial body weight, were reluctant to move, appeared unwell, had ruffled fur, and were diarrheic. These animals were killed, and gross inspection showed the presence of peritonitis with an inflammatory exudate as well as ischemic and hemorrhagic lesions of the gastrointestinal tract. Control mice given AC-CF in MO (10 p.g) or MO alone showed no toxic effects and gained weight normally. The use of AC-CF in these experiments as a negative control was based on initial observations that this derivative substance was not toxic for mice. Triglyceride, glucose, and cachectin levels in plasma of mice with CF-induced cachexia were determined. The triglyceride concentration was 2.1-fold higher in mice injected with CF than in control animals (Fig. 2). The hypertriglyceridemic state was remarkable in view of the severe wasting diathesis that accompanied this experimental model. On the other hand, the glucose concentration in plasma was 2.5 times less in the CF-treated mice than in the controls (Fig. 2). Cytotoxic activity for L929 cells was observed in plasma collected from mice injected with CF but not in plasma collected from control animals. Lytic activity was completely prevented in the presence of a rabbit antibody to recombinant murine TNF-a (Fig. 3A). Recipient mice injected intravenously with plasma from animals with CF-induced cachexia became ill and unkempt in appearance, developed diarrhea, lost 2.1 g of their body weight 48 h after the first injection of plasma (Fig. 4A), and showed increased plasma triglyceride levels (Fig. 5A). Histopathological analysis of the lungs of these mice showed severe interstitial pneumonitis. Previous incubation of plasma with a neutralizing dose of rabbit anti-TNF-ot antiserum prevented weight loss of mice in passive-transfer experiments. The effects of dexamethasone on the weight loss and on the plasma triglyceride and cachectin levels of mice injected with CF were determined. Dexamethasone markedly sup-
VOL. 56, 1988
CACHEXIA INDUCED BY CORD FACTOR
fiT;
4
41 a U)
4
IL
0
.100
loo
- 200
X co
E
2 0
I
"I.z
0
4
0
r-° 0 C] I-
50
100
cn
0
0
MINERAL OIL
AC-CF
CF
FIG. 2. Plasma tryglyceride (open bars) and glucose (hatched bars) concentrations in rnice inoculated i.p. with 10 ,.g of CF in 0.1 ml of MO. Plasma samples were obtained 48 h after injection. Control groups consisted of noninjected mice, mice injected with AC-CF in MO, and mice injected with MO alone. Data representing a typical experiment are expressed as mnean + standard deviation for six mice in each group.
pressed the CF-induced increase in plasma triglyceride and cachectin content (Fig. 5B and 3B, respectively). However, the glucocorticoid partially protected mice against weight loss (Fig. 4B). UNITS OF CACHECTIN RELEASED/ml 2
0
4
6
10
8
A NONINJECTED
CF
C+rTWANTtSERUM MO
I
DEXAMETHASONE
DEXAMETHASONE+CF
e CF
I
CF+rTNFANTtSERLM MEDIUM
serum.
CF-stimulated APC supernatant
was
injected intra-
venously into mice; the average loss of body weight and plasma triglyceride of these animals were still markedly
different in relation to those observed in animals that received CF-stimulated APC supernatant preincubated with TNF-a antiserum, medium from cell cultures without added CF, or medium containing CF without cells (previously delipidated with charcoal particles) (Fig. 4C and SC).
,
-J
NONE
The APC were incubated in vitro in the presence of an aqueous CF suspension (10 ,ug/ml), and the CF-free supernatant was highly toxic for L929 tumor cells, showing the presence of cachectin (Fig. 3B). This figure also show that cytoxicity against L929 cells was completely neutralized with the TNF-a-specific antiserum but not with the control
Cc
Uzn O 0
10
3069
t
FIG. 3. (A) Cachectin content of plasma from mice treated with CF, from CF-treated mice previously incubated with rabbit antimurine TNF-a (rTNF) antiserum, MO, dexamethasone, or dexamethasone plus CF, and from untreated mlice; (B) cachectin content of supernatants from APC exposed to CF, from APC exposed to CF and previously incubated with rabbit anti TNF-a antiserum, and from APC exposed to medium alone for 24 h in culture. Data represent averages from experiments performed in triplicate.
DISCUSSION Several common physiological and biochemical derangements are seen in mammalian hosts responding to a variety of invasive stimuli such as bacterial, viral, and protozoan infections. These responses, which may be initiated by the direct effect of a microorganism or its toxic products on body cells, include a condition known as cachexia, in which animals continuously lose weight even while consuming an adequate diet. It has been widely assumed that CF is directly responsible for the wasting associated with mycobacterial infections (6). The results presented here, taken as whole, indicate that cachectin (TNF) plays a role as a mediator of the wasting induced by CF. Mice injected with CF became profoundly cachectic, losing up to 25% of their body mass within 48 h after inoculation, and, although severely wasted, became remarkably lipemic. Triglycerides were the principal components of the increased plasma lipids, and there was an increase in TNF levels in plasma. Thus, it was postulated that a host factor (cachectin) produced in response to the stimuli was responsible for this clearly defined biochemical alteration. Infectious processes also include alterations in carbohydrate metabolism, such as accelerated glycogenolysis, depressed glycogen synthesis, changes in hepatic gluconeogenesis, and markedly enhanced peripheral use of glucose for both oxidative and nonoxidative catabolism (8). Blood alterations therefore reflect a profound hypoglycemia as a result of peripheral utilization of glucose. Monokines have been implicated in both peripheral and hepatic alterations in glucose metabolism in septic situations or endotoxicoses (8). In response to a variety of invasive stimuli, macrophages and lymphocytes secrete cytokines that are capable of altering host metabolism. One of these cytokines, cachectin (TNF), has been reported to play an important role in inducing shock (2) and in the metabolic processes that lead to cachexia (17). The observation that CF produces cachexia raises a number of questions. Are the host responses to CF dependent on the production of cachectin? Might the plasma collected from animals injected with CF effectively induce CF-producing cachexia and tissue injury when administered to other animals? If so, what is the final mechanism of toxicity? The results obtained in this study show that many of the features of CF toxicity are reliably reproduced by infusion of plasma from CF-primed animals. Lipopolysaccharide appears to induce cachectin byosynthesis at both the transcriptional and posttranscriptional levels. Glucocorticoid hormones, which are highly effective in preventing shock and treating inflammation, inhibit this process at both levels, thus preventing production of cachectin (4). The results we obtained after treatment of mice with
SILVA AND FACCIOLI
3070
INFECT. IMMUN.
22
A
B
C
> 20
-
I
x
I
8 16
--w
T
WI
i16-
14'
10
24
48
24 48 0 24 48 POST INJECTION FIG. 4. Body weights of mice after the following treatments: (A) plasma from noninjected mice (*), plasma from mice injected with CF (A), AC-CF (A), and MO (0), and plasma from CF-treated mice preincubated with TNF-a antiserum (V). (B) Body weights of CF-infected mice treated with dexamethasone (v), mice injected with dexamethasone only (0), mice injected with CF without dexamethasone (A), and noninjected mice (*). (C) Body weights after inoculation of APC supernatants from cells incubated with CF (A), supernatants from CF-treated cells preincubated with TNF-a antiserum (V), culture medium only (0), medium plus CF and without cells (v), and supernatant from CF-free cells (*). Data representing a typical experiment are expressed as mean standard deviation for six mice in each group. 0
HOURS
glucocorticoid demonstrate that cachectin is synthesized by the action of CF. Cachectin is a polypeptide hormone secreted in great abundance by macrophages in response to lipopolysaccharide. It is estimated that cachectin constitutes 1 to 2% of the total secretory protein produced by lipopolysaccharide-activated macrophages (5). Experiments were also undertaken to determine whether APC could be stimulated to produce the mediator through which CF produces cachexia and enhanced plasma triglyceride levels. The soluble factor released from macrophages in vitro had many of the characteristics of the plasma-derived cachexia-inducing factor, showing the same biological effects. This implies that macrophages may be one of the cellular sources of cachexiainducing factor in mice injected with CF. On the basis of _noninjec ted mineral oil
AC-CF CF
binding studies using anti-recombinant murine TNF-a antibodies, the activity found in plasma from CF-treated mice and supernatants from CF-stimulated APC could be attributed to TNF-a. In a previous report, we demonstrated that the presence of cachectin is associated with hypertriglyceridemia in sera of patients with pulmonary tuberculosis (20). The hyperlipidemia and cachexia that accompany this infection could be mediated by the release of cachectin. The impact of the catabolic effects of cachectin on the organism when it is liberated in large amounts by CF stimulation may be sufficient to explain the emaciation and complex metabolic changes that lead to cachexia in tuberculosis patients, with catastrophic consequences for the organism. However, liberation of cachectin in small amounts by the action of CF is very advantageous for the host and could explain, at least in part, the nonspecific immunopotentiating activity of CF.
=medi um
=noni npected
dexamethosone 4
dexamethasone CF
ECF
medium + CF madlumn + cells medium + cells+
Cf
-J
0-
z
ACKNOWLEDGMENTS This research was supported by Conselho Nacional de Pesquisa grants 401007/87 and 300351/81 and by Fundac,o de Amparo a Pesquisa do Estado de Sao Paulo grant 87/1853-0. We thank Izaira Tincani and Sebastiao Lazaro Brandao Filho for technical assistance and Rosangela C. Peral Mesquita for secretarial assistance.
1--
0
FIG 5. Trgycrd cocnrtosi
lam
fmcrae
w
0
(0
FIG.
5.
Triglyceride concentrations
in
plasma of mice treated
with plasma from CF-injected mice (A), dexamethasone (B), and CF-stimulated macrophage supernatants (C) and of the respective control animals. Experimental procedures were as described in the legend to Fig. 4. Data representing a typical experiment are expressed as mean + standard deviation for six mice in each group.
LITERATURE CITED 1. Bekierkunst, A., I. S. Levij, E. Yarkoni, E. Vilkas, and E. Lederer. 1971. Suppression of urethane-induced lung adenomas in mice treated with trehalose-6,6'-dimycolate (cord factor) and living Bacillus Calmette Guerin. Science 174:1240. 2. Beutler, B., and A. Cerami. 1986. Cachextin and tumor necrosis factor as two sides of the same biological coin. Nature (London) 320:584-588. 3. Beutler, A., and A. Cerami. 1987. Cachectin: more than a tumor necrosis factor. N. Engl. J. Med. 316:379-385. 4. 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: 977-980. 5. Beutler, B., J. Mahoney, J. Le Trang, P. Pekala, and A. Cerami. 1985. Purification of cachectin, a lipoprotein lipase-suppressing
VOL. 56, 1988
6. 7.
8. 9.
10. 11. 12.
13. 14.
hormone secreted by endotoxin-induced RAW 264.7 cells. J. Exp. Med. 161:984-995. Bloch, H. 1950. Studies on the virulence of tubercle bacilli. Isolation and biological properties of a constituent of virulent organisms. J. Exp. Med. 91:197-217. Dubois, M., K. A. Gilles, J. Hamilton, P. A. Rebers, and F. Smith. 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28:350-356. Filkins, J. P. 1985. Monokines and the metabolic pathophysiology of septic shock. Fed. Proc. 44:300-304. Hotez, P. J., N. Le Trang, A. H. Fairlamb, and A. Cerami. 1984. Lipoprotein lipase suppression in 3Te-L1 cells by a haematoprotozoan induced mediator from peritoneal exudate cells. Parasite Immunol. 6:203-209. loneda, T., M. Lenz, and J. Pudles. 1963. Chemical constitution of a glycolipid from Corynebacterium diphtheriae PW8. Biochem. Biophys. Res. Commun. 13:110-114. Kato, M. 1967. Procedure for the preparation of aqueous suspension of cord factor. Am. Rev. Respir. Dir. 96:553. Kato, M., and J. Asselineau. 1971. Chemical structure and biochemical activity of cord factor analogs: 6-6'-dimycoloylsucrose and methyl 6-mycoloyl-a-D-glucoside. Eur. J. Biochem. 22:364-370. Kawakami, M., and A. Cerami. 1981. Studies of endotoxininduced decrease in lipoprotein lipase activity. J. Exp. Med. 154:631-639. Minnikin, D. E., S. M. Minnikin, G. Dobson, M. Goodfellow, F. Portaels, L. Van Den Breen, and D. Sesardic. 1983. Mycolic acid patterns of four vaccine strains of Mycobacterium bovis (BCG). J. Gen. Microbiol. 129:889-891.
CACHEXIA INDUCED BY CORD FACTOR
3071
15. Noll, H., H. Bloch, J. Asselineau, and E. Lederer. 1956. Chem16.
17.
18. 19. 20.
21.
22.
ical structure of the cord factor of Mycobacterium tuberculosis. Biochim. Biophys. Acta 20:299-309. Parant, M., F. Parant, L. Chedid, J. C. Drapier, J. F. Petit, J. Wietzerbin, and E. Lederer. 1977. Enhancement of non-specific immunity to bacterial infection by cord factor. J. Infect. Dis. 135:771-777. Rouzer, C. A., and A. Cerami. 1980. Hypertriglyceridemia associated with Trypanosoma brucei infection in rabbits: role of defective triglyceride removal. Mol. Biochem. Parasitol. 2:3138. Ruff, M. R., and G. E. Gifford. 1980. Purification and physicochemical characterization of rabbit tumor necrosis factor. J. Immunol. 125:1671-1677. Senn, M., T. loneda, J. Pudles, and E. Lederer. 1967. Spectrometrie de masse de glycolipids. I. Structure du "cord factor" de Corynebacterium diphtheriae. Eur. J. Biochem. 1:353-356. Silva, C. L., L. H. Faccioli, and G. M. Rocha. 1988. The role of cachectin/TNF in the pathogensis of tuberculosis. Braz. J. Med. Biol. Res. 21:489-492. Silva, C. L., J. L. Gesztesi, and T. loneda. 1979. Trehalose mycolates from Nocardia asteroides, Nocardia farcinica, Gordona lentifragmenta and Gordona bronchialis. Chem. Phys. Lipids 24:17-25. Thomas, D. W., A. K. Matida, C. L. Silva, and T. loneda. 1979. Esters of trehalose from Corynebacterium diphtheriae: a modified purification procedure and studies on the structure of their constituent hydroxylated fatty acids. Chem. Phys. Lipids 24: 267-282.
