BIOLOGY OF REPRODUCTION 53, 814-819 (1995)

Stress-Induced Murine Abortion Associated with Substance P-Dependent Alteration in Cytokines in Maternal Uterine Decidua' Petra C. Arck,3 Fatima S. Merali, 3 Andrzej M. Stanisz, 4 Ron H. Stead,5 Gerard Chaouat,6 Justin Manuel, 3 and David A. Clark 2 ' 3

Departmentsof Medicine, Pathology, Obstetricsand Gynecology, 3 Molecular Virology-Immunology Program McMaster University, Hamilton, Ontario, Canada IntestinalDiseases Research Program,4 Departmentof Pathology McMaster University, Hamilton, Ontario, Canada Departments of Pathology and Medicine,5 McMaster University, Hamilton, Ontario, Canada INSERM CJF92 09,6 Cellular andMolecular Biology of the Materno-FetalRelationship Department of Obstetrics and Gynecology, HospitalAntoine Beclere, Clamart,France ABSTRACT Stress is known to induce abortions, but underlying mechanisms are unknown. Both alloimmunization and injection of antibody to the asialoGM1 determinant of natural killer cells have been shown to prevent stress-triggered abortion in mice. DBA2J-mated CBANJ female mice were used to investigate the influence of stress during early gestation on systemic hormone levels and on cytokines in the decidua that are thought to be relevant to abortion in nonstress-related murine abortion. Lowered levels of progesterone did not occur as a result of stress. In stressed mice, increased levels of the abortogenic cytokine tumor necrosis factor a (TNFa) were associated with decreased levels of pregnancy-protective transforming growth factor p2-related suppressive activity in uterine decidua. Inthe alloimmunized animals where stress failed to boost the abortion rate, these effects were abrogated. Production of TNFa may be stimulated by the neurotransmitter substance P (SP); after injection of an SP receptor antagonist or SP-antibody, stress failed to increase the abortion rate above the background level. The increased levels of TNFa we observed inthe stressed animals were completely abrogated in the animals that had received the SP receptor antagonist; stress also failed to decrease the pregnancy-protective suppressive activity in the decidua of these animals. The data indicate that stress may inhibit protective suppressor mechanisms and promote secretion of abortogenic cytokines such as TNFa via neurotransmitter SR INTRODUCTION

factor a (TNFa) and subnormal levels of transforming growth factor [32 (TGFP32)-mediated suppression in the decidua on Day 13.5 of gestation when the aborted embryos were counted [2]. As in mice, human recurrent pregnancy losses have been linked to the influence of stress, to increased NK cell activity, and to decreased TGFJ32-related suppressive activity in the decidua, and these losses may be preventable by immunization [7-9]. Experiments using mice are therefore potentially relevant to understanding effects of stress in the human species. To further explore the mechanisms responsible for stress-triggered abortions and the mechanism of the protective effect of leukocyte immunization, we studied the alterations produced by a 24-h period of ultrasonic sound stress initiated on Day 5.5 of pregnancy in DBA/2J-mated CBA/J female mice.

Stress interferes with reproduction: it adversely influences implantation and fetal growth, and leads to abortion [1-3]. Stress is also known to alter the immune system; in mammals, abortion may arise because of altered immune function [4-61. It is therefore logical to inquire whether stress interacts with the immune system in a way that may explain reproductive failure and to elucidate pathways through which stress leads to abortion. We recently demonstrated that allogenically pregnant mice subjected to a brief period of restraint or ultrasonic sound stress early in gestation showed a significant increase in the abortion rate [1, 2]. Stress-triggered abortions in mice could be corrected by immunization with allogenic lymphocytes, obtained from spleen cells of BALB/c males, bearing paternal major histocompatibility complex (MHC) antigens (H-2d), or by injection of antibody to the asialoGM1 determinant of natural killer (NK) cells. The stress-triggered abortions were associated with high levels of tumor necrosis

MATERIALS AND METHODS Animals

Accepted May 22, 1995. Received March 7, 1995. 'This research was supported by MRC grant MT-6447; Petra Arck received a fellowship from the Deutsche Forschungsgemeinschaft, and Gerard Chaouat was supported by CEFIPRA and INSERM, France. 2Correspondence: Professor David A. Clark, Rm. 3V39, 1200 Main Street West, Hamilton, Ontario, Canada L8N 3Z5. FAX: (905) 521-4971.

Male mice of the strains DBA/2J, CBA/J, and BALB/c, and female mice of the CBA/J strain were purchased from the Jackson Laboratory (Bar Harbor, ME). The animals were housed as described elsewhere [2]. Animal care and experimental procedures followed institutional ethics guidelines. 814

MECHANISMS OF STRESS-TRIGGERED ABORTIONS

Some groups of CBA/J females (H-2 k) were immunized by injection with 5 X 10 7 BALB/c (H-2d) spleen cells i.p. 7-14 days prior to mating with DBA/2J (H- 2d) males. The reason for including this treatment was to investigate whether immunization against paternal H-2d alloantigens that prevents stress-triggered abortion also altered cytokines in decidua similar to the effect seen with antagonists of the neurotransmitter substance P (SP). Some CBA/J mice were bred locally (CBA/J X CBA/J), and to deplete SP+ neurons, newborn females received s.c. injections of capsaicin (Aldrich Chemical Co., Milwaukee, WI) at 50 mg/kg BW, dissolved in distilled water containing 10% ethanol and 10% Tween 80. Mating and Stress Treatment After overnight cohabitation of the CBA/J females with DBA/2J males, the females with vaginal plugs (Day 0.5 of pregnancy) were segregated and randomized to "no treatment" or to "stress" groups. The two groups of females were respectively exposed to no treatment or to ultrasonic sound stress from a battery-powered rodent repellant device for a single 24-h period beginning at 1100 h on Day 5.5 of pregnancy. On different days after stress, the females were killed. The uteri were removed, and the total number of implantations and resorbing sites was recorded. Preparation of decidual supernatants and mixed-lymphocyte-culture cytotoxic T lymphocyte generation assays to measure the suppressive activity were performed according to our standard protocol as described elsewhere, and the supernatants were pooled per group [2]. 51 Cr-release assays for TNFa activity were performed by use of a 1/4 dilution of decidual supernatant incubated in 96-well flat-bottom culture plates (cat. #76.003.05; Linbro, McLean, VA) with 5 X 10 4 5 1Cr-labeled WEHI 164 target cells for 18 h at 370C in 5% CO2, and a standard curve was generated with recombinant murine TNFa (Genzyme, Boston, MA). Spontaneous and total release of the labeled target cells was determined by the addition of medium alone or 1% Nonidet P-40 (Sigma Chemical Co., St. Louis, MO) to wells containing target cells alone. Percentages of specific release of 8 replicates per group and the lytic activity were calculated as described elsewhere [2]. Some females received injections of SP receptor antagonist CP-96,345-1 (Pfizer Inc., Groton, CT) dissolved in PBS on Days 4.5 (200 jig), 5.5 (200 [tg), and 6.5 (400 pig) of pregnancy, and another group of mice received i.p. injections of antibody specific for SP (ICN, Costa Mesa, CA) dissolved in PBS at a dilution of 1:100; the stress in these groups was given on Day 5.5. Hormone Assays Whole blood samples were taken for all hormone measurements. On the assay day, the female mice received injections of 0.3 ml Somnotol (MTC Pharmaceuticals, Cam-

815

bridge, ON, Canada) and, after anesthesia, were heparinized with 0.1 ml heparin (Leo Laboratories Canada LTD, Ajax, ON). One minute after the heparin injection, the thorax of the mouse was opened carefully to avoid excessive bleeding. The rib cage was cut along the sternum, and the diaphragm was removed to access the heart. A 27-gauge needle on a 1-ml syringe was inserted into the left ventricle, and an average of 0.7-1 ml blood was taken. The mice were killed immediately after the needle was removed from the left ventricle by cervical dislocation. The blood of individuals in each group was pooled, and hormone RIAs were performed by the Section of Laboratory Medicine at McMaster University Hamilton according to standard procedures. Immunohistochemistry Immunohistochemical techniques were used to examine the presence and distribution of SP + cells. In some females, after the placentae were removed, the uteri were immediately fixed in 10% formalin. Within 24 h, the tissue was embedded in paraffin blocks. All immunohistochemistry was conducted at room temperature, with use of the streptavadin-peroxidase method (Histostain-SP Kit for mouse primary antibody; Zymed, S. San Francisco, CA). Two-millimeter sections were cut by ultramicrotome, floated in water for 1 min, and dried at room temperature for 1 h. Tissues were dehydrated in ethanol, cleared in xylene, and rehydrated in PBS. Slides were then placed in a humid chamber, and the nonspecific immunoglobulin binding was blocked with normal goat serum (Dimension Lab. Inc., Mississauga, ON, Canada). Then the sections were covered with antiserum against SP at a predetermined optimal dilution of 1:1000 for 120 min. The secondary antibody was incubated for 10 min, and then the conjugated streptavadin-peroxidase was applied. Color development in aminoethyl carbazole substrate solution was monitored microscopically, and the sections were counterstained with hematoxylin in Scott's tap water, mounted, and examined under a Leitz Othoplan microscope (Leica, Toronto, ON, Canada). Statistics Means and standard errors of the means (± SEM) were calculated for the resorption rates of groups of similarly treated mice killed on the same day of gestation. We used a multiple regression equation to statistically analyze the relationships of the resorption rate to gestation day. Significance was set at p ' 0.05. The statistical significance of differences in hormone levels (Fig. 2) and 51 Cr-release data (% suppression and % lysis WEHI 164/TNFa levels; Figs. 3 and 4) were tested by means of Student's t-test, with the results of the different stressed groups compared to the results of the control group of the same day of gestation. Significance was set at p ' 0.05. For statistical analysis of the

816

ARCK ET AL. -D

40

A

If.& BRAID I

A II

S

o NO STRESS * STRESS DAY 5.5 * STRESS DAY 5.5 BALB/C

w

z

0so

6C

u 0

4C

a

20

a

0 13

cc IC, 0 0

2

0

20

a 0 10

I

I

I

I

I

I

11 S 0 E .

0 /

* STRESS DAY 5.5 o NO STRESS

E r_

(A

a.

10(

0

30

P

L

0.

!.,

/

5

5

6 5 ,6 r

6.5

7.5

4

a

48s,

1

~~

8.5

~ 9.5

*

6

Z

7

~~~a3 5

10

*

6.

' i

i

10.5 11.5

12.5 13.5

e, 23o

40

CD

GESTATION DAY

9

z

7

o.