Veterinary Immunology and Immunopathology, 26 (1990) 171-182
17'1
Elsevier Science Publishers B.V., Amsterdam
Comparative histopathology in BALB/c mice infected with virulent and attenuated strains of Brucella abortus
F.M. Enright a, L.N. Araya b, P.H. Elzer b, G.E.
R o w e b a n d A.J. W i n t e r b
aDepartment of Veterinary Science, Louisiana Agricultural Experiment Station, Louisiana State University Agricultural Center, Louisiana State University, Baton Rouge, Louisiana 70803, U.S.A. bDepartment of Veterinary Microbiology, Immunology, and Parasitology, New York State College of Veterinary Medicine, Cornell University, Ithaca, New York 14853, U.S.A. (Accepted 12 February 1990)
ABSTRACT Enright, F.M., Araya, L.N., Elzer, P.H., Rowe, G.E. and Winter, A.J., 1990. Comparative histopathology in BALB/c mice infected with virulent and attenuated strains of Brucella abortus. Vet. lmmunol. Immunopathol., 26:17 l - 182. The course of infection in BALB/c mice of virulent Brucella abortus strain 2308 (S-2308) and attenuated strain 19 (S- 19 ) varies markedly. Whereas S- 19 is eliminated at an exponential rate beginning at 2 weeks post infection (p.i.), strain 2308 assumes a steady state or plateau during the first 6 weeks p.i. and thereafter is eliminated very slowly over a period exceeding 6 months. Here we compared the initiation and maintenance of inflammatory reactions in spleens and livers of mice infected with either of the two strains of B. abortus for the first 6 weeks p.i. Histological changes in the liver were similar in response to either strain and were characterized by the development of small granulomas and an influx of polymorphonuclear leukocytes (PMN) and monocytes. Tissue reactions in the spleen were similar at weeks 1 and 2 p.i. At 3 weeks p.i. and thereafter, focal granuiomatous responses in S-2308-infected mice exceeded those in mice infected with S-19. Numbers of nonspecific esterase (NSE) positive mononuclear leukocytes in S-19-infected spleens had increased by 3 weeks p.i. and remained elevated. No comparable increase in NSE positive cells occurred in mice infected with S-2308, and numbers were significantly lower. At 4 weeks p.i. the influx of mature neutrophils and the intensity of extramedullary hematopoiesis were significantly greater in S-19-infected spleens. A profound depletion of periarteriolar lymphoid tissue was noted in both infections for the first 3 weeks p.i. However, repopulation of lymphoid sheaths in S-19-infected spleens became significantly greater by 4 weeks p.i. and continued to increase at significantly higher levels during the next 2 weeks. This study demonstrates quantitative differences in splenic inflammatory responses which are temporally related to the more rapid elimination of S- 19. Based upon the lower susceptibility of strain 2308 to the protective effects of immune serum it is hypothesized that the different patterns of infection and inflammation displayed by the 2 strains may relate to the differential capacities of antibody opsonized S-19 and S-2308 to survive in activated macrophages.
0165-2427/90/$03.50
© 1990 - - Elsevier Science Publishers B.V.
172
F.M. ENRIGHT ET AL.
INTRODUCTION
Brucella abortus strain 2308 (S-2308) is representative of virulent field strains which produce abortion in cattle and cause chronic infections in the face of the host's immune response (Subcommittee on Brucellosis Research, 1977 ). Strain 19 (S- 19 ), on the other hand, is an attenuated strain which has for many years been used in cattle as a living vaccine (Subcommittee on Brucellosis Research, 1977). The relative virulence of these strains for cattle is paralleled by their patterns of infection in BALB/c mice, in that S-2308 persists in the spleen at high levels for several months whereas S-19 is cleared much more rapidly (Montaraz and Winter, 1986 ). The host-parasite interactions responsible for the capacity of virulent strains of B. abortus to produce persistent infections are not understood. The purpose of the work reported here was to perform a comparative histopathological study of BALB/ c mice infected with S-2308 and S-19. The studies encompassed the first 6 weeks post infection (p.i.) because the course of infection with the two strains varies markedly during this time (Montaraz and Winter, 1986 ). We hypothesized that the inflammatory reactions engendered in the spleen and liver by S-2308 and S-19 would have distinctive features which would bear upon the different kinetics of infection and could provide further insight into the basis of persistent infection by this facultative intracellular parasite. MATERIALS AND METHODS
Mice Nine-week-old BALB/c ByJ mice were obtained from the Jackson Laboratories (Bar Harbor, ME) and held for 1 week before use. Bacterial strains Stock cultures of B. abortus S- 19 were obtained from Dr. Dale Angus (National Veterinary Service Laboratories, Ames, IA) and from commercial vaccine vials (Professional Biological Co., Denver, CO). B. abortus S-2308 was passaged in BALB/c mice and isolated in pure culture from infected spleens. Stock cultures were prepared by growing strains for 48 h on Schaedler blood agar plates. Bacteria were then suspended in Brucella broth (BBL Microbiology Systems, Cockeysville, M D ) to an estimated concentration of 1 X 109 bacteria per ml, distributed in aliquots, and stored at - 7 0 ° C. Inocula were prepared from contents of freshly thawed vials diluted in sterile phosphate buffered saline (PBS) to an estimated concentration of 5 X 105 CFU per ml. The exact concentration was determined retrospectively by viable counts (Montaraz and Winter, 1986 ). Stocks of S- 19 and S-2308 used in these studies were designated 19.85.1, 19.85.2, 19.87.4, 2308.85.1, 2308.86.1, and 2308.87.2.
HISTOPATHOLOGY OF BALB/c MICE INFECTED WITH BRUCELLA ABORTUS
173
Experimental design Mice were infected intravenously (i.v.) with approximately 5 × 104 CFU of either S-I 9 or S-2308. Groups of mice were sacrificed at weekly intervals from I to 6 weeks p.i. for bacteriological and histological examination. Six such experiments, consisting of three repetitions for each strain, were performed over a period of 18 months. Experiments are designated according to the infecting stock strain as follows: experiment l, 19.85.1; experiment 2, 19.85.2; experiment 3, 19.87.4; experiment 4, 2308.85.1; experiment 5, 2308.86. l; and experiment 6, 2308.87.2. Group sizes were five except in experiments l and 4 in which the group size was three. At the time of sacrifice transverse sections of spleen ( 1.0-1.5 mm) were fixed in a cold solution of 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.5. The remainder of the spleen was homogenized in 10 ml PBS, serially diluted, and plated in triplicate (Montaraz and Winter, 1986 ). Colonies were counted after 3 days of incubation at 37 °C under 10% CO2. In experiments l and 4 the left lateral lobe of the liver was likewise sampled for histological examination and cultured. The fixed tissues were maintained at 4 oC. At the completion of each experiment tissues were coded and shipped refrigerated to a co-investigator (FME). The coded tissues were processed and examined as unknowns. Tissue processing and staining Fixed samples of tissues were placed into infiltrating solutions from one of two kinds of plastic embedding kits and held at 4°C for 4 to 7 days. Infiltrated samples were allowed to polymerize with plastic embedding media (DuPont Sorvall, Wilmington, DE; or LKB, Rockville, MD) at either 4 °C or at room temperature, depending on the manufacturer's instructions. Two or three adjacent sections, 2 to 3 #m thick, were prepared. All sections were stored at 4°C until stained. One section of spleen and liver was stained with hematoxylin and eosin (H&E) and an adjacent section of spleen was stained for nonspecific esterase (NSE) activity by the method of Koski et al. as modified by Jeffers et al. (1987). Sections were reacted with alpha-naphthyl butyrate (Sigma Chemical Co., St. Louis, MO) for 45 rain at 37°C. It was often necessary to examine NSE stained sections with either phase or differential interference microscopy to better visualize the morphological relationships of stained and unstained cell populations. Tissue reaction scores in the spleen The number of granulomas, degree of infiltration with mature polymorphonuclear leukocytes (PMN), degree of hematopoietic activity, and numbers and staining intensity of NSE positive mononuclear leukocytes were determined subjectively and scored on a scale of 0 to 3. Zero was equal to no reaction, 1 was a slight reaction, 2 a moderate reaction, and 3 was an exten-
174
F.M. E N R I G H T ET AL.
sive reaction. The depletion of normal periarteriolar lymphoid sheaths (PALS) resulted in sheath and follicular structures containing sheets of histiocytic cells and fibrous stroma. This change was scored on a scale of 4 to 0. Four represented the normal pattern of thick lymphoid sheaths with prominent follicles. Three was equal to a slight decrease in the size of both sheaths and follicles. Two was indicative of moderate depletion of lymphoid tissues. One was equal to severe depletion with only scattered lymphocytes remaining. Zero represented complete loss of all lymphoid elements. Infection with either strain of B. abortus resulted in focal granulomatous lesions. Granulomas consisted of moderate to large numbers of histiocytic cells mixed with smaller numbers of monocytes, lymphoid cells and granulocytic cells. Larger granulomas with necrotic centers were rarely noted. It was extremely difficult to differentiate between some histiocytic granulomas and the stromal tissue which remained following lymphoid depletion. The most consistent feature in differentiating these two reactions was the close spatial relationship of the few remaining lymphoid cells to the delicate fibrous stroma which delineated the original sheath or follicle. Granulomas were located randomly throughout the red pulp and had indistinct margins. One to 3 granulomas per section of spleen were scored from 0.1 to 1.0, depending on size. Four to 6 granulomas received scores of 1.1 to 2.0, and 7 to 10 granulomas were scored between 2.1 to 3.0. Granuloma scores were most subject to sampling variation due to the focal nature of the lesions within the spleen. Other changes were more generalized in distribution. Examination of sections from age and sex matched noninfected mice allowed the establishment of baseline reaction scores. Spleens of normal mice contained no granulomas (score 0), a modest number of mature P M N within the red pulp (score 0.5 ), a thin subcapsular zone of predominantly mononuclear hematopoietic activity (score 0.5 ), and a very thin zone of intensely NSE stained monocytes adjacent to the PALS (score 1.0). Positively stained foci of 2 to 3 cells were also present within the lymphoid tissue. The normal splenic lymphoid pattern consisted of thick cellular sheaths containing large lymphoid follicles (score 4.0). Tissue reaction scores in the liver The number and size ofgranulomas and the degree of infiltration with mature P M N or monocytes were scored on a scale of 0 to 3, as described above. Statistical analysis A log value of bacteria in each organ was obtained by averaging the triplicate counts following log conversion (Montaraz and Winter, 1986). Liver counts were multiplied by 3 prior to log conversion to adjust numbers for the entire organ. Statistical significance was determined by Fisher's protected least-significant-difference test (Snedecor and Cochran, 1989). Reaction
HISTOPATHOLOGYOF BALB/c MICE INFECTED WITH BRUCELLA ABORTUS
175
scores for individual spleens and mean scores of similarly treated mice were compared by analysis of variance using the general linear model SAS procedure (SAS, 1987). RESULTS
Comparison of growth curves between S-19 and S-2308 Splenic growth curves of S-19 in experiments 1 and 3 (Fig. 1A) were typical (Ho and Cheers, 1982; Montaraz and Winter, 1986) and did not differ significantly from each other at any week p.i. However, counts from experiment 2 were below those of experiments 1 and 3 at 1 week p.i. and were uniformly greater at weeks 3 through 6 p.i. (Fig. 1A). These differences were statistically significant at weeks 1 and 3 ( P < 0.01 ). When organisms isolated at 5 weeks p.i. from spleens of mice in experiment 2 were used to infect another group of mice, the growth curve was typical of S-19 (data not shown). Splenic growth curves of S-2308 in experiments 4, 5, and 6 (Fig. 1A) were typical (Montaraz and Winter, 1986) and were not significantly different from one another, with one exception at week 5 p.i. Growth curves of S-19 and S-2308 in the liver paralleled those in the spleen, but bacterial numbers were always lower by 1 to 2 logs (data not shown). Histological changes in the spleen Statistical comparisons were made of reaction scores of individual S- 19 experiments against individual S-2308 experiments as well as between pooled data from S-19 experiments 1 and 3 and S-2308 experiments 4, 5, and 6 (Fig. 1 ). Pooling of the data was justified because the same results were obtained from pooled as from individual comparisons. The aberrant S-19 infection (experiment 2 ) was analyzed separately (Fig. 1 ). At 1 week p.i. a significant difference between the groups was noted only in the degree of lymphoid depletion, which was moderate in the mice infected with S-2308 but only slight in those infected with S-19 (P
17'1
Elsevier Science Publishers B.V., Amsterdam
Comparative histopathology in BALB/c mice infected with virulent and attenuated strains of Brucella abortus
F.M. Enright a, L.N. Araya b, P.H. Elzer b, G.E.
R o w e b a n d A.J. W i n t e r b
aDepartment of Veterinary Science, Louisiana Agricultural Experiment Station, Louisiana State University Agricultural Center, Louisiana State University, Baton Rouge, Louisiana 70803, U.S.A. bDepartment of Veterinary Microbiology, Immunology, and Parasitology, New York State College of Veterinary Medicine, Cornell University, Ithaca, New York 14853, U.S.A. (Accepted 12 February 1990)
ABSTRACT Enright, F.M., Araya, L.N., Elzer, P.H., Rowe, G.E. and Winter, A.J., 1990. Comparative histopathology in BALB/c mice infected with virulent and attenuated strains of Brucella abortus. Vet. lmmunol. Immunopathol., 26:17 l - 182. The course of infection in BALB/c mice of virulent Brucella abortus strain 2308 (S-2308) and attenuated strain 19 (S- 19 ) varies markedly. Whereas S- 19 is eliminated at an exponential rate beginning at 2 weeks post infection (p.i.), strain 2308 assumes a steady state or plateau during the first 6 weeks p.i. and thereafter is eliminated very slowly over a period exceeding 6 months. Here we compared the initiation and maintenance of inflammatory reactions in spleens and livers of mice infected with either of the two strains of B. abortus for the first 6 weeks p.i. Histological changes in the liver were similar in response to either strain and were characterized by the development of small granulomas and an influx of polymorphonuclear leukocytes (PMN) and monocytes. Tissue reactions in the spleen were similar at weeks 1 and 2 p.i. At 3 weeks p.i. and thereafter, focal granuiomatous responses in S-2308-infected mice exceeded those in mice infected with S-19. Numbers of nonspecific esterase (NSE) positive mononuclear leukocytes in S-19-infected spleens had increased by 3 weeks p.i. and remained elevated. No comparable increase in NSE positive cells occurred in mice infected with S-2308, and numbers were significantly lower. At 4 weeks p.i. the influx of mature neutrophils and the intensity of extramedullary hematopoiesis were significantly greater in S-19-infected spleens. A profound depletion of periarteriolar lymphoid tissue was noted in both infections for the first 3 weeks p.i. However, repopulation of lymphoid sheaths in S-19-infected spleens became significantly greater by 4 weeks p.i. and continued to increase at significantly higher levels during the next 2 weeks. This study demonstrates quantitative differences in splenic inflammatory responses which are temporally related to the more rapid elimination of S- 19. Based upon the lower susceptibility of strain 2308 to the protective effects of immune serum it is hypothesized that the different patterns of infection and inflammation displayed by the 2 strains may relate to the differential capacities of antibody opsonized S-19 and S-2308 to survive in activated macrophages.
0165-2427/90/$03.50
© 1990 - - Elsevier Science Publishers B.V.
172
F.M. ENRIGHT ET AL.
INTRODUCTION
Brucella abortus strain 2308 (S-2308) is representative of virulent field strains which produce abortion in cattle and cause chronic infections in the face of the host's immune response (Subcommittee on Brucellosis Research, 1977 ). Strain 19 (S- 19 ), on the other hand, is an attenuated strain which has for many years been used in cattle as a living vaccine (Subcommittee on Brucellosis Research, 1977). The relative virulence of these strains for cattle is paralleled by their patterns of infection in BALB/c mice, in that S-2308 persists in the spleen at high levels for several months whereas S-19 is cleared much more rapidly (Montaraz and Winter, 1986 ). The host-parasite interactions responsible for the capacity of virulent strains of B. abortus to produce persistent infections are not understood. The purpose of the work reported here was to perform a comparative histopathological study of BALB/ c mice infected with S-2308 and S-19. The studies encompassed the first 6 weeks post infection (p.i.) because the course of infection with the two strains varies markedly during this time (Montaraz and Winter, 1986 ). We hypothesized that the inflammatory reactions engendered in the spleen and liver by S-2308 and S-19 would have distinctive features which would bear upon the different kinetics of infection and could provide further insight into the basis of persistent infection by this facultative intracellular parasite. MATERIALS AND METHODS
Mice Nine-week-old BALB/c ByJ mice were obtained from the Jackson Laboratories (Bar Harbor, ME) and held for 1 week before use. Bacterial strains Stock cultures of B. abortus S- 19 were obtained from Dr. Dale Angus (National Veterinary Service Laboratories, Ames, IA) and from commercial vaccine vials (Professional Biological Co., Denver, CO). B. abortus S-2308 was passaged in BALB/c mice and isolated in pure culture from infected spleens. Stock cultures were prepared by growing strains for 48 h on Schaedler blood agar plates. Bacteria were then suspended in Brucella broth (BBL Microbiology Systems, Cockeysville, M D ) to an estimated concentration of 1 X 109 bacteria per ml, distributed in aliquots, and stored at - 7 0 ° C. Inocula were prepared from contents of freshly thawed vials diluted in sterile phosphate buffered saline (PBS) to an estimated concentration of 5 X 105 CFU per ml. The exact concentration was determined retrospectively by viable counts (Montaraz and Winter, 1986 ). Stocks of S- 19 and S-2308 used in these studies were designated 19.85.1, 19.85.2, 19.87.4, 2308.85.1, 2308.86.1, and 2308.87.2.
HISTOPATHOLOGY OF BALB/c MICE INFECTED WITH BRUCELLA ABORTUS
173
Experimental design Mice were infected intravenously (i.v.) with approximately 5 × 104 CFU of either S-I 9 or S-2308. Groups of mice were sacrificed at weekly intervals from I to 6 weeks p.i. for bacteriological and histological examination. Six such experiments, consisting of three repetitions for each strain, were performed over a period of 18 months. Experiments are designated according to the infecting stock strain as follows: experiment l, 19.85.1; experiment 2, 19.85.2; experiment 3, 19.87.4; experiment 4, 2308.85.1; experiment 5, 2308.86. l; and experiment 6, 2308.87.2. Group sizes were five except in experiments l and 4 in which the group size was three. At the time of sacrifice transverse sections of spleen ( 1.0-1.5 mm) were fixed in a cold solution of 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.5. The remainder of the spleen was homogenized in 10 ml PBS, serially diluted, and plated in triplicate (Montaraz and Winter, 1986 ). Colonies were counted after 3 days of incubation at 37 °C under 10% CO2. In experiments l and 4 the left lateral lobe of the liver was likewise sampled for histological examination and cultured. The fixed tissues were maintained at 4 oC. At the completion of each experiment tissues were coded and shipped refrigerated to a co-investigator (FME). The coded tissues were processed and examined as unknowns. Tissue processing and staining Fixed samples of tissues were placed into infiltrating solutions from one of two kinds of plastic embedding kits and held at 4°C for 4 to 7 days. Infiltrated samples were allowed to polymerize with plastic embedding media (DuPont Sorvall, Wilmington, DE; or LKB, Rockville, MD) at either 4 °C or at room temperature, depending on the manufacturer's instructions. Two or three adjacent sections, 2 to 3 #m thick, were prepared. All sections were stored at 4°C until stained. One section of spleen and liver was stained with hematoxylin and eosin (H&E) and an adjacent section of spleen was stained for nonspecific esterase (NSE) activity by the method of Koski et al. as modified by Jeffers et al. (1987). Sections were reacted with alpha-naphthyl butyrate (Sigma Chemical Co., St. Louis, MO) for 45 rain at 37°C. It was often necessary to examine NSE stained sections with either phase or differential interference microscopy to better visualize the morphological relationships of stained and unstained cell populations. Tissue reaction scores in the spleen The number of granulomas, degree of infiltration with mature polymorphonuclear leukocytes (PMN), degree of hematopoietic activity, and numbers and staining intensity of NSE positive mononuclear leukocytes were determined subjectively and scored on a scale of 0 to 3. Zero was equal to no reaction, 1 was a slight reaction, 2 a moderate reaction, and 3 was an exten-
174
F.M. E N R I G H T ET AL.
sive reaction. The depletion of normal periarteriolar lymphoid sheaths (PALS) resulted in sheath and follicular structures containing sheets of histiocytic cells and fibrous stroma. This change was scored on a scale of 4 to 0. Four represented the normal pattern of thick lymphoid sheaths with prominent follicles. Three was equal to a slight decrease in the size of both sheaths and follicles. Two was indicative of moderate depletion of lymphoid tissues. One was equal to severe depletion with only scattered lymphocytes remaining. Zero represented complete loss of all lymphoid elements. Infection with either strain of B. abortus resulted in focal granulomatous lesions. Granulomas consisted of moderate to large numbers of histiocytic cells mixed with smaller numbers of monocytes, lymphoid cells and granulocytic cells. Larger granulomas with necrotic centers were rarely noted. It was extremely difficult to differentiate between some histiocytic granulomas and the stromal tissue which remained following lymphoid depletion. The most consistent feature in differentiating these two reactions was the close spatial relationship of the few remaining lymphoid cells to the delicate fibrous stroma which delineated the original sheath or follicle. Granulomas were located randomly throughout the red pulp and had indistinct margins. One to 3 granulomas per section of spleen were scored from 0.1 to 1.0, depending on size. Four to 6 granulomas received scores of 1.1 to 2.0, and 7 to 10 granulomas were scored between 2.1 to 3.0. Granuloma scores were most subject to sampling variation due to the focal nature of the lesions within the spleen. Other changes were more generalized in distribution. Examination of sections from age and sex matched noninfected mice allowed the establishment of baseline reaction scores. Spleens of normal mice contained no granulomas (score 0), a modest number of mature P M N within the red pulp (score 0.5 ), a thin subcapsular zone of predominantly mononuclear hematopoietic activity (score 0.5 ), and a very thin zone of intensely NSE stained monocytes adjacent to the PALS (score 1.0). Positively stained foci of 2 to 3 cells were also present within the lymphoid tissue. The normal splenic lymphoid pattern consisted of thick cellular sheaths containing large lymphoid follicles (score 4.0). Tissue reaction scores in the liver The number and size ofgranulomas and the degree of infiltration with mature P M N or monocytes were scored on a scale of 0 to 3, as described above. Statistical analysis A log value of bacteria in each organ was obtained by averaging the triplicate counts following log conversion (Montaraz and Winter, 1986). Liver counts were multiplied by 3 prior to log conversion to adjust numbers for the entire organ. Statistical significance was determined by Fisher's protected least-significant-difference test (Snedecor and Cochran, 1989). Reaction
HISTOPATHOLOGYOF BALB/c MICE INFECTED WITH BRUCELLA ABORTUS
175
scores for individual spleens and mean scores of similarly treated mice were compared by analysis of variance using the general linear model SAS procedure (SAS, 1987). RESULTS
Comparison of growth curves between S-19 and S-2308 Splenic growth curves of S-19 in experiments 1 and 3 (Fig. 1A) were typical (Ho and Cheers, 1982; Montaraz and Winter, 1986) and did not differ significantly from each other at any week p.i. However, counts from experiment 2 were below those of experiments 1 and 3 at 1 week p.i. and were uniformly greater at weeks 3 through 6 p.i. (Fig. 1A). These differences were statistically significant at weeks 1 and 3 ( P < 0.01 ). When organisms isolated at 5 weeks p.i. from spleens of mice in experiment 2 were used to infect another group of mice, the growth curve was typical of S-19 (data not shown). Splenic growth curves of S-2308 in experiments 4, 5, and 6 (Fig. 1A) were typical (Montaraz and Winter, 1986) and were not significantly different from one another, with one exception at week 5 p.i. Growth curves of S-19 and S-2308 in the liver paralleled those in the spleen, but bacterial numbers were always lower by 1 to 2 logs (data not shown). Histological changes in the spleen Statistical comparisons were made of reaction scores of individual S- 19 experiments against individual S-2308 experiments as well as between pooled data from S-19 experiments 1 and 3 and S-2308 experiments 4, 5, and 6 (Fig. 1 ). Pooling of the data was justified because the same results were obtained from pooled as from individual comparisons. The aberrant S-19 infection (experiment 2 ) was analyzed separately (Fig. 1 ). At 1 week p.i. a significant difference between the groups was noted only in the degree of lymphoid depletion, which was moderate in the mice infected with S-2308 but only slight in those infected with S-19 (P
