National Animal Disease Center, Brucellosis Research Unit, US Department of Agriculture, Agriculture Research Service, Ames, IA

Vet Pathol 33:282-289 (1996) Morphometric and Histopathologic Analysis of Lymphoid Depletion in Murine Spleens Following Infection with Brucella abortus strains 2308 or RB51 or an htra Deletion Mutant M. V. PALMER, N. F. CHEVILLE, AND F. M. TATUM National Animal Disease Center, Brucellosis Research Unit, US Department of Agriculture, Agriculture Research Service, Ames, IA Abstract. BALBK mice were inoculated intraperitoneally with suspensions of Brucella abortus strains 2308 or RB5 1 or an htra mutant. Spleens were examined on postinoculation day (PID) 2, 4, 7, 10, 15, 21, 30, and 60. Brucellae were cultured in high numbers from the spleens of mice infected with strains 2308 or htra through PID 60; however, mice infected with strain RB51 cleared the infection between PID 30 and PID 60. Histopathologic changes in spleens from 2308-infected mice were characterized by marked accumulations of macrophages, which expanded marginal zones beginning as early as PID 7 and persisting through PID 60. Morphometnc analysis showed a decrease in splenic white pulp in 2308-infected mice at PID 10, which correlated with the peak of bacterial infection. Although this decrease was significant (P < 0.05) when compared with values at the previous (PID 7) and the following (PID 15) time periods, it was not significantly different from white pulp values noted at PID 2 or PID 4 or the values for control spleens. Spleens from RB5 1-infected mice showed only mild to moderate accumulations of macrophages in marginal zone areas during the peak of RB5 1 infection (PID 7-10). Morphometric analysis of RB5 1-infected spleens showed a decrease in white pulp area, which coincided with peak bacterial numbers. However, this decrease was not significant (P > 0.05). Spleens from mice infected with the htra mutant showed moderate to marked accumulations of macrophages in marginal zone areas, which persisted through PID 60. Multifocal necrosis in lymphoid follicles as early as PID 4 was seen in both htra and 2308 infection. Morphometric analysis of htra-infected spleens revealed no significant decrease in white pulp and no obvious correlation with bacterial numbers in the spleen. These results suggest that virulent B. abortus does not induce lymphoid depletion significantly below those values seen in noninfected mice; thus, the possible role of lymphoid depletion in the pathogenesis of brucellosis remains questionable. Key words: Brucella spp.; lymphoid depletion; mice; morphometry; spleen. Lymphoid depletion may play an important role in the pathogenesis of brucellosis. Lymphoid depletion has been described in T-cell-dependent areas of lymph nodes of cattle inoculated with virulent Brucella abortus2 and in mouse spleens during the first 3 weeks following infection with B. abortus S2308 or S 19.6 Such changes in lymphoid organs could delay host immune response by causing a transient immunodeficiency contributing to persistence of virulent B. abortus.* Different strains of Brucella may vary in their ability to induce lymphoid depletion and to colonize host tissue and produce disease. Therefore, potential vaccines should not deplete lymphoid tissue to the point of immunosuppression. The pathologic potential of various strains of Brucella and potential vaccines have been related to bacterial structural or genetic alteration^,^ e.g., composition of bacterial lipopolysaccharide (LPS). l4 A rough variant of B. abortus, RB5 1, which lacks much of the 0-side chain of the LPS, has been developed and in- 282 vestigated as an alternative vaccine to S19.I8 This strain has been shown to induce protection in micei8 and did not induce abortion in pregnant goats. Testing has also shown that RB5 1 does not induce production of antibodies to the 0-side chain of Brucella LPS and therefore does not cross-react with standard serologic tests.19 Genetic manipulation has also been used to produce mutants of B. abortus that are candidates for attenuated live vaccines and are also useful in pathogenesis studies. The gene encoding the htra stress protein has been deleted from the parental strain 2308 to create an htra deletion mutant.16%2 The htra protein in Brucella shows identity with the htra of Escherichia coli and Salmonella typhimurium. In E. coli, the htra protein is a heat-inducible periplasmic serine protease necessary for bacterial survival at temperatures above 42 C.*I Studies in mice to determine the survivability of the htra mutant have shown that it persists in a manner similar to that of the parental strain,*i although one report showed fewer bacteria in spleens from htra-

Vet Pathol 33:3, 1996 Murine Spleen Morphometry Following Infection with B. abortus 283 infected mice at 1 week postinfection than in 2308- infected mice.4 The pathology in murine spleens following htra infection, however, has not been detailed. Strains such as RB51 and the htra mutant must be tested in vivo for their capacity to produce disease and to nonspecifically suppress host immune responses. The purpose of this study was to morphometrically and histopathologically describe and compare murine splenic changes following infection with three different strains of Brucella: RB5 1, 2308, and the htra mutant. Smooth strain 2308 is considered the wild-type, virulent strain and has been used extensively as the challenge strain in numerous experimental studie~.~j Morphologic changes in mouse spleens following infection with some B. abortus strains have been described previously.6 Morphometric analysis of splenic compartments, including white pulp, marginal zone, and red pulp, can be used to detect depletion or expansion of splenic compartments through the course of infection. Inference from the relative changes in splenic compartments permits assessment of the immunologic significance and virulence potential of the LPS 0-side chain and the htra protein. Animals Materials and Methods One hundred thirty-two healthy male BALB/C mice were divided into four groups: mice infected with B. abortus RB5 1 (n = 42), mice infected with B. abortus 2308 (n = 42), mice infected with B. abortus htra mutant (n = 42), and salineinjected negative control mice (n = 6). Inoculation Brucella organisms were grown on tryptose agar at 37 C for 48 hours, suspended in saline (0.85% NaCl) solution, and diluted to the appropriate concentration. The 42 mice in each of the three treatment groups were injected intraperitoneally with a suspension containing 5 x lo7 colony-forming-units (CFU) B. abortus RB51 in 0.2 ml saline, 1 x lo4 CFU B. abortus S2308 in 0.2 ml sterile saline, and 5 x lo4 CFU B. abortus htra mutant in 0.2 ml saline, respectively. The six control mice were injected i.p. with 0.2 ml sterile saline. Inoculum doses of each strain were similar to those used in previous studies.20,2l Actual inoculum doses were determined retrospectively: 1.24 x 1 O7 CFW for RB5 1, 1.64 x 1 O4 CFU for 2308, and 1.84 x lo4 CFU for the htra mutant. Necropsy At postinfection day (PID) 2, 4, 7, 10, 15, 21, 30, and 60, five mice from each of the treatment groups were euthanatized by subcutaneous injection of ketamine (100 mg/kg) and acepromazine (2.5 mg/kg) followed by cervical dislocation. Spleens from three of these five mice were collected for quantitative bacterial analysis; the spleens from the other two mice were fixed in 10% neutral buffered formalin. Spleens from two saline-injected control mice were collected at PID 2 and 60. At each of these time periods, one spleen was collected for quantitative bacterial analysis and one spleen was fixed in 10% neutral buffered formalin. Bacteriologic analysis Spleens were weighed, triturated in saline solution, and serially diluted 1 0-fold with saline solution. Bacteria were grown at 37 C on tryptose agar with 5% bovine serum, cyclohexamide (30 ydml), bacitracin (7.5 U/ml), polymyxin B sulfate (1.8 U/ml), and ethylviolet (1 : 300,000). Bacterial cell counts were made from each dilution by standard plate counts. Brucella colonies were identified by colony morphology and growth characteristics. Light microscopy Tissues were fixed for light microscopy in 10% neutral buffered formalin and processed by routine paraffin-embedding techniques. Sections were cut at thicknesses of 4-6 ym and stained with hematoxylin and eosin (HE). Unstained sections were used for immunohistochemical staining of brucellar antigen using an avidin-biotin-peroxidase complex staining technique as previously describedi3 and a commercial staining kit (HistoMark Steptavidin-HRP, Kirkegaard & Perry Laboratories, Gaithersburg, MD). Control slides were stained with serum from clinically normal rabbits. Sections of Brucella-infected BALB/C mouse spleen and liver were used as control slides. Primary anti-brucella antibody was prepared by hyperimmunization of rabbits with whole cell irradiated Brucella strains 2308 or RB5 1. Anti-2308 antiserum was used as a primary antibody at a dilution of 1 : 10,000 to identify bacterial antigen in both 2308- and htrainfected spleens, and anti-rb51 antiserum was used at a dilution of 1 : 500 on RB5 1-infected spleens. Optimal dilutions for each primary antibody had previously been determined using sections of Brucella-infected BALB/C mouse liver and spleen. Morphometr y Prior to paraffin embedment, the distal third of the spleen was transected, and the remaining two thirds was divided longitudinally along the median plane, providing a cross section across the widest point ofthe spleen and two longitudinal sections for morphometric analysis of splenic compartment areas. HE-stained sections of the spleen samples were examined using a Zeiss Axioplan microscope (Carl Zeiss, Oberkochen, West Germany) with image input into a computerized image processing system (Kontron Elektronik, Image Analysis Division, Carl Zeiss Co., New York, NY) using Vidas 2.1 image analysis software (Carl Zeiss Co., New York, NY). Areas of splenic red pulp, white pulp, and marginal zone on each slide were examined for two mice at each time period from each Brucella treatment group. Uninfected control spleens from PID 2 and PID 60 were measured in the same fashion. The marginal zone was defined as the zone of macrophages and some B lymphocytes that surrounds the periarteriolar lymphoid sheath (PALS).lo Statistical analysis Student s t-test was used to evaluate differences between means. Differences were considered significant at P < 0.05.

284 Palmer. Cheville, and Tatum Vet Pathol 33:3, 1996 106 l o 7 I l %, -0- htra 105- g 104- u 103-102- 101 - - + RB51 2308 loo I I I I I I 0 10 20 30 40 50 60 PID Fig. 1. Kinetics of splenic infection with B. abortus S2308, RB5 1, or htra expressed in CFu/g spleen. Each point represents the mean CFU/g from three mice from each strain at each time point. Standard errors are shown. Culture Results Numbers of live bacteria reached peak levels in spleens at PID 7-10 in mice infected with all three strains and then began to decline. RB5 1 levels peaked 1-2 log values lower than did htra or 2308 levels and then began to decline at a steady rate, being cleared from the spleen between PID 30 and PID 60. In contrast, after a brief decline in CFU/g spleen, 2308 and htra counts began to increase and by PID 30 were as high as those recorded at PID 10. Strain 2308 and htra CFU/g spleen remained high through the termination of the experiment at PID 60 (Fig. 1). Histopathology Evidence of splenic change in RB5 1 -infected mice was first seen at PID 7 and was characterized by multifocal accumulations of macrophages and mature segmented neutrophils located adjacent to the PALS in the region of the marginal zone and in solitary foci within the red pulp (Fig. 2, 3). By PID 15, spleens from RB5 1 -infected mice lacked discrete foci of inflammatory cells and showed only a mild expansion of the marginal zone by large macrophages interspersed between lymphocytes and occasional plasma cells. At PID 21, the spleens appeared similar to those at PID 15 with the exception of prominent multifocal germinal centers composed of lymphoblasts, mitotic figures, and numerous intracellular residual bodies. Diffuse splenic congestion and moderate numbers of megakaryocytes (3-4/high power field [HPF]) were seen in the red pulp of spleens examined at PID 30. Mild marginal zone expansion persisted through PID 30. By PID 60, spleens from RB51-infected mice did not differ significantly from those spleens obtained from saline-injected mice. As early as PID 4, spleens from mice infected with 2308 showed multifocal, well-delineated areas of lymphoid necrosis not seen in RB5 1-infected mice. These areas ranged in size from 50 to 200 ym and were found within PALS, often adjacent to marginal zones. They were further characterized by pyknotic and karyorrhectic nuclei among accumulations of large macrophages with abundant eosinophilic cytoplasm and lesser numbers of mature segmented neutrophils. Moderate numbers of tingible body macrophages containing nuclear debris were also seen. By PID 7, multifocal accumulations of macrophages and lesser numbers of neutrophils were present adjacent to PALS. These accumulations were associated with marginal zone expansion up to 50 cell layers thick. Inflammatory cell accumulations were marked in spleens from 2308-infected mice sampled at PID 15. Multifocal to bridging accumulations of macrophages and neutrophils were present. Marginal zones remained markedly expanded (Fig. 4). The marked inflammatory cell presence persisted in 2308-infected mice through PID 60. Moreover, focal abscesses were occasionally seen in spleens from PID 30 and PID 60. These areas were characterized by numerous degenerate neutrophils and pyknotic and karyorrhectic nuclei among cellular debris surrounded by lightly eosinophilic extracellular material. The red pulp from 2308- infected mice showed increased extramedullary hematopoiesis, with megakaryocytes present (5-1OIHPF) at PID 60. Spleens from mice infected with the htra mutant exhibited signs of lymphoid necrosis similar to those seen in 2308-infected mice at PID 4 (Fig. 5). However, at PID 7 minimal changes were noted, with only mild marginal zone expansion up to 10 cell layers thick. From PID 10 through PID 60, htra-infected mice showed moderate multifocal accumulations of macrophages and neutrophils within marginal zones; however, the marked bridging accumulations of inflammatory cells seen in 2308-infected mice were not seen. Immunostaining Immunoperoxidase staining did not reveal Brucella antigen until PID 4-7 in any of the Brucella-infected mice. In RB5 1 -infected mice, only spleens collected at PID 7 showed positive staining. In contrast, spleens from both 2308- and htra-infected mice showed positive staining from PID 4 through the conclusion of

Vet Pathol 33:3, 1996 Murine Spleen Morphometry Following Infection with B. abortus 285 Fig. 2. Spleen; RB5 1-infected mouse, PID 7. Note multifocal accumulations of inflammatory cells that surround PALS (between arrows) and expand marginal zones. HE. Bar = 300 pm. Fig. 3. Spleen, RB51-infected mouse, PID 7. Higher magnification of Fig. 2. Inflammatory infiltrate is composed of numerous large macrophages and fewer neutrophils. HE. Bar = 50 pm. Fig. 4. Spleen; S2308-infected mouse, PID 15. Note marked multifocal-bridging accumulations of macrophages that surround PALS and expand marginal zones (arrows). Extensive macrophagic accumulations persisted in 2308- and htrainfected mice through PID 60. HE. Bar = 300 Fm. Fig. 5. Spleen; htra-infected mouse, PID 4. Note focal area of necrosis within PALS characterized by pyknotic and karyorrhectic nuclei among numerous macrophages and lesser numbers of neutrophils. Also note tingible body macrophages containing nuclear cell debris. HE. Bar = 50 pm.

286 Palmer, Cheville, and Tatum Vet Pathol 33:3, 1996 I RB51 A"- 2308 Fig. 6. Spleen; S2308-infected mouse. Note positive immunoperoxidase staining for brucellar antigen within macrophage cytoplasm (arrows) and occasionally within neutrophils and stellate or spindloid cells. Gills hematoxylin counterstain. Bar = 50 Fm. the study at PID 60. Positive staining for all strains was seen within the cytoplasm of macrophages, occasionally within the cytoplasm of neutrophils, and to a lesser degree in spindloid or stellate cells. All cell types were located within marginal zones and areas of inflammatory cell accumulation surrounding PALS (Fig. 6). Additionally, areas of lymphoid necrosis within PALS seen at PID 4 in both 2308- and htra-infected tissues showed positive staining for Brucella antigen within macrophage cytoplasm. Morphometr y RBSI-infected mice. RB5 1 -infected spleens showed initial increases in areas of white pulp, red pulp, and marginal zones, followed by decreases in all compartments at PID 7-1 5; however, these decreases were not significant (P > 0.05, t-test; Figs. 7, 8; data for red pulp not shown). Following a transient increase in white pulp area at PID 21, all compartments in RB51-infected spleens remained in ranges consistent with normal spleen values between PID 30 and PID 60. Decreases in white pulp in RB5 1 -infected spleens, seen at PID 10-15, occurred at or shortly after the initial peak in bacterial numbers were reached. However, this decrease was not significant (P > 0.05). Peak bacterial numbers were also followed by a significant increase in the size of marginal zones as compared with that seen at the previous and following time periods. 2308 RB51 2 n htra k c 107 88 gb 3 -i E + 106.I Dam L- lo5 2 4 7 10 15 21 30 60 PID Fig. 7. Comparative white pulp changes in mouse spleen following infection with B. abortus S2308, RB51, or htra mutant. Uninfected control spleens measured at PID 2 and PID 60 had white pulp areas of 5.3 x lo6 and 7.5 x lo6 pm2, respectively. Standard errors are shown. Each bar represents the mean of measurements recorded from longitudinal and cross sections of spleens from two mice. Fig. 8. Comparative marginal zone changes in mouse spleen following infection with B. abortus S2308, RB51, or htra mutant. Uninfected control spleens measured at PID 2 and PID 60 had marginal zone areas of 6.8 x lo5 and 5.9 x lo5 Fm2, respectively. Standard errors are shown. Each bar represents the mean of measurements recorded from longitudinal and cross sections of spleens from two mice. 2308-infected mice. In 2308-infected spleens, white pulp values initially increased twofold from PID 2 to PID 7; however, by PID 10 these values had decreased significantly (P < 0.05) from the value recorded at PID 7. This transient decrease was followed by a significant increase in white pulp values by PID 15. Following PID 15, white pulp values decreased slightly at PID 2 1; however, this decrease was not significant. White pulp values remained elevated from PID 30 through the termination of the experiment at PID 60 (Fig. 7). Marginal zone measurements revealed a steady increase beginning at PID 10, remaining elevated through PID 60, and ending with a 63-fold increase over marginal zone values at PID 2 (Fig. 8). A threefold increase

Vet Pathol 33:3, 1996 Murine Spleen Morphometry Following Infection with B. abortus 287 in the area of the red pulp compartment was also measured over the course of the experiment from PID 2 to PID 60. However, a decrease in red pulp area was seen at PID 10 which coincided with a significant decrease in white pulp (data not shown). This decrease in red pulp was not statistically significant (P > 0.05). Decreased white pulp values at PID 10 coincided with the initial peak in bacterial numbers. A decrease in bacterial numbers from PID 10 to PID 20 was coincident with a significant rise in white pulp area and marginal zone size. htra-infected mice. Measurements of white pulp from htra-infected mice revealed increases in the early phases of infection that persisted with minimal decreases (Fig. 7). At the termination of the experiment, htra-infected spleens showed a threefold increase in white pulp area compared with areas measured at PID 2. The largest absolute areas of white pulp were measured in htra-infected spleens. This difference among strains, however, was not significant (P < 0.05). Similar to 2308-infected spleens, the largest increase in total spleen area (from PID 2 to PID 60) came from increased marginal zone size; however, this increase was only eightfold, compared with the 63-fold increase seen with 2308 infection. Significant rises or declines in white pulp or marginal zones that corresponded to changes in bacterial numbers were not obvious in htra-infected spleens. However, as bacterial numbers increased from PID 15 to PID 60, there was also a trend toward increasing areas of both the white pulp and marginal zone. Discussion Lymphoid depletion has not been described previously in RB5 1 -infected mice. The measurable, but not statistically significant, decrease in white pulp seen at PID 10-1 5 was not visible as a histopathologic change in this experiment. RB51 may cause mild lymphoid depletion that is not discernible without the application of morphometric techniques. Rapid clearance of RB5 1 infection and the return to an apparently normal spleen are consistent with previous findings in mice6j8,20 and in bovine lymph nodes.2 The early increases in white pulp and marginal zones suggest that there is significant splenic response to RB5 1 in mice. This finding is important for any potential vaccine because it is morphologic evidence of an immune response. However, any further conclusions concerning immune response to RB5l can only be reached by monitoring antibody and cell-mediated responses and ultimately demonstrating protection against challenge with a virulent strain. These factors have been investigated el~ewhere~.~j~-~~ and are the subject of ongoing research. Although the largest decrease in white pulp area was seen in 2308-infected mice at PID 10, coincident with the peak of infection, this decrease is not significantly different from values noted at PID 2 or PID 4 or from values recorded in uninfected mouse spleens. The decreases in white pulp that occurred coincident with peak bacterial numbers in both 2308- and RB5 1 -infected spleens suggest some cause-and-effect relationship. There may be some factor(s) elaborated by certain strains of Brucella that causes lymphocyte death or decreased lymphocyte proliferation. Our findings suggest that this decrease in white pulp probably is not enough to result in immunosuppression. However, no specific tests of immune function were performed. More precise conclusions concerning the effect on immune response could be drawn by immunostaining for specific lymphocyte subsets. This research is ongoing in our lab. In studies of human brucellosis patients diagnosed by hemoculture, a decrease in the percentage of circulating CD4-t lymphocytes and an increase in the percentage of CD8-t lymphocytes with an inverted CD4+/CD8+ ratio was observed in the early stages of infection. These percentages normalized slowly over the course of the observation peri~d.~ Greater expansion of the marginal zone in 2308 infection as compared with RB51 infection may reflect an increased immunogenicity due to the complete LPS of strain 2308. The decreased level of this expansion in infection with the htra mutant suggests an important immunogenic role for the htra protein that is absent in the htra mutant. Expansion of the marginal zone is to be expected with any hematogenous immunogen. The marginal zone is where incoming blood first leaves the arteriole and encounters antigen-presenting cells of the mononuclear phagocyte system and the dendritic cell family. Both infiltration of circulating inflammatory cells and local proliferation of resident macrophages may contribute to marginal zone expansion. Increases in numbers of macrophages within marginal zones and expansion of red and white pulp con- tribute to splenomegaly during the first 60 days postinfection. Only in 2308 infection was the marginal zone the largest compartment in spleens at any time period. In RB51 and htra infection, red pulp made up the majority of total spleen area. However, proliferating cellular populations within the red pulp were not specifically examined. Increases in red pulp have been described for various systemic diseases in laboratory animals resulting in ~plenomegaly.~~ Splenomegaly has also been described in B. abortus S19-infected mice due to massive numbers of macrophages.' Culture results suggest that the htra mutant replicates in a manner similar to that of the parental strain 2308 and is able to persist in the mouse spleen, as has

288 Palmer, Cheville, and Tatum Vet Pathol 33:3, 1996 been found previously.21 The commonalities in histopathologic findings also suggest similarity to the parental strain. Although the S. typhimurium htra mutant shows a significant increase in sensitivity to killing by H202 and 0,- in vitro and decreased vir~lence,~ a similar decreased virulence does not appear to occur in vivo with the Brucellu htra mutant. Early lymphocyte necrosis in both 2308- and htra-infected mice is also significant. However, lymphocyte necrosis without measurable lymphoid depletion in htra-infected spleens is curious. Differences in the severity of necrosis were not evident histologically. Furthermore, differences in the ability of lymphocytes to proliferate and replenish dying populations may differ in 2308 infection and htra infection. Lymphoid necrosis may not be the only mechanism of lymphoid depletion in 2308- infected spleens. Lymphoid tissues undergo accelerated programed cell death in response to several infectious agents. Specific bacterial pathogens possess certain virulence factors described as superantigens, such as the staphylococcal enterotoxin B (SEB). Lymph nodes draining the site of SEB injection in mice undergo up to 50% deletion of specific T cells.22 The mechanism of deletion of these T cells is thought to be apoptosis (programed cell death). Whether Brucellu 2308 LPS or some other unknown factor acts in this manner as a superantigen is unknown. Immunostaining results suggest a threshold level of CFU/g below which results are unreliable, possibly resulting in false-negative findings. Immunostaining for Brucellu antigen was negative prior to PID 4-7 in all strain groups, although bacteria were cultured from these spleens in relatively high numbers. Immunostaining for RB5 1 was positive only when bacterial CFU/g was highest, which suggests a threshold level ofapproximately lo5 CFU/g. Therefore, in any attempt to identify Brucellu antigen in tissue, bacterial culture should accompany immunostaining. Although our results suggest that it is doubtful that lymphoid depletion in 2308-infected mice is of such severity as to cause significant immunosuppression, the relatively mild splenic changes observed in spleens from mice infected with RB5 1 may avoid any transient inhibitory or deleterious affects that may occur with strain 2308 or related strains that are able to cause lymphoid necrosis. These mild changes combined with the lack of persistence seen with RB51 increase the attractiveness of RB51 as a potential candidate vaccine. Caution should be used, however, in drawing conclusions on bovine brucellosis from studies performed in mice, especially those that involve lymphocyte function. The bovine immune system differs from that of the mouse in being characterized by a relatively high number (up to 60 /~)12 of CD4- CD8- T cells that express the y/6 T cell receptor.8 Although one of the proposed functions of y/6 T cells is the protection of epithelial surfaces, they represent a high percentage of circulating T cells.*j1j2 In the ruminant spleen, y/6 T cells localize to regions of cellular traffic, namely the marginal zones and red pulp. y/6 T cells occur in only very small numbers in the conventional T cell domains of spleen.ilj2 In lymph nodes, they are located chiefly in the medullary cords and the subcapsular and par- atrabecular region of the cortex.11j2 How this significant difference between cattle and mice would affect splenic pathology and immune response is important to consider but is as yet unclear. Acknowledgements We thank T. Krausman, R. Capsel, A. Duit, A. Jensen, and R. Hornsby for their technical assistance. No endorsements are herein implied. Brand names are necessary to report factually on available data; however, the USDA neither guarantees nor warrants the standards of the products, and the use of the names by the USDA implies no approval of the products to the exclusion of others that may also be available. References 1 Cheers C, Young AM: Serum colony stimulating activity and colony forming cells in murine brucellosis: relationship to immunopathology. Microb Pathog 3: 185-1 94, 1987 2 Cheville NF, Jensen AJ, Halling SM, Tatum FM, Morfitt DC, Hennager SG, Frerichs WM, Schurig G: Bacterial survival, lymph node changes, and immunologic responses of cattle vaccinated with standard and mutant strains of Brucella abortus. Am J Vet Res 53: 1881-1888, 1992 3 Cheville NF, Stevens MG, Jensen AE, Tatum FM, Halling SM: Immune response and protection against infection and abortion in cattle experimentally vaccinated with mutant strains of Brucella abortus. Am J Vet Res 54~1591-159, 1993 4 Elzer PH, Phillips RW, Kovach ME, Peterson KM, Roop RM: Characterization and genetic complementation of a Brucella abortus high temperature-requirement A (htra) deletion mutant. Infect Immun 62:4 1354139, 1994 5 Enright FE: The pathogenesis and pathobiology of Brucella infection in domestic animals. In: Animal Brucellosis, ed. Nielsen K. and Duncan JR, pp. 301-320. CRC Press, Boca Raton, FL, 1990 6 Enright FE, Araya LN, Elzer PH, Rome GE, Winter AJ: Comparative histopathology in BALB/c mice infected with virulent and attenuated strains of Brucella abortus. Vet Immunol Immunopathol26: 171-1 82, 1990 7 Gazapo E, Gonzalez-Lahoz J, Subiza JL, Baquero M, Figueredo MA, Gomez-de-la-Concha E: Subpoblaciones linfocitarias en la evolucion de la brucelosis. Rev Clin Esp 186:369-373, 1990 8 Hein WR, Mackay CR: Prominence of y6 T cells in the

Vet Pathol 33:3, 1996 Murine Spleen Morphometry Following Infection with B. abortus 289 ruminant immune system. Immunol Today 12:30-34, 1991 9 Johnson K, Charles I, Dougan G, Pickard D, O Gaora P, Costa G, Ali T, Miller I, Hormaeche C: The role of a stress-response protein in Salmonella typhimurium virulence. Mol Microbiol5:401-407, 199 1 10 Klein J: Immunology, pp. 63-64. Blackwell Scientific, Cambridge, MA, 1990 11 Mackay CR, Beya MF, Matzinger P: r/6 T cells express a unique surface molecule appearing late during thymic development. Eur J Immunoll9: 1477-1483, 1989 12 McClure SJ, Hein WR, Yamaguchi K, Dudler L, Beya MF, Miyasaka M: Ontogeny, morphology and tissue distribution of a unique subset of CD4-CD8-sheep T lymphocytes. Immunol Cell Biol 622 15-22 1, 1989 13 Meador VP, Tabatabai LB, Hagemoser WA, Deyoe BL Identification of Brucella abortus in formalin-fixed paraffin-embedded tissues of cows, goats, and mice with an avidin-biotin-peroxidase complex immunoenzymatic staining technique. Am J Vet Res 47:2147-2150, 1986 14 Moreno E, Berman DT, Boettcher LA: Biological activities of Brucella abortus lipopolysaccharides. Infect Immun 31:362-370, 1981 15 Percy DH, Barthold SW: Pathology of Laboratory Rodents and Rabbits, p.6. Iowa State University Press, Ames, IA, 1993 16 Roop RM, Fletcher TW, Sriranganathan NM, Boyle SM, Schurig GG: Identification of an immunoreactive Brucella abortus htra stress response protein homolog. Infect Immun 62:lOOO-1007, 1994 17 Roop RM, Jeffers G, Bagchi T, Walker J, Enright FM, Schurig GG: Experimental infection of goat fetuses in utero with a stable, rough mutant of Brucella abortus. Res Vet Sci 51:123-127, 1991 18 Schurig GG, Roop RM, Bagchi T, Boyle S, Buhrman D, Sriranganathan N: Biological properties of RB5 1; a stable rough strain of Brucella abortus. Vet Microbiol 28: 171-188, 1991 19 Stevens MG, Hennager SG, Olsen SC, Cheville NF: Serologic responses in diagnostic tests for brucellosis in cattle vaccinated with Brucella abortus strain 19 or RB5 1. J Clin Microbiol32:1065-1066, 1994 20 Stevens MG, Olsen SC, Pugh GW, Palmer MV: Immune and pathologic responses in mice infected with Brucella abortus strain 19, RB5 1 or 2308. Infect Immun 62:3206-3212, 1994 21 Tatum FM, Cheville NF, Morfitt D: Cloning, characterization and construction of htra and htra-like mutants of Brucella abortus and their survival in BALBK mice. Microb Pathog 17:23-26, 1994 22 Wahl C, Miethke T, Heeg K, Wagner H: Clonal deletion as direct consequence of an in vivo T cell response to bacterial superantigen. Eur J Immunol 23: 1197-2000, 1993 Request reprints from Dr. M. V. Palmer, National Animal Disease Center, USDA, ARS, 2300 Dayton Avenue, Ames, IA 50010 (USA).