Effect of pregnancy on the immune response of cattle to a Brucella vaccine

Journal of Reproductive Immunology, 9 (1986) 313-325 Elsevier

313

JRI 00425

Effect of pregnancy on the immune response of cattle to a Brucella vaccine A.J. Winter 1,., C.E. Hall 1, R.H. Jacobson 1, D.R. Verstreate M.P. Meredith 2 and W.L. Castleman 1

1,

J New York State College of Veterinary Medicine and 2 Biometrics Unit, Cornell University, Ithaca, N Y 14853, U.S.A. (Accepted for publication 2 May 1986)

An experiment was performed to determine whether humoral- or cell-mediated immune responses of cattle to a Brucella abortus vaccine were influenced by the stage of gestation. Heifers were vaccinated 2 mth before and 2 mth after breeding with cell envelopes of B. abortus in an oil adjuvant containing trehalose dimycolate and muramyl dipeptide. Control groups received adjuvant alone or no vaccine. Following breeding, vaccinated animals were divided into pregnant and nonpregnant subgroups. Immune responses to two outer membrane proteins were measured at monthly intervals by ELISA and lymphocyte blastogenesis tests. Skin tests were performed during the ninth month of gestation. Vaccination induced sustained immune responses, but few differences were detected between pregnant and non-pregnant animals. The relative increase in IgA antibodies to group 3 protein in nonpregnant heifers exceeded that in pregnant heifers during months 4 and 6 of gestation ( P < 0.05). Dermal hypersensitivity, measured by changes in double skin thickness, was significantly greater in nonpregnant heifers to porin ( P < 0.01) and group 3 ( P < 0.05) antigens at 24 h post-injection, but no significant differences in skin thicknesses or in the nature of the lesions were observed at 48 h. Animals which received adjuvant alone demonstrated negligible responses. Pregnancy had no significant effect on the responses of lymphocytes to phytohemagglutinin (PHA) or Concanavalin A (Con A). However, plasmas from nonvaccinated pregnant heifers taken during the sixth and seventh (but not eighth or ninth) months of pregnancy decreased responses of normal donor cells to PHA and Con A when compared with those in autologous plasma ( P < 0.05). Key words: cattle, immune response, pregnancy, Brucella vaccine

Introduction B r u c e l l o s i s o f c a t t l e is c a u s e d b y B r u c e l l a a b o r t u s , a f a c u l t a t i v e i n t r a c e l l u l a r parasite in which cell-mediated immune (CMI) responses are believed to have a c r i t i c a l r o l e i n p r o t e c t i v e i m m u n i t y ( M a c k a n e s s , 19.64; C h e e r s , 1984). P r e g n a n t cattle become progressively more susceptible to bKicellosis and undergo a more s e v e r e f o r m o f t h e d i s e a s e ( N a t i o n a l A c a d e m y o f S c i e n c e s , 1977). T h i s p h e n o m e n o n r e m a i n s l a r g e l y u n e x p l a i n e d , a l t h o u g h it m a y b e a c c o u n t e d f o r i n p a r t b y t h e h i g h

* To whom correspondence should be addressed (Room 225, Veterinary Research Tower). 0165-0378/86/$03.50 © 1986 Elsevier Science Publishers B.V.

314

concentrations of erythritol in the fetal placenta and fetal fluids (National Academy of Sciences, 1977). It is well recognized in several species, including man, that certain pathogens produce more severe disease in pregnant females (Weinberg, 1984; Brabin, 1985). Many, although not all, such pathogens are facultative intracellular parasites, and for diseases produced by this category of pathogens it has been hypothesized that the nonspecific depression of CMI responses which occurs during pregnancy contributes to diminished resistance to disease (Weinberg, 1984). The occurrence of humoral immunosuppressive factors during pregnancy has been well documented in a number of species, including cattle (Griffin, 1981; Splitter and Everlith, 1982; Lloyd, 1983; Weinberg, 1984). However, controlled studies have not been performed to test the effect of pregnancy on the specific immune response of cattle. An important aim of our research is to produce a nonliving vaccine for bovine brucellosis. Such products were evaluated previously in long-term studies in nonpregnant heifers and short-term (10-wk) trials in cows (Winter et al., 1983). Some of the cows were pregnant, at maximum in the fourth month of gestation (Winter et al., i983). However, to be effective a vaccine against B. abortus must produce immunity which is sustained throughout gestation. The principal objective of the work reported here was to determine whether humoral or CMI responses of cattle to a B. abortus vaccine were altered quantitatively or qualitatively during the last two trimesters. We believed that the occurrence of such altered responses to vaccine antigens might also aid in understanding the basis for increased susceptibility of pregnant cattle to infection and disease caused by B. abortus. The design of the experiment was based on a schedule of immunization being employed in vaccination-challenge trials in cattle (L.G. Adams and A.J. Winter, unpublished data).

Materials and Methods

Antigens for vaccines and immunological assays B. abortus strain 45/20 was grown and extracted as described previously (Winter et al., 1983). Cell envelopes were employed in vaccines and porins and group 3 proteins were used for ELISA, blastogenesis, and skin tests.

Adjuvants Trehalose dimycolate (TDM) extracted from Mycobacterium boris was purchased from Choay Chimie Reactifs (Paris, France). The derivative of muramyl dipeptide (MDP), N-acetylmuramyl-L-a-aminobutyryl-D-isoglutamine, used previously (Winter et al., 1983), was employed in this study. Preparation and administration of vaccines The described method (Winter et al., 1983) was followed, except that spermidine was included in the mixture and distilled water containing 0.5% formalin was substituted for phosphate-buffered saline. Each dose of vaccine contained cell

315 envelopes (5 mg), T D M (5 rag), MDP (5 mg), spermidine (100/~1), mineral oil (1.8 ml) and Tween-water solution (8.2 ml).

Blood sampling Blood samples were collected from the tail vein into heparinized or untreated vacutainer tubes. Serum and plasma were stored at - 2 0 ° C. Lymphocyte blastogenesis Tests were performed by cell titration assays described previously (Winter et al., 1983; Baldwin et al., 1985a,b) with minor modifications. Porin and group 3 antigens were tested with two cell concentrations (2 × 105 and 4 x 105 cells per well) and Con A (Sigma Chemical Co.) and phytohemagglutinin-P (PHA) (Difco Laboratories) were tested with four concentrations (1.25 × 104, 2.5 × 104, 5 × 10 4 and 1 × 105) of cells per well. Medium was supplemented weekly with 2 mM L-glutamine, and 2-mercaptoethanol was omitted from medium used to test antigens. Data from month 14 were incomplete due to a contamination problem and were not included in the analyses. An experiment was performed to test the effect of plasma obtained during the last 4 mth of pregnancy on lymphocyte blastogenesis in response to PHA and Con A. The response of cells from the same two normal nonpregnant donor heifers was tested. Plasma from an individual donor was used as a standard for its own cells. Donor plasmas had been frozen and held at - 2 0 ° C , just as were those of the animals under test. Autologous or heterologous plasmas, at a final concentration of 10%, were substituted in the medium for 10% fetal calf serum. Heterologous plasmas were always tested separately, rather than as pools. ELISA Kinetics-based ELISA was performed with porin and group 3 antigens, and data were expressed as percentages of increase in slope values over control values ([slope of postvaccination sample - mean slope of prevaccination samples]/[mean slope of prevaccination samples]) (Winter et al., 1983). Sera from selected heifers were also tested with group 3 antigen for the distribution of antibody isotypes. Horseradish peroxidase conjugates of rabbit antisera (IgG fractions) specific for bovine IgG1, IgG2, IgM and IgA were obtained from Dr. Kiaus Nielsen (A.D.R.I., Nepean, Ontario, Canada). The specificity of each conjugate was tested in ELISA against 200 ng of purified bovine isotype per well. At working dilutions the four conjugates produced OD414 n m > 1.64 with the homologous isotype and < 0.10 with heterologous isotypes (K. Nielsen, pers. comm., 1984). Skin tests Skin tests were performed and punch biopsies were taken from each site and processed as described before (Winter et al., 1983). Tissue sections were coded and scored by one person (W.L.C.) using the following criteria: nature of the cellular infiltrate; quantity of cellular infiltrate and its distribution within the dermis and perivascularly; and the degree of dermal edema.

316

Experimental animals and procedures Twenty-two Holstein-Friesian heifers were used. Animals were 15-17 months old at the beginning of the experiment. Preimmunization blood samples were tested five times over a period of 4 months. Animals were then divided randomly into vaccination groups A, B and C (Table 1). In month 4, group A was given the complete vaccine, group B received adjuvant without antigen, and group C was not vaccinated. All heifers were synchronized to estrus by treatment with prostaglandin F2a (Lutalyze, Upjohn Co.) and bred by artificial insemination 2 mth after the first immunization. Five heifers (one in group B and four in group C) returned to estrus and were rebred. One animal from group C which failed to become pregnant after the second breeding was dropped from the experiment. Two months after the first breeding (month 9 of the experiment) heifers in groups A and B were divided into nonpregnant (1 and 3) and pregnant (2 and 4) groups (Table 1). Groups 1 and 3 were composed of animals which had failed to become pregnant or in which pregnancy was terminated at this time by prostaglandin F2ct (required for three heifers in group 1 and one in group 3). Heifers in groups A and B were revaccinated with complete vaccine or adjuvant alone, respectively, 4½ mth after the first vaccination. Blood samples were tested at monthly intervals from the time of the first vaccination until the ninth month of gestation. For this purpose heifers were divided into three groups which were bled within the period of 1 wk. Groups for bleeding were composed insofar as possible of equal numbers of heifers from each treatment group. Statistical analyses, unless otherwise stated, were performed on data obtained during the last 7 mth of gestation (months 9 to 15 of the experiment). Skin tests were performed during the ninth month of gestation, after the last blood samples had been drawn.

Statistical analyses Lymphocyte blastogenesis Responses were measured as the geometric mean of the triplicate tests at each cell concentration. The logarithm of the geometric means was used in the analyses to assure approximate normality and variance homogeneity. Background counts were ignored for mitogen assays. Data for antigens were analyzed on the basis of both the logarithm of the stimulation index (SI) and the

TABLE 1 Treatment groups of heifers to determine effect of pregnancy on the immune response Vaccination group

No. of heifers

Vaccine

Pregnancy group

No. of heifers

Pregnancy status

A

11

Antigen + adjuvant

B

6

Adjuvant only

C

5

None

1 2 3 4 5

5 6 3 3 4

NP P NP P P

P, pregnant; NP, nonpregnant.

317 logarithm of the difference in counts per minute between antigen-treated and control wells (ACPM) (Baldwin et al., 1985b). Analyses of responses to antigens and mitogens were always restricted to horizontal (within-month) comparisons. The treatment structure for the antigen responses utilized a 23 factorial combination of pregnancy (pregnant or not pregnant) by vaccination (adjuvant with and without antigen) by cell concentration (2 x 105 and 4 X 105). The group of pregnant control animals (group 5) was always included at both levels of cell concentration, giving a total of 10 treatment combinations. The pregnancy and vaccination factors were allocated to heifers in a randomized design and each cell concentration was done for each animal. Responses were analyzed as a split-plot experiment (Cochran and Cox, 1957). Because pregnancy and vaccination factors were found to behave similarly at both cell concentrations (interaction, P > 0.20), comparisons were performed with values averaged over the two cell concentrations. The same analyses were performed on the mitogen responses, except that four cell concentrations were utilized. The treatment structure was thus a 2 X 2 X 4 factorial combination of treatments plus the pregnant control animals at the four cell concentrations. However, to assure a straight line response to increasing cell concentrations the highest concentration was omitted from the analysis.

ELISA

Data from heifers in groups 1 and 2 were averaged from each month's tests, t-tests were performed to determine whether levels of antibodies between the two groups differed at each respective time period.

Results

Antigen-induced lymphocyte blastogenesis Fig. 1 illustrates lymphocyte transformation stimulated by porin in the different treatment groups. The first vaccination produced sustained increases in response (groups 1 and 2). The second vaccination produced another rise in response, most evident in month 11. Reactions declined slowly over the next 4 months. In groups injected with adjuvant alone (3 and 4) or untreated (5) the mean response was, with few exceptions, below 10,000 ACPM (Fig. 1). There were no significant differences between responses of group 1 (vaccinated, nonpregnant) and group 2 (vaccinated, pregnant) at any time from months 9 to 15. This was so for both antigens and held true whether data were analyzed on the basis of stimulation index or ACPM. At each time period the magnitude of responses from groups 1 and 2 was significantly greater than that of groups 3 and 4 (adjuvant alone) when analyzed on the basis of stimulation index ( P < 0.01) or ACPM ( P < 0.01). No significant difference in the response occurred at any time period between groups 3 and 4, and when values of groups 3 and 4 were pooled they did not differ significantly from group 5 (untreated pregnant controls).

Mitogen induced lymphocyte blastogenesis No significant differences in response to Con A or PHA occurred within treatment groups 1 to 5 at any time between months 9 and 15. Pregnancy did not

318 /

./' 0

o~q

/

/

.."

~

..'

v,.

.."

./ 0

....

.

..." z771

,"

~"

.."

~

....

." ...

/ z~

,

//

......

/"

/

/'

~

/

,..." z~" i I

....,....

/

/'

/

/

.." I ~ 1I

,'

/.,'

-

,' ~"

/

,, ......

'

r 4

A i I

6

/

IA [J

i 7

i 8 Month

6 91 of

,"¢c~l /

i

r IO

i II

r 12

| 13

/ 15

experimenf

Fig. 1. Response in blastogenesis tests to the porin of B. abortus by lymphocytes from heifers in the five treatment groups during months 4 through 15 of the experiment. Each block represents approximately 10,000 ACPM from tests with 4 × 10 5 cells per well. For identification of groups see Table 1. Times of vaccination (black arrows) and breeding (white arrow) are indicated.

alter the magnitude of the responses. Thus, responses to either mitogen of pregnant heifers (combined data of groups 2, 4 and 5) did not differ from those of nonpregnant animals (combined data from groups 1 and 3) at any time period.

Effects of bovine plasmas on blastogenesis tests Plasmas were tested from three of the four heifers in group 5 (pregnant, nonvaccinated). The magnitude of the responses to both PHA and Con A in cultures containing plasmas from the sixth month of pregnancy was significantly lower ( P < 0.05) than in cultures containing autologous plasma. This held true for cells from both donors. Plasmas from the seventh month of pregnancy produced a significant reduction in responses to PHA and Con A ( P < 0.05) in cultures from only one of the two donors. No significant differences were detected with plasmas from the eighth or ninth months of pregnancy. Antibody responses Antibodies were detectable only in groups 1 and 2. Titers rose after each vaccination and then gradually declined (Fig. 2). Similar responses occurred to both antigens, although the magnitude of response to group 3 antigen was almost always greater (Fig. 2). The mean responses of groups 1 and 2 did not differ significantly from each other at any time period from months 10 through 15 of the experiment. Antibodies of each isotype were measure d against the group 3 antigen during the last 6 mth of the experiment. The percentage of increase of IgG l, IgG z, IgM and IgA antibodies generally occurred in a decreasing order (Fig. 3). However, the relative increase of IgA antibodies exceeded that of IgM antibodies in group 1 (vaccinated, nonpregnant) at months 10 and 11 (Fig. 3) and the percentage increase of IgA in group 1 was significantly greater than that in group 2 (vaccinated, pregnant) durin~ months 10 and 12 ( P < 0.05). No other significant differences in

319 0.20 ..-'O-...O

o. t5

,,,o"

o-,,

O

-~

/ ,'

//

~o . , o

'

"'-o, •

~0..-0---0 /

"0--0

0--.0

°--°--°--°-"°I"%'0-0

0.00

i '2

' 3 '41~ 5 ' 6 '071 8 '91 l l 0 t

titl2113'141

15 I

Month of experiment

Fig. 2. Antibody responses measured by ELISA of one heifer in treatment group 1 to porin (solid line) and group 3 (broken line) antigens of B. abortus strain 45/20. Tests were performed over the 15 mth course of the experiment. Times of vaccination (black arrows), breeding (white arrow), and skin testing (striped arrow) are indicated.

concentrations of antibodies of individual isotypes occurred between groups 1 and 2 at any time period. S k i n tests

A double skin thickness >/2.0 m m was considered to represent a positive reaction (Nicoletti, 1983/1984). At 24 h after injection all reactions in groups 1 and

800

o A O o

GROUP 1 °----------o~

-

= IgG=

= igGt = IgM = IgA

600

e

Q. 0

400

.c_ 2 0 0 o

=o 800

GROUP 2

o----._.__ o ~ D , ,

~

600 o_

40O

200 0

I

I0

I

II Month

t

12

15

t

of experiment

Fig. 3. Isotype specific antibody responses to group 3 antigen measured by ELISA during the last 6 mth of the experiment. Points represent mean responses of heifers within treatment groups i and 2, respectively.

320 Skin test Antigen

48 hr post injection

24 hr post injection

Treatment Group I

mp

2 PORIN 5 4 5

3

) ?

I

1P'°°5

2 GROUP 5

3 ) )

:5 4

S3

5 I

I

I

I

I

I

I

0 2 4 6 8 1 0

[ ~

I I

0

I

I

4

I

I I

6

I I

I

8 I0

Difference in skin thickness (mm)

Fig. 4. Skin test responses to porin and group 3 antigens by heifers in the treatment groups. Statistically significant differences between groups 1 and 2 are indicated. Bars, standard error.

2 were positive and, except for reactions of one heifer in group 3 and one in group 5 to group 3 antigen, all reactions in groups 3, 4 and 5 were negative (Fig. 4). Reactions in group 1 were significantly larger than those in group 2 for porin ( P < 0.01) and group 3 ( P < 0.05). By 48 h this difference was no longer significant, although mean responsees of group 1 were slightly greater to each antigen (Fig. 4). At 48 h all reactions were negative in groups 3, 4 and 5. Histological lesions were the same as those described previously (Winter et al., 1983) and were consistent with changes referable primarily to states of delayed type hypersensitivity. The cellular infiltrate was composed of macrophages and lymphocytes, with variable numbers of neutrophils and eosinophils. It was not possible on the basis of scores assigned to the lesions to distinguish between heifers of groups 1 and 2.

Discussion

The development of subunit vaccines for bovine brucellosis for use in place of attenuated strain 19 constitutes an important objective in several laboratories (Bosseray et al., 1978; Winter et al., 1983; Tabatabai and Deyoe, 1984; Wu et al.,

321 1984; Sowa et al., 1985). It is tacitly assumed that the immune response to vaccines will not be altered during pregnancy, although this point cannot be tested in studies which require challenge with a virulent strain of B. abortus. The data presented here demonstrate that the magnitude, quality and duration of humoral and CMI responses to a Brucella vaccine were, with few exceptions, uninfluenced by pregnancy. Such a conclusion must, of course, be qualified in that it is based on a limited set of assays in relatively small experimental groups. Furthermore, a powerful adjuvant was utilized, so that generalizations to other vaccines must also be limited. In several studies in women it has been concluded from quantitative measures that antibody responses to vaccines during pregnancy were satisfactory, although in none of these were simultaneous comparisons made over time with matched nonpregnant individuals (Brown and Carroll, 1958; Carvalho et al., 1977; Sumaya and Gibbs, 1979; Brabin et al., 1984). Prolonged immunization of rabbits with killed B. abortus cells leads to the development of IgA antibodies which block complement-mediated killing of the organism (Hall et al., 1971). IgA-blocking antibodies have also been detected in sera of patients convalescing from infection with Neisseria meningitidis (Griffiss, 1975), and IgG-blocking antibodies have been proposed to contribute to the development of disseminated gonococcal infection in pregnant women (McCutchan et al., 1978). A shift in isotypes of antibodies formed during pregnancy could reflect altered effector functions. No shift in isotypes of IgG antibodies to Brucella group 3 antigen occurred during the last 6 months of pregnancy (Fig. 3). The biological relevance of the small but significant decrease in the proportion of IgA antibodies during the second trimester (Fig. 3) is unknown. Recent reports provide evidence that in rodents protective immunity to the facultative intracellular pathogen Listeria monocytogenes is mediated by the cooperative interaction of two T lymphocyte subsets, of which one is responsible for DTH (Chen-Woan et al., 1985; Kaufmann et al., 1985). It is therefore noteworthy that DTH responses to Brucella antigens were sustained in heifers in their ninth month of gestation, 6 mth after the second vaccination. DTH responses in pregnant heifers were equivalent to those in nonpregnant controls, except for the kinetics of their development (Fig. 4). The reason for this is not apparent. It seems unlikely, based on histological findings and the use of denatured skin test antigens unreactive with antibodies to native antigens (Winter et al., 1983), that reactions at 24 h were attributable to antibody-mediated Arthus reactions. It is conceivable that the difference in 24 h reactions reflected a slower mobilization of DTH in pregnant cattle, and that this could constitute a partial explanation for the increased susceptibility of pregnant cattle to brucellosis. The effect of pregnancy on DTH in women is controversial. In some studies DTH responses have been reported to be normal (Montgomery et al., 1968; Rocklin et al., 1979; Hawes et al., 1981), but others have found tuberculin responses to be depressed in the late stages of pregnancy (Lichtenstein, 1942; Finn et al., 1972; Anderson et al., 1974). Although blastogenic transformation of lymphocytes does not in itself constitute a strict correlate of CMI or even of T cell function, our use of denatured protein antigens ensured that responding cells were largely or exclusively T cells requiting

322 antigen processing and presentation by macrophages or other accessory cells. These include T helper/inducer cells and cells mediating DTH (Tamura et al., 1983; Unanue et al., 1984; Morita et al., 1985). The magnitude of the cellular response to either of the antigens tested was not influenced by pregnancy during the last two trimesters, although the presence in pregnant heifers of humoral factors which may have inhibited the cellular response in vivo cannot be excluded. The presence in plasma of factors suppressing mitogen responses during the second trimester may bear on this point. Pregnancy-associated inhibition of lymphocyte blastogenesis to antigens of infectious agents has been reported in some trials conducted in women. In various studies inhibition appeared to result from humoral factors (Leikin, 1972; Lopatin et al., 1980; Kumar et al., 1984), intrinsic alterations of cell function (Covelli and Wilson, 1978; Birkeland and Kristoffersen, 1980; Lopatin et al., 1980; Gehrz et al., 1981), or to a combination of these (Lopatin et al., 1980). In some studies inhibition was demonstrated toward some, but not all, of the antigens tested (Kumar et al., 1984) or failed to occur (Hawes et al., 1981). Comparison of those data to ours must be qualified because of differences in experimental design, in particular in the types of controls employed. The effect of adjuvant alone was tested because in a prior study some cattle given the same adjuvant produced substantial responses to antigens in the blastogenesis test (L.G. Admas and A.J. Winter, unpubl, data). TDM and MDP produce powerful nonspecific effects (Lederer, 1980) which through a possible indirect stimulation of interleukin 2 could affect the immune response (Grimm et al., 1983; LeFrancois et al., 1984). In this study, however, the reactivities of heifers given adjuvant alone (groups 3 and 4) were no different from those which were untreated (group 5) by any of the measurements employed. Numerous studies exist in man and other species on mitogen-induced transformation of lymphocytes during pregnancy (for reviews and representative studies in women see Leikin, 1972; Birkeland and Kristoffersen, 1977, 1980; Fujisaki et al., 1979; Lopatin et al., 1980; Hawes et al., 1981; Lloyd, 1983; Valdimarsson et al., 1983; Kumar et al., 1984). Although in many studies responses to one or more mitogens were found to be inhibited during pregnancy, conflicting reports exist. For example, responses to PHA during pregnancy have been stated to be depressed (Leikin, 1972; Fujisaki et al., 1979; Lopatin et al., 1980; Valdimarsson et al., 1983; Kumar et al.,, 1984), unaffected (Birkeland and Kristoffersen, 1977, 1980; Gehrz et al., 1981), or enhanced (Hawes et al., 1981). In most studies suppression has been ascribed to humoral factors, but an intrinsic alteration in cellular reactivity has also been demonstrated (Lopatin et al., 1980). In a well designed study in cattle, Manak (1982) tested the blastogenic responses to Con A, PHA, and pokeweed mitogen of lymphocytes during successive stages of pregnancy. Comparisons were made with lymphocytes from ovariectomized heifers, and cells were cultured in parallel in the presence of serum from pregnant and ovariectomized heifers. A pattern was observed with all three mitogens, although less clearly with PHA, in which responses of lymphocytes from pregnant heifers were greater than those from ovariectomized animals during the latter half of gestation. Furthermore, sera from 3-7 mth pregnant heifers produced a suppression of mitogenic activity (Manak, 1982). In our

323 experiment comparisons were made between lymphocytes of pregnant and normal, rather than ovariectomized, heifers using 10% fetal calf serum in the medium, and no significant differences in reactivity to P H A or C o n A were noted during the last 7 months of gestation. However, in agreement with Manak, plasmas from normal heifers (group 5) during the sixth and seventh m o n t h s of gestation inhibited mitogenic responses to C o n A and PHA. The relationship of the responsible inhibitory factor or factors to immunoregulatory proteins or steroid hormones which have been defined in cattle and other species (Griffin, 1981; Lloyd, 1983; Weinberg, 1984) remains to be established. It is not certain whether the minor differences in i m m u n e responsiveness demonstrated in this study between pregnant and n o n p r e g n a n t cattle will prove to be relevant in explaining the increased susceptibility of pregnant cattle to brucellosis. A n explanation of this p h e n o m e n o n is far from complete in any species, and is n o w regarded as having a multifactorial basis (Weinberg, 1984; Brabin, 1985). The recent finding that activated macrophages from pregnant mice were deficient in killing Toxoplasma gondii (Luft and Remington, 1984) exemplifies a potentially important advance in understanding this complex phenomenon.

Acknowledgements The authors are grateful to W.G. TenHagen, G.E. Rowe, and R.L. Vickstrom for technical assistance and to J. R e y n a for preparation of the manuscript. This research was supported in part by U.S. D e p a r t m e n t of Agriculture grant 59-2361-02-080-0.

References Anderson, F.D., Ushijima, R.N. and Larson, C.L. (1974) Recurrent herpes genitalis. Treatment with Mycobacterium boris (BCG). Obstet. Gynecol. 43, 797-805. Baldwin, C.L., Antczak, D.F. and Winter, A.J. (1985a) Cell titration assay for measuring blastogenesis of bovine lymphocytes. Vet. Immunol. Immunopathol. 9, 319-333. Baldwin, C.L., Verstreate, D.L. and Winter, A.J. (1985b) Immune response of cattle to Brucella abortus outer membrane proteins measured by lymphocyte blastogenesis. Vet. Immunol. Immunopathol. 9, 383~-396. Birkeland, S.A. and Kristoffersen, K. (1977) Cellular immunity in pregnancy: blast transformation and rosette formation of maternal T and B lymphocytes. Clin. Exp. Immunol. 30, 408-412. Birkeland, S.A. and Kristoffersen, K. (1980) Lymphocyte transformation with mitogens and antigens during normal human pregnancy: a longitudinal study. Scand. J. Immunol. 11, 321-325. Bosseray, N., Plommet, M. and Dubray, G. (1978) Immunogenic activity of a cell wall fraction extracted from Brucella abortus in guinea-pigs. Ann. Microbiol. (Inst. Pasteur) 129B, 571-579. Brabin, B.J. (1985) Epidemiology of infection in pregnancy. Rev. Infect. Dis. 7, 579-603. Brabin, B.J., Nagel, J., Hagenaars, A.M., Rutenberg, E. and Van Tilborgh, A.M.J.C. (1984) The influence of malaria and gestation on the immune response to one and two doses of adsorbed tetanus toxoid in pregnancy. Bull. WHO 62, 919-930. Brown, G.C. and Carroll, C.J. (1958) Antibody response of pregnant women to poliomyelitis vaccine and passive transfer to infants. J. Immunol. 81, 389-395.

324 Carvalho, A.A., Giampaglia, C.M.S., Kimura, H., Pereira, O.A.C., Farhat, C.K., Neves, J.C., Prandini, R., Carvalho, E.S. and Zarvos, A.M. (1977) Maternal and infant antibody response to meningococcal vaccination in pregnancy. Lancet ii, 809-811. Cheers, C. (1984) Pathogenesis and cellular immunity in experimental murine brucellosis. In 3rd International Symposium on Brucellosis, Algiers, Algeria, 1983. Develop. Biol. Standard. Vol. 56 (Valette, L. and Hennessen, W., eds.), pp. 237-246, S. Karger, Basel. Chen-Woan, M., Sajewski, D.H. and McGregor, D.D. (1985) T cell cooperation in the mediation of acquired resistance to Listeria monocytogenes. Immunol. 56, 33-42. Cochran, W.G. and Cox, G.M. (1957) Experimental Designs. John Wiley and Sons, Inc., New York. Covelli, H.D. and Wilson, R.T. (1978) Immunologic and medical considerations in tuberculin-sensitized pregnant patients. Am. J. Obstet. Gynecol. 132, 256-259. Finn, R., St. Hill, C.A., Govan, A.J., Ralfs, I.G., Gurney, F.J. and Denye, V. (1972) Immunological responses in pregnancy and survival of fetal homograft. Br. Med. J. 3,150-152. Fujisaki, S., Moil, N., Sasaki, T. and Maeyama, M. (1979) Cell-mediated immunity in human pregnancy. Microbiol. Immunol. 23, 899-907. Gehrz, R.C., Christianson, W.R., Linner, K.M., Conroy, M.M., McCue, S.A. and Balfour Jr., H.H. (1981) Cytomegalovirus-specific humoral and cellular immune responses in human pregnancy. J. Infect. Dis. 143, 391-395. Griffin, J.F.T. (1981) Materno-fetal tolerance: nature's gift to viviparity. Irish Vet. J. 35, 194-204. Griffiss, J.M. (1975) Bactericidal activity of meningococcal antisera: blocking by IgA of lytic antibody in human convalescent sera. J. Immunol. 114, 1779-1784. Grimm, E.A., Robb, R.J., Roth, J.A., Neckers, L.M., Lachman, L.B., Wilson, D.J. and Rosenberg, S.A. (1983) Lymphokine-activated killer cell phenomenon III. Evidence that IL-2 is sufficient for direct activation of peripheral blood lymphocytes into lymphokine-activated killer cells. J. Exp. Med. 158, 1356-1361. Hall, W.H., Manion, R.E. and Zinneman, H.H. (1971) Blocking serum lysis of BruceUa abortus by hyperimmune rabbit immunoglobulin A. J. Immunol. 107, 41-46. Hawes, C., Kemp, A.S., Jones, W.R. and Need, J.A. (1981) A longitudinal study of cell-mediated immunity in human pregnancy. J. Repro Immunol. 3, 165-173. Kanfmann, S.H.E., Hug, E., Vath, U. and Muller, I. (1985) Effective p~otection against Listeria monocytogenes and delayed-type hypersensitivity to listerial antigens depend on cooperation between specific L3T4 + and Lyt2 + T cells. Infect. Immun. 48, 263-266. Kumar, A., Madden, D.L. and Nankervis, G.A. (1984) Humoral and cell-mediated immune responses to herpesvirus antigens during pregnancy - - a longitudinal study. J. Chn. Immun. 4, 12-17. Lederer, E. (1980) Immunostimulation: Recent progress in the study of natural and synthetic immunomodulators derived from the bacterial cell wall. In Progress in Immunology IV (Fougereau, M. and Dausset, J., eds.), pp. 1194-1211. LeFrancois, L., Klein, J.R., Paetkau, V. and Bevan, M.J. (1984) Antigen-independent activation of memory cytotoxic T cells by interleuldn 2. J. Immunol. 132, 1845-1850. Leikin, S. (1972) Depressed maternal lymphocyte response to phytohaemagglutinin in pregnancy. Lancet ii, 43. Lichtenstein, M.R. (1942) Tuberculin reaction in tuberculosis during pregnancy. Am. Rev. Tuber. 46, 89-92. Lloyd, S. (1983) Effect of pregnancy and lactation upon infection. Vet. Immunol. Immunopathol. 4, 153-176. Lopatin, D.E., Korman, K.S. and Loesche, W.J. (1980) Modulation of immunoreactivity to periodontal disease-associated microorganisms during pregnancy. Infect. Immun. 28, 713-718. Luft, B.J. and Remington, J.S. (1984) The adverse effect of pregnancy on macrophage activation. Cell Immunol. 85, 94-99. Mackaness, G.B. (1964) The immunological basis of acquired cellular resistance. J. Exp. Med. 120, 105-120. Manak, R.C. (1982) Mitogenic responses of peripheral blood lymphocytes from pregnant and ovariectomized heifers and their modulation by serum. J. Reprod. Immunol. 4, 263-276. McCutchan, J.A., Katzenstein, D., Norquist, D., Chikami, G., Wunderlich, A. and Braude, A.I. (1978) Role of blocking antibody in disseminated gonococcal infection. J. Immunol. 121, 1884-1888.

325 Montgomery, W.P., Young Jr., R.C., Allen, M.P. and Harden, K.A. (1968) The tuberculin test in pregnancy. Am. J. Obstet. Gynecol. 100, 829-831. Morita, C.T., Goodman, J.W. and Lewis, G.K. (1985) Arsonate-specific murine T cell clones II. Delayed-type hypersensitivity induced by p-azobenzenearsonate-L-tyrosine (ABA-Tyr). J. Immunol. 134, 2894-2899. National Academy of Sciences (1977) Brucellosis research: an evaluation. Report of the subcommittee on Brucellosis Research, Washington, D.C. Nicoletti, P. (1983/1984) The use of a brucella protein antigen in dermal hypersensitivity as an adjunct method to diagnose bovine brucellosis. Vet. Immunol. Immunopathol. 5, 27-31. Rocklin, R.E., Kitzmiller, J.L. and Kaye, M.D. (1979) Imrnunobiology of the maternal-fetal relationship. Annu. Rev. Med. 30, 375-404. Sowa, B.A., Crawford, R.P., Heck, F.C., Williams, J.D., Wu, A.M., Kelly, K.A. and Adams, L.G. (1985) Size, charge, and structural heterogeneity of Brueella abortus lipopolysaccharides demonstrated by 2-D gel electrophoresis and monoclonal antibodies. Hybridoma 4, 81. Splitter, G.A. and Everlith. K.M. (1982) Suppression of bovine T- and B-lymphocyte responses by fetuin, a bovine glycoprotein. Cell. Immunol. 70, 205-218. Sumaya, C.V. and Gibbs, R.S. (1979) Immunization of pregnant women with influenza A/New Jersey/76 virus vaccine: Reactogenicity and immunogenicity in mother and infant. J. Infect. Dis. 140, 141-146. Tabatabai, L.B. and Deyoe, B.L. (1984) Characterization of salt-extractable protein antigens from Brucella abortus by crossed immunoelectrophoresis and isoelectric focusing. Vet. Micro. 9, 549-560. Tamura, S., Chiba, J., Kojima, A. and Uchida, N. (1983) Properties of cloned T cells that mediate delayed-type hypersensitivity against ovalbumin in mice. Cell. Immunol. 76, 156-170. Unanue, E.R., Bcller, D.I., Lu, C.Y. and Allen, P.M. (1984) Antigen presentation: comments on its regulation and mechanism. J. Immunol. 132, 1-5. Valdimarsson, H., Mulholland, C., Fridriksdottir, V. and Coleman, D.V. (1983) A longitudinal study of leucocyte blood counts and lymphocyte responses in pregnancy: a marked early increase of monocyte-lymphocyte ratio. Clin. Exp. Immunol. 53, 437-443. Weinberg. E.D. (1984) Pregnancy-associated depression of cell-mediated immunity. Rev. Infect. Dis. 6, 814-831. Winter, A.J., Verstreate, D.R., Hall, C.E., Jacobson, R.H., Castleman, W.L., Meredith, M.P. and McLaughlin, C.A. (1983) Immune response to porin in cattle immunized with whole cell, outer membrane, and outer membrane protein antigens of Brueella abortus combined with trehalose dimycolate and muramyl dipeptide adjuvants. Infect. Immun. 42, 1159-1167. Wu, A.M., Heck, F.C., Adams, E.G. and Jones, K. (1984). Immunochemical studies on the binding properties of Brueella abortus lipopolysaccharides to bovine precipitating antibodies. Mol. Immunol. 21, 1123-1129.