Vet. Pathol. 26:369-375 (1989)

Effect of Nursing on Brucella abortus Infection of Mammary Glands of Goats V. P. MEADOR, B. L. DEYOE, AND N. F. CHEVILLE US Department of Agriculture, Agricultural Research Service, National Animal Disease Center, Ames, IA Abstract. Eight 1-year-old, goats were inoculated intravenously with Brucella abortus (B. abortus) on the day of parturition and necropsied at 28 days after inoculation. Four nursed their kids and four did not (milk was not removed from the udders). Tissues and fluids were examined by bacterial isolation, light microscopy, and serologic methods. Nonnursing goats had high titers of brucellae (I 1On organisms/ml) in milk (brucellae were isolated from four of four udders), had marked enlargement of supramammary lymph nodes, and had lymphoplasmacytic and histiocytic interstitial mastitis. Immunoperoxidasestainingrevealed that brucellae were primarily in macrophages and neutrophils of the mammary alveolar and ductal lumens and in macrophages of the subcapsular sinuses of the supramammary lymph node. In contrast, nursing goats excreted brucellae intermittently at low concentrations (< lo3 organisms/ml) in milk; brucellae were isolated at necropsy from one of four udders; supramammary lymph nodes were not enlarged; and mammary lesions were not seen. Brucellae were detected in more tissues other than the udder, and serum anti-Brucella antibody titers were higher in nonnursing goats than in nursing goats. The present study indicates that the failure to nurse or release milk enhances localization and replication of B. abortus in mammary glands of goats after parturition, and that mammary gland infection may result in increased systemic spread and persistence of brucellae in the host.

Brucella abortus, a Gram-negative, facultative-intracellular bacterium, has a marked affinity for mammary glands and supramammary lymph nodes in persistently infected cows.4,7,11,12,28 M a m m a r y gland infection is typically clinically inapparent, and gross lesions are not detected. Microscopically, there is a lobular and periductal, lymphoplasmacytic and histiocytic interstitial mastitis with leukocytes in alveoli and duct^.^,^,^^ Lesions are not seen in all infected mammae In experimentally inof both cows and fected goats, brucellae localize and replicate primarily in macrophages and neutrophils in mammary ductal and alveolar lumens; fewer brucellae are free in mammary secretions or in phagocytes in the interstitium.1° Supramammary lymph nodes may be enlarged and are characterized by lymphofollicular hyperplasia, medullary plasmacytosis, and sinus h i s t i o c y t ~ s i s . ~ ~ J ~ Susceptibility of the mammae to become infected with B. abortus may vary with the stage of lactation. Excretion of brucellae from the mammary gland can occur throughout lactation but is most frequent in colostrum and during late 1 a ~ t a t i o n . Enhanced l~ excretion may be caused by reduced milk flow at these stages, either due to concentration of brucellae in smaller quantities of milk or due to increased replication of brucellae in mammary glands. Lactating mammae are less susceptible to bacterial pathogens that invade 369

through the teat canal, apparently due to physical removal of pathogens with the milk.20,27Brucellae are disseminated hematogenously to mammary glands; however, because organisms localize primarily in phagocytes in mammary secretions, milk expulsion may reduce the severity of infection by flushing organisms from mammary glands. This study examines the effect of nursing on B. abortus infection of the ruminant mammary gland. Goats were used because they are an excellent model for studying bovine brucellosis, and mammary gland changes from B. abortus infection are similar in both goats and cows.2,3,5J5-17

Materials and Methods Eight 1-year-old, mixed-breed goats were infected with Brucella abortus within 16 hours after parturition. Goats were given 1.0 ml of B. abortus strain 2308 suspended in 0.85% NaCl soluton in the jugular vein, such that the dose was 5.0 x lo9 colony-forming units (CFU) B. abortuslgoat. Four 1-year-old goats served as noninfected controls. Kids were removed from four infected and two noninfected goats (nonnursinggoats) on the day of parturition. The remaining goats, two noninfected and four infected nursing goats, were allowed to nurse their kids throughout the experiment. Mammary secretions were not removed from mammary glands of nonnursing goats throughout the experiment, except dur-

Meador, Deyoe, and Cheville

370

12800 .

6400

3200 I600 cc w

800

d

3 400.

a

k2

200-

I-

& 100 50. 25, N0NDECTA BLE

I

I

I

I

I

2

3

4

WEEKS AFTER INFECTION

1

‘2800

ing weekly sample collection. Goats were housed in isolation from 2 weeks before parturition until necropsy. Serum and milk (< 10 ml/mammary gland) were collected from each goat on the day of parturition and at I-day intervals until necropsy. Goats were killed with intravenous sodium pentobarbital 28 days after infection. Supramammary lymph nodes were trimmed of excess tissue and weighed. Milk, mammary gland, spleen, bone marrow (from proximal femur), and supramammary, medial iliac, and medial retropharyngeal lymph nodes were collected for bacterial culture. Tissues from mammary glands, supramammary lymph nodes, kidney, liver, lung, spleen, and uterus were collected for examination by light microscopy. Serum samples were tested for anti-Brucella antibody by the card test, standard tube agglutination test, and mercaptoethanol agglutination test.z6 The milk ring test was performed on milk samples that did not solidify after collection. Fluids and tissues were cultured for B. abortus.16The number of brucellae in milk and supramammary lymph nodes were determined quantitatively, but qualitative evaluations were completed for the presence of brucellae in mammary glands, spleen, uterus, bone marrow, and medial retropharyngeal and medial iliac lymph nodes. Tissues for light microscopic examination were fixed in 10% neutral buffered formalin. Tissues were processed by routine paraffin embedment, sectioned at 3-4 microns, and stained with hematoxylin and e o ~ i nTissue . ~ ~ sections were stained for brucellar antigen (mammary gland and supramammary lymph node) and keratin (mammary gland) with avidin-biotin-peroxidase immunohistochemical staining techniques. I 6 , l 8 Statistical analysis, using Student’s t-test, was done on serologic data.

Results All goats delivered live, full-term kids. Each of the four infected, nursing goats had twins. The two noninfected, nursing goats had a set of twins and a single kid. All goats remained clinically normal throughout this study. Serology

Before infection, anti-Brucella antibody titers were not detected or were at low levels; the card test and mercaptoethanol agglutination (MET) titers were not detected in 12 of 12 goats; standard tube agglutination test (STT) titers were I 1 : 25 in 11 of 12 goats; and the milk ring test (MRT) titers were 5 1 : 10 in 11 of 12 goats. Goat No. 7 (nonnursing, infected group) had elevated titers of STT = 1 :200 and MRT = 1 :40 before infection. Titers before infection were considI I I I ered nonspecific and did not indicate Brucella infecI 2 3 4 tion. In control goats, serologic titers remained at the WEEKS AFTER INFECTION same levels throughout the study. Figs. 1, 2. Serologic anti-Brucella antibody titers, infectBlood anti-Brucella antibodies, as detected by all ed, nursing and nonnursing goats. Bar = standard error. Fig. 1. Mean reciprocal standard tube agglutination test serologic tests, were elevated in all infected goats at 1 week after infection and remained elevated throughout titers (STT). Fig. 2. Mean reciprocal mercaptoethanol agglutination this study. The STT and MET titers were consistently test titers (MET).

B. abortus Infection of Goats

NONNURSING NURSING

T

T

T

37 1

infection at approximately 1O8 colony-forming units (CFU)/ml milk. Brucellae were isolated intermittently from milk of nursing, infected goats. Four milk samples collected from two goats contained brucellae; concentrations ranged from 20 to 570 CFU/ml milk. Brucellae were not detectable at levels < 10 CFU/ml milk. At necropsy, brucellae were isolated from supramammary, medial retropharyngeal, and medial iliac lymph nodes, the spleen, and the uterus from every inoculated goat. Mammary glands and bone marrow from all nonnursing goats contained brucellae, while in nursing goats, brucellae were isolated from mammary glands from one goat and were not isolated from bone marrow. Supramammary lymph nodes of nonnursing goats contained higher titers of B. abortus than supramammary lymph nodes of nursing goats (Table 1). Necropsy examination

I

-.-. -. Vio 1

2

3

4

WEEKS AFTER INFECTION ...... ......, ,.,--, .,.

Mean I."."'mlnnv-fnrmino iinitc lrFTT) nf -. R nhnrtrir ."&

isolated from milk of infected, nursing and nonnursing goats. Bar = range of values. Concentrations of brucellae < 10 CFU/ ml milk were not detectable.

higher in nonnursing, infected goats than in nursing, infected goats (Figs. 1,2). The STT titers between these two groups were significantly ( P < 0.05) different for weeks 1 through 4,while MET titers were not significantly ( P > 0.05) different. The MRT titers remained at low levels (< 1 :80) but differed slightly between nursing and nonnursing, infected goats; titers in nursing goats were highest at week 1, while titers in nonnursing goats were highest at weeks 2 and 4.

Gross lesions were not found in mammary glands of Brucella-inoculated goats. Mammae of nonnursing goats were firmer than those of nursing goats, but there were no palpable differences between udders of infected and control goats of the same nursing status. Supramammary and medial iliac lymph nodes were markedly enlarged in nonnursing, infected goats (Table 1). Medial retropharyngeal lymph nodes of five goats contained large abscesses, which were attributed to Actinomyces pyogenes infection based on gross features and previous history of infection in the parent herd. Gross lesions were not detected in other organs. Microscopic examination

Lesions were seen in mammary glands from all nonnursing, infected goats. There were marked, diffuse, lymphoplasmacytic and histiocytic infiltrates, dense lymphocytic foci, and individual and small foci of neutrophils and eosinophils in intralobular septa and periductal stroma (Fig. 4). Many ducts and alveoli were filled with macrophages, fewer neutrophils and lymBrucella isolation phocytes, cell debris, and rare, sloughed, epithelial cells. Brucellae were isolated from all milk samples col- Duct epithelium was multifocally hyperplastic and had lected from nonnursing, infected goats (Fig. 3). Con- numerous intraepithelial leukocytes that often formed centrations of brucellae progressively increased from small nests. There were few small erosions and rare weeks 1 through 3, and peaked at 3 to 4 weeks after sites of squamous metaplasia of duct epithelium. Most Table 1. Supramammary lymph nodes of Brucellu-infected goats: titers of Brucella abortus isolated and lymph node weights. Group Status

Nursing Nonnursing Nursing Nonnursing

Infected Infected Noninfected Noninfected

Supramammary Lymph Node Weight: Mean (range)

4.7 g (3.5 to 18.7 g (14.0 to 4.7 g (4.0 to 7.1 g (5.5 to

6.4) 27.2) 5.3) 8.7)

B. abortuslg Supramammary Lymph Nodes Mean (range)

5.0 x lO'(8.9 x 10'to 1.6 x lo4) 9.1 x 105(6.7 x 1 0 4 to 2.9 x 106) Not applicable (control goats) Not applicable (control goats)

B. abortus Infection of Goats

373

remaining lymph node parenchyma. Subcapsular granulomatous foci contained abundant brucellar antigen (Fig. 10). Rare foci of brucellar antigen were in germinal centers and paracortices.

Discussion The enhanced susceptibility of nonnursing mammae to Brucella abortus infection likely is due to stasis of mammary secretions in alveoli and ducts. Since brucellae localize and replicate primarily in macrophages and neutrophils in alveolar and ductal lumens and do not attach to or localize in epithelial cells, milk expulsion should physically remove brucellae from the mammae, thus reducing the chances of udder infection.16 Brucellae entering mammae of nursing goats in this study were probably excreted in milk before significant replication could occur. Conversely, cessation of milk flow would allow organisms to persist in an environment compatible with replication. Because brucellae were not detected in mammae and mammary secretions from all nursing goats, it may be that fewer organisms were transported into mammary glands of nursing goats as compared with those of nonnursing goats. Migration of infected phagocytes from the blood stream into mammary parenchyma and secretions is the primary route by which brucellae enter the mammae.16,22 In cattle, there is a decline in the number of leukocytes in milk at the onset of lactation and an increase when lactation ceases.I3J4If similar changes in leukocyte numbers occur in goats, nursing goats in this study may have had fewer infected phagocytes migrate into the mammae as compared with nonnursing goats, thereby partially accounting for the low numbers of brucellae isolated from milk. Additionally, dilution of brucellae due to continued milk production and excretion in nursing goats could also reduce detection of brucellae in mammary secretions. In the nonnursing goats in this study, leukocyte numbers in mammary secretions may be comparable with those in cows at the cessation of lactation, because mammary glands were developed at parturition and then underwent involution. The increased susceptibility of mammary glands of nonnursing goats to become infected with B. abortus is probably not attributable to decreased competency of immune defense mechanisms. Generally, cows are most susceptible to development of mastitis in early lactation.21Leukocyte functions are impaired in the lactating dairy cow just after parturition, which suggests that a compromised immune function increases susceptibility to mastiti~.~JO If lactation-associated immunosuppression is a major contributing factor in determining susceptibility of mammae to B. abortus infection, mammae of nursing goats in this study would have been expected to be susceptible to infection. Infected mammary glands can serve as a nidus from

Fig. 6. Supramammary lymph nodes from both infected, nursing (small lymph node) and nonnursing goats. Sections were obtained from the lymph node center near the hilus. Lymph node from the nonnursing goat contains numerous germinal centers in the outer cortex, has expanded paracortices, and has prominent medullary cords. Lymph node from the nursing goat is comparatively smaller, and has a relatively thin cortex, few small germinal centers, and sparsely populated medulla. HE. Bar = 3 mm.

which brucellae disseminate to other tissues in the host. The tissues in the goats ofthis study were likely infected both by hematogenous distribution of brucellae following intravenous inoculation and by dissemination of brucellae from mammary glands after establishment of mammary infection. Because supramammary lymph nodes of goats with severely infected mammae contained markedly higher concentrations of brucellae than those draining relatively noninfected mammae, it appears that brucellae were transported from mammary glands to supramammary lymph nodes through lymphatics. Localization of brucellae primarily in sites of

314

Meador. Deyoe, and Cheville

Figs. 7- 10. Immunoperoxidase-stained brucellar antigen (brown) in mammary glands and supramammary lymph nodes, nonnursing, infected goats. Hematoxylin counterstain. Fig. 7. Duct lumen (DL) filled with macrophages and neutrophils that contain abundant brucellar antigen. Few infected cells are in the ductal epithelial layer (short arrow) and periductal interstitium (long arrows). Fig. 8. Infected phagocytes in alveolar lumens. Fig. 9. Macrophages containing brucellae (arrows) in the subcapsular sinus of a supramammary lymph node. Fig. 10. Small focus of macrophages in subcapsular sinus of supramammary lymph node. Abundant brucellar antigen in macrophages.

afferent lymph drainage in lymph nodes (i.e., subcapsular sinuses) is a further indication of dissemination of Brucella by lymph. Brucellae have been seen in macrophages and neutrophils in lymphatics draining sites of inoculation. Isolation of brucellae from more tissue sites peripheral to the udder in nonnursing goats, accompanied by higher serum standard tube agglutination test titers, suggests that there was increased systemic circulation of brucellae as compared with nursing goats. Since lymph from the mammary gland eventually enters the blood, transportation of infected

phagocytes in lymph could disseminate organisms throughout the host. BruceIIa localization in mammary glands is not completely inhibited by nursing. Milk expulsion during lactation is not always uniform throughout the udder; interference with complete expulsion of milk due to functional variation of individual lobules during lactation occurs normally and as a sequel to previous incidences of mastitis. Mammary glands of cows nursing single or twin calves would probably be unequally nursed, causing variations in the degree of milk stasis

B. abortus Infec:tion of Goats

among mammae. Brucella may be expected to localize more frequently in the least nursed glands and in inactive lobules in actively lactating mammary glands. In nursing goats of this study, indications of mammary gland infection may have been more apparent if each goat had nursed a single kid instead of twins; flushing of mammary glands and removal of brucellae would have been less complete. Brucellae also localize in phagocytic cells in mammary gland interstitium.I6 It seems probable that at least some infected cells would localize in the interstitium regardless of lactational status. Nursing and nonnursing goats in this study most closely resemble lactating, milked dairy cows in early lactation and cows delivering dead fetuses in late gestation, respectively. Our findings, therefore, may be more applicable to B. abortus infection in these two groups of cows. If findings in goats are applicable to cattle, cows that abort in late gestation and are not subsequently milked should have a high incidence of mammary gland infection and may be likely to excrete organisms in subsequent lactations. Acknowledgements We thank Judi Stasko, Kathryn Meredith, and Carol Irvin for technical assistance.

References 1 Ackermann MR, Cheville NF, Deyoe BL: Bovine ileal dome lymphoepithelial cells: endocytosis and transport ofBrucella abortus strain 19. Vet Pathol2528-35, 1988 2 Anderson TD, Cheville NF, Meador VP: Pathogenesis of placentitis in the goat inoculated with Brucella abortus. 11. Ultrastructural studies. Vet Pathol 23:227-239, 1986 3 Anderson TD, Meador VP, Cheville NF: Pathogenesis ofpalcentitis in the goat inoculated with Brucella abortus. I. Gross and histologic lesions. Vet Pathol 23:2 19-226, 1986 4 Corner LA, Alton GG, Iyer H: Distribution of Brucella abortus in infected cattle. Aust Vet J 64:241-244, 1987 5 Doyle TM: Brucella abortus infection of goats. J Comp Pathol Ther 52:89-115, 1939 6 Emminger AC, Schalm OW: The effect of Brucella abortus on the bovine udder and its secretion. Am J Vet Res 4~100-109, 1943 7 Fensterbank R: Some aspects of experimental brucellosis. Ann Rech Vet 18:42 1-428, 1987 8 Hofstad MS: The changes produced by Brucella abortus in the milk and udder of cows infected with Bang’s disease. Cornell Vet 32:289-294, 1942 9 Kehrli ME, Nonnecke BJ, Roth JA: Alterations in bovine peripheral blood neutrophil function during the penparturient period. Am J Vet Res 50:207-214, 1989 10 Kehrli ME, Nonnecke BJ, Roth JA: Alterations in bovine peripheral blood lymphocyte function during the periparturient period. Am J Vet Res 50:2 15-220, 1989

375

11 Lambert G, Amerault TE, Manthei CA, Goode ER: Further studies on the persistence of Brucella abortus infection in cattle. Proc Annu Meet US Livestock Sanit ASSOC64:109-117, 1960 12 Manthei CA, Carter WR: Persistence of Brucella abortus infection in cattle. Am J Vet Res 11:173-180, 1950 13 McDonald JS, Anderson AJ: Total and differential somatic cell counts in secretions from noninfected bovine mammary glands; the early nonlactating period. Am J Vet Res 42:1360-1365, 1981 14 McDonald JS, Anderson AJ: Total and differential somatic cell counts in secretions from noninfected bovine mammary glands; the peripartum period. Am J Vet Res 41~1366-1368, 1981 15 Meador VP, Deyoe BL: Experimentally induced Brucella abortus infection in pregnant goats. Am J Vet Res 47: 2337-2342, 1986 16 Meador VP, Deyoe BL, Cheville N F Pathogenesis of Brucella abortus infection of the mammary gland and supramammary lynph node of the goat. Vet Pathol 26: 357-368, 1989 17 Meador VP, Hagemoser WA, Deyoe BL: Histologic changes in Brucella abortus-infected, pregnant goats. Am J Vet Res 49:274-280, 1988 18 Meador VP, Tabatabai LB, Hagemoser WA, Deyoe B L 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 19 Morgan WJB, McDiarmid A: The excretion of Brucella abortus in the milk of experimentally infected cattle. Res Vet Sci 1:53-56, 1960 20 Neave FK, Oliver J, Dodd FH, Higgs TM: Rate of infection of milked and unmilked udders. J Dairy Res 35: 127-134, 1968 21 Oliver J, Dodd F, Neave FK, Bailey GL: Variations in the incidence of udder infection and mastitis with stage in lactation, age, and season of the year. J Dairy Res 23: 181-193, 1956 22 Payne JM: The pathogenesis of experimental brucellosis in the pregnant cow. J Pathol Bacteriol78:447463, 1959 23 Runnells RA, Huddleson I F The nature of Bacterium abortus infection of the udder of the bovine. Cornell Vet 15~376-390, 1925 24 Sheehan DC, Hrapchak BB: Theory and Practice of Histotechnology. CV Mosby Company, St. Louis, MO, 1980 25 Sholl LB, Torrey JP: A contribution to the bacteriology and pathology of the bovine udder. Mich Agric Exp Sta Tech Bull, p. 110, 1931 26 US Department of Agriculture, Animal Plant and Health Inspection Service: Supplemental test procedures for the diagnosis of brucellosis. Diagnostic Reagents Manual 65E 27 Thomas CL, Neave FK, Dodd FH, Higgs TM: The susceptibility of milked and unmilked udder quarters to intra-mammary infection. J Dairy Res 39: 1 13-1 3 1, 1972 28 Washko FV,Donham CR, Hutchings LM, Heimlich A: Recovery of Brucella from tissues of cattle exposed to Brucella abortus. J Am Vet Med Assoc 120:82-84, 1952

Request reprints from Dr. V. P. Meador, National Animal Disease Center, PO Box 70, Dayton Road, Ames, I A 50010 (USA).