George K. Mutwiri, Daniel G. Butler, S0ren Rosendal and Bill Woodward. ABSTRACT. A paired feeding experiment was conducted to investigate if reduced.
The Role of Restricted Food Intake in the Pathogenesis of Cachexia in Severe Combined Immunodeficient Beige Mice Infected with Mycobacterium paratuberculosis George K. Mutwiri, Daniel G. Butler, S0ren Rosendal and Bill Woodward
ABSTRACT A paired feeding experiment was conducted to investigate if reduced food intake is a reason for the body weight loss previously observed in severe combined immunodeficient beige (SCID bg) mice infected with Mycobacterium paratuberculosis. Mice were paired on the basis of age, litter and sex. One of each pair was injected intraperitoneally with 105 viable M. paratuberculosis organisms. The remainder served as uninfected pairfed mates. Each uninfected mouse was restricted to the amount of food (per gram body weight) that its infected paired mate ate in the previous 24 hour period starting at four weeks postinfection until 12 weeks postinfection when the mice were necropsied. The mean body weights of the two groups were not significantly different (p < 0.05) at the start of the experiment (infected 27.6 2.1 g, pairfed 27.3 3.4 g) but the pairfed group weighed less after 12 weeks of restricted food intake. Mycobacterium paratuberculosis was isolated from the spleen, liver, gut and fecal pellets of the infected but not the uninfected mice. Acid-fast bacilli were seen histologically in the liver, spleen and intestines of the infected mice only. Analysis of carcass compositions indicated that both infected and pairfed mice lost dry matter. Despite the loss in dry matter, the infected mice appeared to have maintained their body weights due to an increased retention of body water (presumably due to edema of inflammation). These ±
±
results suggest that infection of SCID bg mice with M. paratuberculosis causes a reduction in their food intake (presumably due to reduced appetite) which, in turn, contributes to a loss in dry matter. We suggest that this loss in dry matter is one of the initial events that eventually lead to cachexia, and that it precedes the body weight loss that inevitably occurs in SCID bg mice chronically infected with M. paratuberculosis.
RESUME
l'experimentation mais le groupe ayant eu une alimentation restreinte pesait moins a la fin de l'experience. Mycobacterium paratuberculosis a ete isole de la rate, du foie, de l'intestin et des feces des souris infectees seulement. A l'examen histologique du foie, de la rate et des intestins des animaux infectes on pouvait voir des micro-organismes alcoolo-acido-resistants. L'analyse de la composition des carcasses indiquait une perte de matiere seche autant chez les animaux infectes que chez les non-infectes. Malgre la perte de matiere seche, les souris infectees semblaient avoir maintenues leur poids corporel a cause d'une augmentation de la retention de liquide corporel, due probablement a l'oedeme de la reaction inflammatoire. Ces resultats indiquent qu'une infection par M. paratuberculosis chez les souris souffrant d'immunodeficience combinee entraine une reduction de la prise alimentaire (associee probablement a une diminution d'appetit), ce qui contribue par la suite a une perte de matiere seche. Cette perte de matiere seche precederait la perte de poids corporel et la cachexie observe'e chez les souris souffrant d'immunodeficience combinee infectee par M. paratuberculosis. (Traduit par Dr Serge Messier)
La presente etude avait pour but de determiner si la diminution de poids corporel observee lors de l'infection par Mycobacterium paratuberculosis chez des souris souffrant d'immunodeficience combinee etait causee par une diminution de la prise de nourriture. Les souris etaient pairees en fonction de l'age, la portee et du sexe. Une des souris de chaque paire recevait 105 cellules viables de M. paratuberculosis par injection intra-peritoneale, alors que les autres servaient de temoins non-infectes. A partir de la quatrieme semaine apres l'infection jusqu'au moment de la necropsie a la douzieme semaine post-infection, chaque souris non-infectee ne recevait en nourriture que la quantite (par gramme de poids corporel) consommee par l'autre moitie de la INTRODUCTION paire durant la periode de 24 heures Paratuberculosis is an insidious qui precedait. Aucune diffe'rence significative dans le poids corporel chronic enteritis of ruminants caused moyen des animaux des deux by infection with Mycobacterium groupes n'etaient notee au debut de paratuberculosis (MPTB). It occurs
Department of Veterinary Microbiology and Immunology (Mutwiri, Rosendal), Department of Clinical Studies (Butler) and Department of Nutritional Sciences (Woodward), University of Guelph, Guelph, Ontario NIG 2W1. This work was supported by OMAF Red meat II project, the CIDA-Kenya GTF project and NSERC STRGP 45396. Submitted August 16, 1993.
40
Can J Vet Res 1995; 59: 40-45
world wide and causes significant economic losses (1). The immunology and pathogenesis of this disease are poorly understood (2,3). A consistent characteristic of clinical paratuberculosis in ruminants is a progressive loss of body condition which leads to terminal cachexia. It has been proposed that this cachexia is either due to a protein-losing enteritis (3) or to immunological mechanisms involving competent lymphocytes (4,5), but neither has been conclusively shown to be the cause. A more recent hypothesis suggests that the cachexia of paratuberculosis is caused by tumor necrosis factor-alpha (TNF-a), a peptide known to cause cachexia in laboratory animals (6,7). Mycobacterial infections, and more recently M. paratuberculosis have been shown to prime macrophages to produce high levels of TNF-ot (8,9,10). Nonetheless a more reasonable assumption would be that several factors are involved in the pathogenesis of the cachexia. It is thought that one reason for the slow progress in the understanding of the itnmunopathogenesis of this disease is the lack of a suitable laboratory animal model. We have described a severe combined immunodeficient beige (SCID bg) mouse model for the study of paratuberculosis (11). The SCID bg mouse strain has two mutations that make it particulary susceptible to infections. The SCID mutatioi, results in a failure to generate mature and functional T and B lymphocytes (12) while the beige mutation causes reduced natural killer cell activity (13). In our previous research, SCID bg mice following a single intraperitoneal (IP) injection of 105 colony-forming units (CFU) of M. paratuberculosis became infected and developed cachexia similar to clinical paratuberculosis. No TNF-ao was detected in the plasma of these mice. We hypothesized that malnutrition could contribute to the cachexia reasoning that since mice infected with M. paratuberculosis are known to develop anorexia, their food intake would be expected to fall. This would in turn be reflected as a loss in body weight or in one or more of the body constituents (i.e. dry matter, protein, lipid). Therefore, this study was conducted (i) to determine, whether restriction of normal SCID bg mice to the weight of food per
g body weight consumed by matched but M. paratuberculosis-infected SCID bg mice littermates would be sufficient to maintain their body weights, (ii) to determine, by carcass analysis, the changes in body composition that occur during infection and in uninfected mice restricted to the dietary intake of infected paired mates, and (iii) to measure the levels of plasma TNF-ao of mice at various times during the course of infection.
MATERIALS AND METHODS
-20°C until processed. Plasma from blood samples collected every four weeks from the orbital plexus was stored at -20°C for TNF-ot analysis. All mice were weighed to determine final body weights at killing 12 weeks postinoculation. Once killed, five mice from each group were stored in individual air-tight sealed bags at -20°C until carcass analysis was done. The remaining mice in each group were necropsied and samples for bacteriology and histopathology taken as previously described (1 1). CONTROL ANIMALS
EXPERIMENTAL ANIMALS
Severe combined immunodeficient beige mice (16-19 weeks old) were housed individually in microisolator cages within a Horsfall unit, in the isolation facility of the Ontario Veterinary College. The protocol met the guidelines set by the Canadian Council on Animal care (Guide to the Care and Use of Experimental Animals Volumes 1 and 2), and was approved by the Animal Care Committee of the University of Guelph. The mice were fed the same batch of feed (Autoclaved Mouse Chow, Charles River, St. Constant, Quebec). Clean autoclaved tap water was available to the mice ad libitum. Mice did not receive any prophylactic antibiotic treatment. Eight infected mice injected IP at time 0 with 105 CFU of MPTB of bovine origin (11) and another eight uninfected (pairfed) mice were paired on the basis of age, sex and litter. The animals were observed for clinical signs daily and weighed each week. Starting four weeks after infection (by which time the infection is known to be established), pairfeeding was started. It was conducted as described in (14). Briefly, each infected mouse was weighed each day and its food intake (per g body weight) over a 24 h period was determined at 1400 h each day by the difference between the weight of food provided and the weight of food left after 24 h. Each pairfed mouse was weighed daily and provided with the amount of food per g body weight (BW) that its paired infected mate ate during the preceding 24 h period. Fecal samples were collected from all cages once every two weeks for MPTB culture and stored at
In order to obtain control carcass composition values for comparison with the two experimental groups, 20 SCID bg mice, matched for age were obtained from the same colony. Ten of these mice were weighed at time 0 (t = 0), and killed and their carcasses were analyzed to provide baseline carcass composition data (baseline controls). This was necessary as it provided data with which to predict carcass compositions at t = 0 for the three groups of mice kept for the 12 week experimental period. The remaining ten mice were fed ad libitum (food intake was not determined) on a diet similar to that fed to the experimental animals, weighed weekly and killed after 12 weeks comprising the final controls. ANALYSES
Plasma (for TNF) and fecal pellets (for the culture of MPTB) and necropsy samples (liver, spleen and small intestine for light microscopy) were processed and analyzed as previously described (11). Briefly, TNF-ao was analyzed using a commercial ELISA (Genzyme Corp., Boston, Massachusetts). Fecal pellets and organs were decontaminated with 0.75% hexadecylpyridinium chloride (Sigma Chemical Co., St. Louis, Missouri), inoculated on Herrold's egg yolk medium (HEYM) with or without mycobactin J and incubated at 37°C for at least eight weeks. Serial sections of formalin-fixed tissues (liver, spleen and small intestine) were stained with hematoxylin and eosin, and using the ZiehlNeelsen technique for acid-fast organisms and examined under a light
microscope.
41
fected pairfed and final controls over the entire 12 week period. Although the mean body weights of the infected (27.6 ± 2.1 g) and pairfed (27.3 ± 3.4 g) T T T,& T groups were similar at the start of the experiment, after the food intake of the uninfected mice was restricted to T that of their infected paired mates, the mean body weight of the former fell 4 and remained below that of the T T L* infected group (Fig. 1). Although the mean body weights of the two control groups were higher than the mean weights of the experimental groups at t = 0 (Table I), this difference reflects the usual variation in body weight for age observed in this recently established mutant line. TNF-a was not detected in plasma of any mice (data not shown). 0, Mycobactin J dependent mycobacteria were isolated from feces and organs of infected mice only. Mycobacterium paratuberculosis were detected in the feces of infected mice 8 to 12 weeks after infection. At necropsy the livers of the infected mice were mottled and the spleens enlarged compared to those of the pairfed mice. Macrophages in livers, spleens and lamina propria of small intestine of infected mice in contrast to pairfed mice contained acid-fast (AF) bacilli and were found in association with many mononuI I I I I I clear cells and a few polymorphonu'1 o 111 2 clear 4 C)7 8 Theneutrophils. body weights (initial and final and carcass composition values only) for the four groups of mice are shown in Table I. Comparing the two control groups, it is evident that the carcass Fig. 1. Mean weekly body weights (+ SD) of three gro tups of SCID bg mice. Infected mice (0) copsiin of th baeieania were injected IP with 105 MPTB at t = 0. Uninfecte d pairfed mice (V) were matched with infected on the basis of a2e. sex and litter Drior to t = O and their food intake restricted to controls did not change during the that of infected beginning at four weeks after t = 0 up to 12 weeks when all the mice were 12 week period. killed. Final controls (O) were age matched with the other two groups and fed ad libitum for The infected mice had the lowest 12 weeks and killed. The mean body weights of the infected and pairfed groups were statisti- carcass DM %, followed by the two cally different starting at week 7 postinfection and onwards. control groups. The pairfed mice had the highest DM %. However, a comData were analyzed using SAS (17) parison of the absolute DM values of To determine carcass dry matter (DM), each entire carcass (including and the means of the various groups the infected with that of the pairfed the tail) was initially freeze-dried, were compared using Duncan's new indicated that their carcass dry matter ground to powder and dried overnight multiple range test. The level of sig- contents were not significantly different (infected DM was 8.7 ± 1.4 g, to constant weight in a vacuum oven nificance was set at p = 0.05. pairfed DM was 9.3 ± 3.5 g). at 90°C. Lipid and crude protein conBy using the DM % value of basetents were determined in duplicate RESULTS line controls (36.3 ± 3.1%) to estimate analyses (1.5-2.0 g each) and the the DM of pairfed and infected mice at One infected and two pairfed mice results expressed as percentages of dry carcass weight. The methods of died of unknown causes during the t = 0 (baseline DM % x body wt at t = 0), it was possible to determine the Bligh and Dyer (15) and Kjeldahl (16) experiment. were used for the analysis of total carFigure 1 shows the mean weekly DM changes in each of the three cass lipid and protein respectively. body weights of the infected, unin- groups of mice at the end of the experi-
35
T
0L
30
M0 25
120
4 15
10
0 1 2 3 4 5 Postinfection Time (weeks)
42
TABLE I. Comparisons of body weights and carcass compositions of four groups of SCID bg mice: SCID bg mice infected with 105 CFU of MPTB (Infected, n = 5); uninfected SCID bg mice paired to infected mice by age, sex and litter and feed intake restricted to that of infected mates (Pairfed, n = 5); uninfected SCID bg mice fed ad lib (Final controls, n = 9). The infected, pairfed and final control groups were killed after 12 weeks. Uninfected SCID bg mice were age-matched with those in the experimental groups, but were killed at time = 0 (Baseline controls, n = 10)
Pairfed Infected 27.3 ± 3.4a 27.6 ± 2.1 a Initial BW (g/mouse) 23.2 ± 6.8b 27.5 ± 1.2a *Final BW (g/mouse) 39.5 ± 3.31 31.5 ± 4.1a Dry matter(% w wt) 68.8 ± 1.6" 74.4 ± 3.9a Moisture (% fat free wt) 12.0 ± 6.3 9.1 ± 4.7 Lipid (% wet wt) 17.8 ± 4.2 15.9 ± 3.9 Crude Protein (% wet wt) Abbreviations: ± SD, BW; body weight, wwt; wet weight Superscripts denote statistical significance. Values with different cantly different (p < 0.05) * Final body weight just before mice were killed
Final controls 32.1 ± 0.9b 32.7 ± 1. Ia 35.7 ± 2.3c 72.0 ± 1.8c 10.6 ± 1.8 18.1 ± 0.9
Baseline controls 30.2 ± 2.4b 36.3 ± 3.1"' 72.0 ± 1.3c 11.5 ± 3.0 18.1 ± 0.6
superscript letters are signifi-
TABLE II. The dry matter changes in three groups of mice (infected, pairfed and final controls) after a 12 week period. Dry matter (g/mouse) of the mice at t = 0 was estimated by multiplying mean DM % of baseline controls (36.3%) with their respective initial weights and determining the group means (± SD). The difference between initial and final DM values indicates the estimated change in DM Initial Final Difference
Infected 10.0 ± 0.8 8.7± 1.4
1.3
mental period (Table II). Infected and pairfed mice appeared to have lost DM while control mice maintained their carcass DM during this period. Percent carcass moisture on a fatfree basis (carcass water/wet weightlipid weight X 100) was significantly highest in the infected group, followed by the two control groups and then the pairfed mice. Carcass lipid % and crude protein % (as percentages of their final body weights) was not statistically different in any of the four groups, though both appear to be lower in the infected group (Table I). This is a reflection of the high water content of carcasses from infected mice. Based on absolute carcass content of lipid and crude protein (CP) (infected: lipid was 2.5 ± 1.3, CP was 4.4 ± 1.2; pairfed: lipid was 3.1 ± 2.4, CP was 3.9 ± 0.3), the losses of these constituents were not statistically different in the infected and the pairfed mice.
Pairfed 9.9 ± 1.3 9.3 ± 3.5 0.6
Final controls 11.7 ±0.3 11.7± 1.0 0.0
DISCUSSION In previous experiments, we have established that SCID bg mice infected with viable MPTB lose weight after 12 weeks of infection (1 1). In this experiment, infected animals did not appear to lose body weight during 12 weeks of infection. A very significant observation, however, was that they did lose DM (Table II). One possible explanation for this difference is that in this experiment we used older mice than previously, a factor that needs further investigation. Also, we think that if the experimental period had been extended beyond 12 weeks, infected mice in this experiment would also, eventually have lost weight. In our laboratory, mice inoculated with 102 as opposed to 105 (in this study) M. paratuberculosis, lost considerable body condition by 32 weeks after infection (unpublished observations). Pairfeeding has been used by others in animal nutrition experiments (14, 18,19) where one animal of a pair
(presumably the one that consumes less) determines the intake of the pairfed mate. In this experiment, we hypothesized that M. paratuberculosis infected mice would eat less. Accordingly, their actual daily food intake (per g/BW) was determined and their pairfed mates were restricted to this rate of food intake. When uninfected pairfed mice were restricted to the food intake of their infected mates, they lost body weight compared to the infected mice (Fig. 1). This implies that M. paratuberculosis infected SCID bg mice ate less (presumably due to reduced appetite), such that their food intake was at a level below that necessary to maintain the body weight of mice with normal active behavior. Had the feed intake of the infected mice been at a normal level, then the pairfed mice would not have lost body weight. Rather, they would have maintained their body weight as did the final controls. These findings concur with those of Madge (20), who noted that some C57 mice infected with M. paratuberculosis lost appetite and consequently body weight. The purpose of carcass analysis was to actually assess how the body content of water, protein and fat in the SCID bg mouse changed in response to feed restriction and infection. We consider that using DM % of the baseline controls (same age and from same colony) for estimating DM of infected, pairfed and final controls at t = 0 is reasonable since control mice not restricted in terms of food intake, maintained their carcass compositions during the experimental period. According to these data (Table II), although both infected and pairfed groups lost DM, infected mice seem to have lost almost twice as much DM. All the DM lost by the pairfed mice can be attributed to restricted food intake (final controls did not lose any DM). Since their intake was determined by the amount of food consumed by the infected mice, it is reasonable to conclude that M. paratuberculosis infection blunts the appetite of mice and was responsible, at least in part, for the loss in carcass DM. However since infected mice lost more DM than the pairfed, we can only assume that another factor(s), in 43
addition to nutrition, acting independently or together, must also be at work in the infected mice. At the moment, we can only speculate that this could be due to the hypermetabolism induced by infection. Chronic infection tends to induce a hypermetabolic or catabolic state in which most of the increased energy expended is derived from fat (80%), and the rest from protein catabolism (reviewed in 21,22). However, it is imperative to remember that in this experiment, although the animals were maintained as gnotobiotics, other subclinical infections cannot be completely ruled out since they did not receive any prophylactic treatment. Loss of carcass DM (1.31 g) by the infected mice without a proportionate loss in body weight (initial mean BW 27.6 g, final 27.5 g) implies that these mice must have accumulated another body constituent. The only body constituents (DM) that were not measured in this experiment were nitrogen-free extract (NFE) and ash. However, when the percentages of these carcass constituents were derived by calculation [NFE + ash % = DM % -(lipid % + CP %)] and compared between the three groups, they were not significantly different (NFE + ash: infected 6.5 ± 1.1%; pairfeds 8.9 ± 2.9% final controls 7.0 ± 1.6%). Thus we can only attribute this apparent maintenance of body weight (despite a loss in carcass DM) to a gain in body water by infected mice. This is evident since body water (reported as moisture % on a fat free basis) was increased relative to the baseline group (Table I). Increase in carcass moisture (% fat free basis) implies that there was water retention in tissues other than fat and can reasonably be attributed to edema of inflamed tissues. Since ash was not actually measured, there is not sufficient information to explain the rather low moisture (on a % fat free basis) of the pairfed mice (68.8 ± 1.6%) compared to both control groups whose moisture contents were identical (baseline controls 72.0 ± 1.8%; final controls 72.0 ± 1.3%). Nevertheless, an increase in ash in the pairfed (which likely occurred) would be reflected as an increase in DM % and a decrease in moisture %. 44
Mycobacterium paratuberculosisinfected and diet restricted pairfed SCID bg mice apparently mobilize similar types of tissue since their carcass fat and crude protein contents were not different. It would seem reasonable to attribute the gain in carcass water by infected mice to local inflammation (presumably as edema) in target organs particularly the intestine. We have observed that in the advanced stages of this mycobacterial infection in SCID bg mice, the ileum appears grossly thickened and the muscle layers more fragile and subject to tearing during routine manipulation (unpublished observations). However, we also noted in our earlier report that in advanced stages of the infection, other body tissues (skin, muscles, peritoneum) appeared very dry, especially the muscular tissues (1 ). We suggest that these observations are not in conflict with our current results because it is possible that some organs (e.g. muscle) could lose water together with protein and fat while other organs such as the gut could retain water as a result of inflammatory edema leading to a net increase in carcass water. In fact, Madge (20) found that some C57 mice infected with M. paratuberculosis lost body weight while the small intestine had gained weight and water, but the protein content of this organ was decreased. No TNF-ot was detected in the plasma of any mice, a finding that confirms our previous observation (1 1). It has been reported that in BCG infection in immunocompetent mice, local TNF synthesis occurs in granulomas in the liver but is not detectable systemically (23). This suggests how TNF-oL could be involved in weight loss, exert a local effect and yet be undetectable in the systemic circulation. Further, the chronic nature of MPTB infections and the extremely short half-life of TNF-a (24), may also preclude its detection in the plasma. Undernourishment and infection, if concurrent, make intestinal injury more severe and chronic (25). More work needs to be done to understand the intricate relationship that exists between MPTB infection, malnutrition, intestinal injury and the cachexia of paratuberculosis.
Although information derived from the infection in SCID bg mice cannot be directly extrapolated to the spontaneous disease in domestic ruminants, it does provide insights into the pathogenesis of M. paratuberculosis infections. In conclusion, restricting noninfected pairfed SCID bg mice to the food intake of their infected mates implicated reduced food intake as one of the initial triggers in the overall pathogenesis of weight loss in experimental MPTB infection. As we have clearly demonstrated, caution must be exercised in interpreting changes in body weights in such experiments. The changes in body constituents that determine changes in body weight are apparently different in health, restricted food intake and in M. paratuberculosis infection due to differences in metabolism.
ACKNOWLEDGMENTS We thank M.P. Ruka, A. Bashir, 0. Oliver and B. Blake for the help with the care and feeding of mice, M. Shoukri for help with the statistical analysis, and S. Atmore, H. Hunter and S. Tatarski for their technical assistance.
REFERENCES 1. BLOOD DC, RADOSTITS OM. Johne's disease. Veterinary Medicine. 7th ed. Bailliere and Tindall 1989: 722-729. 2. BENDIXEN PH. Immunologic reactions caused by infections with Mycobacterium paratuberculosis. A review. Nord Vet Med 1978; 30: 163-168. 3. CHIODINI RJ, VAN KRUININGEN HJ, MERKAL RS. Ruminant paratuberculosis (Johne's disease): The current status and future prospects. Cornell Vet 1984; 74: 218-262. 4. MERKAL RS, KOPECKY KE, LARSEN AB, NESS RD. Immunologic mechanisms in bovine paratuberculosis. Am J Vet Res 1970; 31: 475-485. 5. MERKAL RS, WITZEL DA. Passive transfer of Johnin hypersensitivity with reactivity resembling clinical paratuberculosis in calves. Infect Immun 1973; 7:196-198. 6. OLIFF A, DEFEO-JONES D, BOYER M, MARTINEZ D, KEIFER D, VUOCOLO G, WOLFE A, SOCHER SH. Tumour secreting human TNF/cachectin induce cachexia in mice. Cell 1987; 50: 555 -563.
7. TRACEY KJ, WEI H, MANOGUE KR, FONG Y, HESSE DG, NGUYEN HT, KUO GC, BEUTLER B, COTRAN RS, CERAMI A, LOWRY SF. Cachectin/ tumor necrosis factor induced cachexia, anemia, and inflammation. J Exp Med 1988; 167: 1211-1227 8. SILVA CL, FACCIOLI LH. Tumor necrosis factor (cachectin) mediates induction of cachexia by cord factor from mycobacteria. Infect Immun 1988; 56: 3067-3071. 9. VALONE SE, RICH EA, WALLIS RS, ELLNER JJ. Expression of tumour necrosis factor in vitro by human mono nuclear phagocytes stimulated whole Mycobacterium bovis BCG and mycobacterial antigens. Infect Immun 1988; 56: 3313-3315. 10. HAMILTON HL, COOLEY AJ, ADAMS JL, CZUPRYNSKI CJ. Mycobacterium paratuberculosis monoassociated nude mice as a paratuberculosis model. Vet Pathol 1991; 28: 146-155. 11. MUTWIRI GK, BUTLER DG, ROSENDAL S, YAGER J. Experimental infection of severe combined immunodeficient/ beige mice with Mvcobacterium paratuberculosis of bovine origin. Infect Immun 1992; 60: 4074 -4079. 12. BOSMA GC, CUSTER RP, BOSMA MJ. A severe combined immunodeficiency mutation in mice. Nature (Lond) 1983; 301: 527-530.
13. CROY BA, CHAPEAU C. Evaluation of pregnancy immunotropism hypothesis by assessment of reproductive performance of young adult mice of genotype SCID. SCID.bg.bg. J Reprod Fertil 1990; 88: 23 1-239. 14. MAYNARD LA, LOOSLI JK. Feeding experiments. The determination of digestibility. In: Animal Nutrition. McGraw-Hill, 1962: 283-309. 15. BLIGH EG, DYER WJ. A rapid method of total lipid extraction and purification. Can J Biochem Pysiol 1959; 37: 911-917. 16. HORWIZT W. Official Methods of Analysis of the Association of Analytical Chemists. I Ith ed. Washington DC: AOAC, 1970. 17. SAS INSTITUTE. SAS version 6. Cary, North Carolina: SAS Institute, Inc., 1985. 18. SHERMAN P, FORSTNER J, ROOMIE N, KHATRI I, FORSTNER G. Mucin depletion in the intestine of malnourished rats. Am J Physiol 1985; 248: G418-G423. 19. COHEN MB, NOGUEIRA J, LANEY Jnr DW, CONTI TR. The jejunal secretory response to Escherichia coli heatstable enterotoxin is prolonged in malnourished rats. Pediatr Res 1992; 31: 228-233. 20. MADGE DS. Malabsorption in C57 mice experimentally infected with Johne's disease. Comp Biochem Physiol [A] 1970; 40: 649-648.
21. FRYAN KN. Hormonal control of metabolism in trauma and sepsis. Clin Endocrinol 1985; 24: 577-599. 22. STONER HB. Interpretation of the metabolic effects of trauma and sepsis. J Clin Pathol 1987; 40: 1108 -1117. 23. KINDLER V, SAPPINO AP, GRAU GE, PIGUET PF, VASSALLI P. The inducing role of Tumour necrosis factor in the development of bactericidal granulomas during BCG infections. Cell 1989; 56: 731-740. 24. BEUTLER B, CERAMI A. Cachectin: A monokine implicated as a mediator of cachexia and shock. In: Lymphokines. Academic Press, 1987: 203-222. 25. BUTZNER JD, HALL DG. Effects of chronic protein-calorie malnutrition on small intestinal repair after an acute bacterial enteritis: A study in infant rabbits. Pediatr Res 1988; 23: 408-413.
45
