Government Support Support for this work was provided by grants HD00815, HD23547 and HD23775 from the National Institutes of Health.

Field of the Invention This invention relates to methods and compositions for diagnosing and treating immunologic reproductive failure. The methods include the administration of an immunomodulating agent for suppressing a cellular immune response.

Background of the Invention Pregnancy is a unique immunologic state in which a natural homeostasis exists between antigenically different tissues (Feinberg, B. and Gonik, B., Clinical Obstetrics and Gynecology 34(1) 3-16 (1991). In one out of five pregnancies, this homeostatic state is disrupted and the fetus is prematurely aborted. In one out of 300 pregnancies, this homeostatic state is repeatedly disrupted, i.e., two or more pregnancy losses before 20 weeks gestation, and the fetus is prematurely aborted. Such recurrent spontaneous abortions are attributed to a variety of causes.

Chromosomal defects are found in 3-7 % of couples with a history of recurrent spontaneous abortion and are the only well-established cause of recurrent abortion. Non-genetic causes of recurrent abortion are difficult to ascribe, although numerous associations have been made with anatomic, endocrine, infectious, and humoral immune abnormalities. To date, the basis for recurrent spontaneous abortion remains unexplained in 40% to 60% of couples, although immunologic mechanisms have been proposed in many unexplained cases of recurrent spontaneous abortion.

The human immune system has two effector arms: humoral (antibody mediated) immunity that is highly effective at neutralizing pathogens, and cellular (T cell mediated) immunity that principally functions to destroy foreign and infected cells. This dichotomous nature of the immune system is at least in part determined by regulatory CD4 T helper/inducer lymphocytes. Following activation by antigen, CD4 T lymphocytes produce one of two distinctive cytokine profiles which has led to their classification as T helper 1 (T„l) and T helper 2 (T„2) cells (Mosmann T. et al . ,

1986, J. Immunol. 136:2348; Cherwinski, H. et al. , 1987, J. Exp. Med. 166:1229; Kurt-Jones, E. et al. , 1987, 166:1774). The different cytokines produced by these cells lead to differences in immune function (Cher, D. et al. , 1987, J. Immunol. 138:3688; Mosmann, T. et al. , 1989, Ann. Rev. Immunol. 7:145). T _H 1 cells primarily secrete interferon-gamma (IFN-gamma), but also interleukin-2 (IL-2) and tumor necrosis factor-beta (TNF-beta, Lymphotoxin) , and induce cellular immunity. T„2 cells primarily secrete interleukin-4 (IL-4), but also IL-5 and IL-10, and downregulate cellular immunity while playing a major role in the induction of antibody responses mediated by plasma cells (Kurt-Jones, E. et al. , 1987, 166:1774; Cher, D. et al . ,

1987, J. Immunol. 138:3688; Mosmann, T. et al., 1989, Ann. Rev. Immunol. 7:145; Mosmann, T. et al. , 1987, Immunol. Today 8:223; Abbas, A. et al . , 1990, J. Immunol. 144:2031; Powrie, F. et al., 1993, Immunol. Today 14:270). The cytokine TNF- can be produced during both T„l and T„2 immune responses, although levels are higher in T„l responses and this cytokine is known for cytolytic effects that contribute to the efficacy of cellular immunity (Mosmann, T. et al.,

1989, Adv. Immunol. 46:11; Romagnani, S. et al . , 1992, Int.

Arch. Allergy Immunol. 4:279). Most of the original work on T„l and T„2 lymphocyte subsets was performed with mouse helper T cell clones, however, it is now recognized that human CD4 T cells secrete similar cytokine profiles and can

also be classified into T„l and T _H 2 subsets (Powrie, F. et al . , 1993, Immunol. Today 14:270; Romagnani, S. et al. , 1992, Int. Arch. Allergy Immunol. 4:279; Romagnani, S. et al . , 1991, Immunol. Today 8:256).

The precise nature and extent of immunologic mechanisms for recurrent abortion are not well understood. Although it is generally known that the cellular mediators of the cellular immune response, T lymphocytes and macrophages, are present in the uterine endometrium in the secretory phase of the menstrual cycle (the time of implantation) and throughout pregnancy, the role of these cells in mediating spontaneous abortion has not been precisely defined. As a consequence, therapies for treating immunologic spontaneous abortion have not been adequately explored.

Summary of the Invention

The present invention provides methods and compositions for diagnosing and preventing immunologic reproductive failure.

According to one aspect of the invention, a method for diagnosing a predisposition to immunologic reproductive failure is provided. The method involves contacting a sample containing a plurality of leukocytes derived from the mammal with a reproductive antigen under conditions to stimulate release by the plurality of leukocytes of an extracellular embryotoxic factor, determining the concentration of the embryotoxic factor in the sample and diagnosing a predisposition to immunologic reproductive failure in the mammal by comparing the sample embryotoxic factor concentration to the concentration of embryotoxic factor present in at least one standard. The standard can be a "positive" standard (which contains a concentration of the embryotoxic factor indicative of a predisposition to immunologic reproductive failure) or a "negative" standard (which contains a concentration of the embryotoxic factor indicative of the absence of a predisposition to immunologic

reproductive failure) . A diagnosis of a predisposition to immunologic reproductive failure is made if the sample embryotoxic factor concentration is elevated compared to the negative standard or is substantially the same as the positive standard.

According to another aspect of the invention, a method for diagnosing a predisposition to immunologic reproductive failure in a non-pregnant mammal is provided. The method involves obtaining from the mammal a leukocyte secretion-containing sample, determining the concentration of an embryotoxic factor in the sample and diagnosing a predisposition to immunologic reproductive failure by comparing the sample embryotoxic factor concentration to the concentration of embryotoxic factor present in at least one standard as described above. The sample concentration of embryotoxic factor is determined by, for example, an immunoassay that specifically recognizes the embryotoxic factor, an embryo development bioassay, a trophoblast assay or by a lymphocyte proliferation assay.

According to yet another aspect of the invention, a third method for diagnosing a predisposition to immunologic reproductive failure is provided. The method involves stimulating a sample containing a plurality of leukocytes as described above to release a cytokine, determining the concentration of the cytokine to obtain a sample cytokine concentration and comparing the sample cytokine concentration to the concentration of cytokine present in at least one standard. The standard can be a positive standard (which contains a concentration of cytokine indicative of a predisposition to immunologic reproductive failure) or a negative standard (which contains a concentration of cytokine indicative of the absence of a predisposition to immunologic reproductive failure) . For a cytokine that is a T„-l cytokine, a diagnosis of immunologic reproductive failure is made if the sample cytokine concentration is elevated compared to the negative cytokine standard or is

substantially the same as the positive cytokine standard. For a cytokine that is a T _H ~2 cytokine, a diagnosis of immunologic reproductive failure is made if the sample cytokine concentration is reduced compared to the negative cytokine standard or is substantially the same as the positive cytokine standard. Thus, a diagnosis of immunologic reproductive failure can be made by observing an increase in the concentration of Tr„t-l cytokines, e.g., interferon-gamma, tumor necrosis factor-alpha, tumor necrosis factor-beta or by observing a reduction in the level of certain cytokines, e.g., T„-2 cytokines such as interleukin-4 and interleukin-10, in stimulated leukocyte- or leukocyte secretion-containing samples. Accordingly, a determination of the ratio of embryotoxic factor concentration to the concentration of, for example, interleukin-4 or interleukin-10, also can be used to predict a woman's predisposition to immunologic reproductive failure.

According to another aspect of the invention, a method for preventing immunologic reproductive failure is provided. The method includes administering a therapeutically effective dose of an immunomodulating agent to a mammal diagnosed as having a predisposition to immunologic reproductive failure. According to one embodiment, the method further includes isolating from the mammal a leukocyte secretion-containing sample; determining the concentration of embryotoxic factor present in the sample; and administering one or more subsequent doses of the immunomodulating agent. Alternatively, a leukocyte-containing sample can be isolated from the mammal following initial administration of the immunomodulating agent. The leukocyte-containing sample then is "stimulated", i.e., contacted with a plurality of reproductive antigens to release embryotoxic factor(s) in vitro. The amount of released factor then is determined using one or more of the above-mentioned assays. The

subsequent dose of the immunomodulating agent is selected to cause a reduction in the amount of the embryotoxic factor released in vivo by the leukocytes.

The immunomodulating agent can be an agent that downregulates a T„-l immune response (e.g., a T _H ~2 cytokine or an antibody which specifically binds to or otherwise regulates a T„-l cytokine, thereby abrogating the cytokine's biological activity) or an agent that upregulates a T _u π-2 immune response or otherwise increases the concentration of a T _H ~2 cytokine in vivo. Preferably, the immunomodulating agent specifically modulates a cellular immune response to a reproductive antigen. Accordingly, the preferred immunomodulating agent is a vaccine containing a reproductive antigen (or an antigen structurally-related thereto) in an adjuvant that downregulates a T„-l immune response or in an adjuvant that upregulates a Tr„l-2, response. Nonspecific immunomodulating agents also are useful for the purposes of the invention, e.g., progesterone, glucocorticoids, cyclosporins, nifidipine, pentoxiphylline transforming growth factor-beta, intravenous immunoglobulin, anticytokines, cytokine receptor blockers, and antibodies to cytokine producing cells. Progesterone is a preferred nonspecific immunomodulating agent.

Also provided are kits for determining the concentration of an extracellular embryotoxic factor in a sample. The kits include a first vial containing an antibody to the embryotoxic factor; a second vial containing a positive or negative standard; and instructions for using the antibody to determine the concentration of embryotoxic factor in the sample and for comparing the sample embryotoxic factor concentration to the standard embryotoxic factor concentration to determine whether the mammal has a predisposition to recurrent immunologic reproductive failure. Preferably, the first vial includes a plurality of antibodies, each antibody being directed against a different embryotoxic factor. In this manner, the kits can be used to

detect a plurality of different embryotoxic factors. According to one embodiment, the second vial contains a positive standard containing an amount of the embryotoxic factor indicative of a predisposition to immunologic reproductive failure and the kit further includes a third vial containing a negative standard. The negative standard contains an amount of the embryotoxic factor indicative of the absence of a predisposition to immunologic reproductive failure. Kits that are to be used to determine the amount of embryotoxic factor present in a leukocyte-containing sample further contain an additional vial including a plurality of reproductive antigens for stimulating the leukocytes to release the embryotoxic factor(s) in vitro.

According to yet another aspect of the invention, a vial containing an amount of one or more isolated reproductive antigens capable of stimulating a plurality of leukocytes to release an embryotoxic factor also is provided. Exemplary reproductive antigens include sperm antigens, rophoblast antigens and choriocarcinoma cell line-derived antigens. Preferably, the vial containing the reproductive antigens targeted for administration to a mammalian recipient (i.e., for stimulating leukocytes in vivo) further includes a pharmaceutically acceptable carrier. It is believed that administration of such reproductive antigens (or structurally-related antigens) in the appropriate type of adjuvant (e.g., an adjuvant that downregulates a T„-l cytokine immune response or that upregulates a T„-2 cytokine response to the reproductive antigen) to a patient can reduce a T-.-1 cytokine immune response, thereby preventing immunologic pregnancy loss.

These and other aspects of the invention as well as various advantages and utilities will be more apparent with reference to the detailed description of the preferred embodiments.

Brief Description of the Drawings

Fig. 1 shows the results of an active immunization experiment with various control groups indicating reduced fertility and increased fetal resorptions in sperm-immunized mice.

Fig. 2 shows the effects of passive transfer of T lymphocytes from immunized mice (from experiment depicred in Fig. 1), on fertility of recipient mice.

Fig. 3 shows the mean numbers + S.D. of T lymphocyte subpopulations and macrophages in uterine sections of mice after immunizations (from experiment depicted in Fig. 1).

Fig. 4 shows the mean numbers + S.D. of T lymphocyte subpopulations and macrophages in uterine sections of mice after passive transfer of T lymphocytes from immunized mice (same experiment as depicted in Figs. 1-3).

Fig. 5 shows preimplantation embryo recovery in mice following allogenic trophoblast immunization. Immunization groups: f, saline; g, adjuvant; adjuvant plus trophoblast.

Fig. 6 shows the number of viable offspring (1) and fetal resorption (2) in mice following allogenic trophoblast immunization. The groups are as described in Fig. 4.

Fig. 7 shows the dose response curve of Jeg-3 antigen in a lymphocyte proliferation assay.

Fig. 8 shows the trophoblast antigen fractions responsible for lymphocyte proliferation (SI>3).

Fig. 9 shows the correlation between lymphocyte proliferation (SI) and embryotoxic factors (% blastocysts) in response to trophoblast antigen stimulation.

Detailed Description of the Invention The prior art reveals the difficulties of diagnosing non-genetic causes of recurrent spontaneous abortion. The present invention provides a solution to this problem by disclosing a method for diagnosing in a patient a predisposition to immunologic reproductive failure. The phrase "immunologic reproductive failure" refers to those

reproductive failures (abortions occurring at any time post-conception) that are attributable to a cell-mediated immune response. Accordingly, the accurate diagnosis of a predisposition to immunologic reproductive failure is essential for designing effective therapies for preventing or reducing the occurrence of such abortions.

According to one aspect of the invention, a method is provided for diagnosing in a mammal a predisposition to immunologic reproductive failure. The method involves contacting a sample containing a plurality of leukocytes derived from the mammal with a reproductive antigen under conditions to stimulate release by the plurality of leukocytes of an extracellular embryotoxic factor, determining the concentration of the embryotoxic factor in the sample to obtain a sample embryotoxic factor concentration, and diagnosing a predisposition to immunologic reproductive failure in the mammal by comparing the sample embryotoxic factor concentration to the concentration of embryotoxic factor present in at least one standard (described below) .

As used herein, "leukocytes" embraces lymphocytes (e.g., T-lymphocyteε and natural killer cells) and macrophages, i.e., the cellular mediators of the cellular immune response). As used herein, the term "leukocyte secretion-containing sample" refers to a preparation containing soluble factors that have been released by lymphocytes or macrophages, in vivo or in vitro. These extracellular factors are released by the leukocytes in response to stimulation by reproductive antigens. Exemplary leukocyte and leukocyte secretion-containing samples include peripheral blood and serum, peritoneal fluid, endometrial tissue, vaginal secretions and saliva. Methods for isolating the above-described samples from a patient are known to one of ordinary skill in the art. The preferred leukocyte (e.g., lymphocyte/macrophage) and leukocyte secretion-containing sample is peripheral blood. The sample containing a

plurality of leukocytes derived (i.e., obtained) from the mammal can be used in the methods of the invention with or without prior culturing. Preferably, the leukocyte-containing sample is isolated from the mammal, cultured and stimulated in vitro (by contacting the leukocytes with a plurality of reproductive antigens) to release one or more embryotoxic factor(s). By contacting, it is meant that the leukocytes are exposed to the plurality of reproductive antigens, under conditions such that the leukocytes are stimulated to release soluble factors. Exemplary conditions are provided in the Examples.

The term "reproductive antigens" refers to sperm antigens, trophoblast antigens, and/or choriocarcinoma cell line-derived antigens. The sperm antigens are isolated from mammalian (e.g., human) motile sperm from fertile donors. Antigen isolation from tissue or cells involves subjecting the tissue or cells to homogenization, followed by washing and dividing the preparation into aliquots containing about 300 ug/ml protein, (see also Examples 1, 3 and 5). Endometrial antigens may be isolated from eutopic (intrauterine) or ectopic (endometriosis) endometrium. The preferred reproductive antigens are derived from a trophoblast cell line, such as choriocarcinoma cell lines BeWo or Jeg-3.

As disclosed in Example 1 ("Diagnosis of Immunologic Recurrent Spontaneous Abortion"), both sperm and trophoblast antigens stimulate peripheral blood lymphocytes and macrophages from women diagnosed with recurrent reproductive failure to secrete embryotoxic factors in vitro. Embryotoxicity was measured by bioassay, e.g., observing the adverse effect of the released factors on embryo development and trophoblast proliferation in vitro.

The trophoblast antigens used in Example 1 were derived from the choriocarcinoma cells lines BeWo and Jeg-3. These cell lines were chosen as the trophoblast antigen model for three reasons: (1) they are homogeneous trophoblast cell

lines uncontaminated by other cellular constituents such as stromal mesenchymal cells or lymphoid and myeloid populations that are always associated with normal placental trophoblast; (2) like normal trophoblast, they are devoid of classic major histocompatability complex class I and II antigens; and (3) they are established antigenic models of early normal trophoblast because of their invasive capabilities and cell surface antigen profiles that are similar to those of early normal human trophoblast. There were no observed differences between BeWo and Jeg-3 antigen preparations in terms of their ability to stimulate embryotoxic factor release from activated leukocytes of susceptible individuals. Accordingly, these two trophoblast antigen sources were used interchangeably for leukocyte stimulation in Examples 1-3.

The reproductive antigen(s) derived from the Jeg-3 cell line have been fractionated to determine their cellular location (see Example 5). Further purification of the Jeg-3 antigen is within the ordinary skill in the art, using conventional purification and characterization techniques. Thus, for example, gel filtration chromatography is used initially to fractionate a preparation based upon differences in molecular size and various additional chromatographic separation techniques (e.g. , ion-exchange HPLC) are used to isolate reproductive antigenic proteins that are capable of stimulating leukocytes to release an embryotoxic factor in vitro. The Jeg-3 reproductive antigen proteins are isolated and cloned using procedures known to the artisan of ordinary skill in the art. The isolation and cloning of a trophoblast membrane protein antigen from human placenta has recently been reported (see International Patent Application having publication number WO 93/06857 and publication date April 15, 1993, the contents of which application are incorporated herein by reference, for disclosure of a trophoblast membrane expressed protein derived from normal human placenta) .

The relative amount or absolute concentration of embryotoxic factor present in a reproductive antigen-stimulated leukocyte-containing sample or in a leukocyte secretion-containing sample is determined using one or more of the following exemplary assays: an immunoassay that specifically detects the embryotoxic factor, an embryo development assay, a trophoblast proliferation assay and/or a lymphocyte proliferation assay (see Examples 1, 4 and 5 for assay protocols). To diagnose a predisposition to immunologic reproductive failure, the "sample embryotoxic factor concentration" (i.e., the concentration of embryotoxic factor in the sample) is compared to the concentration of embryotoxic factor present in at least one standard. As used herein, a "positive standard" refers to a standard which contains a concentration of the embryotoxic factor that is indicative of a predisposition to immunologic reproductive failure. Conversely, a "negative standard" refers to a standard which contains a concentration of the embryotoxic factor that is indicative of the absence of a predisposition to immunologic reproductive failure.

Although various assays can be used to determine the concentration of embryotoxic factor, the preferred assays are im unoassays which specifically determine the amount of an embryotoxic factor present in the sample (see e.g., Example 4) and the lymphocyte proliferation assay (see, e.g. , Example 5). In addition to the specific immunoassays disclosed in Example 4, the preferred immunoassays for the purposes of the invention are those which specifically recognize the following antigens: gamma-interferon, tumor necrosis factor-alpha, interleukin-1, interleukin-2, interleukin- , inte leukin-6, interleukin 10 and transforming growth factor-beta. As discussed in more detail below, in some instances, elevated levels of cytokines (e.g., T _u rl-1 cytokines) are indicative of immunologic reproductive

failure. In yet other instances, reduced levels of some cytokines (e.g., T-.-2 cytokines such as interleukins 4 and 10), are indicative of this condition.

As used herein, the phrase "elevated concentration" in reference to an embryotoxic factor or cytokine refers to a concentration which is elevated with respect to the concentration of the embryotoxic factor or cytokine in the negative standard. In general, diagnosis of a predisposition to immunologic reproductive failure is made by an immunoassay specific for the embryc oxic factor if the ratio of the sample embryotoxic factor concentration to the negative standard embryotoxic factor concentration is at least about 2:1, i.e., the sample embryotoxic factor concentration is elevated two-fold compared to the concentration of embryotoxic factor in the negative standard. Alternatively, or additionally, a diagnosis of a predisposition to immunologic reproductive failure is made if the sample embryotoxic factor concentration is substantially the same as the concentration of embryotoxic factor present in the positive standard. By "substantially the same" it is meant that the sample embryotoxic factor concentration falls within the margin of error of the concentration of embryotoxic factor in the positive standard for the particular assay performed.

Immunoassays have been used to determine the concentrations of three different embryotoxic factors: interferon-gamma (IFN-gamma), tumor necrosis factor-alpha (TNF-alpha) and tumor necrosis factor-beta (TNF-beta) (Example 4). IFN-gamma was detected in reproductive antigen-stimulatei leukocyte supernatants from 125 of 255 recurrent abortion patients and significantly correlated with embryotoxicity. TNF-alpha and TNF-beta also were detected. Thus, Applicants have discovered that a majority of women with unexplained recurrent abortion release T„-l cytokines following stimulation with trophoblast extract reproductive antigens. In contrast, antigen-stimulated lymphocytes from

parous women with normal reproductive histories and men release T„-2 cytokines (Example 4). Although not intending to be bound to a particular theory, Applicants believe that

T„π-l cytokines play a role in reproductive failure by mediating a cellular immune response to reproductive antigens, whereas Tr.i,-2 cytokines are associated with a successful normal pregnancy and presumably play a role in downregulating a cellular immune response to reproductive antigens in vivo.

As used herein, "T.,-1 cytokines" refers to the cytokines that are released by leukocytes in vivo as part of a T„-l type immune response. Exemplary T„-l cytokines include interferon-gamma, TNF-beta, IL-2, TNF-alpha, and potentially IL-12. As used herein "T _H -2 cytokines" refers to the cytokines released by leukocytes in vivo as part of a T„-2 type immune response. Exemplary T„-2 cytokines include IL-4, IL-5, IL-6, IL-10 and potentially TGF-beta and colony stimulating factors. Accordingly, the invention is not limited in scope to cytokines that are presently known to be T„-l or T„-2 cytokines but rather, embraces cytokines that are identified in future to fall within these categories.

Alternatively, the diagnosis of a predisposition to immunologic reproductive failure can be made by performing a lymphocyte proliferation assay and observing lymphocyte proliferation in response to stimulating the lymphocytes with the above-described trophoblast antigen. Example 5 demonstrates that antigen-stimulated lymphocyte proliferation was significantly higher in women with recurrent abortion of unknown etiology than in fertile controls and that lymphocyte proliferation significantly correlated with embryotoxic factor activity in stimulated culture supernatants. A diagnosis of a predisposition to immunologic reproductive failure was made if the stimulated leukocyte supernatant had a Stimulation Index greater than about 3.0. Fertile control subjects were observed to have a stimulation index that was less than 3.0. As used herein in reference to the lymphocyte

proϋferation assay, the term "Stimulation Index" refers to the ratio of counts per minute (cpm) observed for unstimulated cultures to the cpm observed for antigen-stimulated cultures.

The term "embryotoxic factor" refers to an extracellular factor that is released by leukocytes (i.e., lymphocytes and/or macrophages) in response to stimulation by a reproductive antigen. The embryotoxic factor is further characterized as being toxic to developing mouse embryos and human placental cell lines. Toxicity is determined using the above-recited bioassays. Applicants disclose herein at least three cytokines that are embryotoxic factors.

A preparation containing one or more embryotoxic factors has been partially purified and characterized (see Example 1). The preparation is heat-sensitive and, based upon molecular weight dialysis experiments, contains a factor(s) having a molecular weight between about 10 and about 30 kD. By "heat-sensitive", it is meant that the ability of the embryotoxic factor to inhibit mouse embryo development is abrogated following exposure of the factor to 56°C for 1 hour. Passage of the embryotoxic factor preparation through an affinity column containing anti-interferon gamma-coated beads (Endogen, Boston, MA) abrogated approximately 80 % of the embryotoxic factor activity, as measured in the above-mentioned embryo mouse development bioassay (Example 1), suggesting that IFN-gamma is indeed an embryotoxic factor. An immunoassay which specifically detects interferon-gamma has been used to definitively demonstrate the presence of interferon-gamma in trophoblast antigen-stimulated leukocytes derived from women having a history of recurrent spontaneous abortion (Example 4). Stimulated leukocytes from fertile parous controls did not release IFN-gamma in response to trophcblast antigen stimulation.

Immunoassays also have been used to detect an elevated concentration of two other T„-l cytokines (TNF-alpha, TNF-beta) in stimulated leukocytes from women having a history of recurrent spontaneous abortion compared to stimulated leukocytes from fertile controls (Example 4). Accordingly, at least gamma-interferon (one form of which has a molecular weight of 17,000) or a soluble molecule which closely associates (or works in synergy) with gamma-interferon, TNF-alpha and TNF-beta have been identified as embryotoxic factors. Soluble products of activated macrophages (e.g., tumor necrosis factor (TNF)) also have been reported to have toxic effects on preimplantation embryos and trophoblast cell lines in vitro. Accordingly, the term embryotoxic factors is not limited to gamma-interferon, but rather embraces any and all soluble factors including cytokines (e.g., T„-l cytokines) that are involved in a cell mediated immune response to a reproductive antigen. Exemplary cytokines that modulate a cellular immune response are shown in Table I .

Example 4 also demonstrates that trophoblast antigen-stimulated leukocytes from women with recurrent reproductive failure do not release T„-2 cytokines such as IL-2, IL-4 and IL-10 but that trophoblast antigen-stimulated leukocytes obtained from fertile parous controls do release these T„-2 cytokines, as measured in immunoassays which specifically detect each of the above-mentioned cytokines. The implications of this discovery with respect to methods and compositions for diagnosing and/or preventing immunologic reproductive failure are discussed below.

According to another aspect of the invention, yet another method for diagnosing a predisposition to immunologic reproductive failure is provided. The method involves stimulating a sample containing a plurality of leukocytes as described above to release at least one cytokine, determining the concentration of the cytokine to obtain a sample cytokine concentration and comparing the sample cytokine concentration

TABLE 1

Immunorequlatory Cytokines

Cytokine Function

Cytokines involved in Cell Mediated Immune Responses

Interleukin-2 Proliferation of activated B and T cells

Migration Inhibitory Factor Inhibition of macrophage migration

Cytotoxic and Cytostatic Kill or inhibit growth of susceptible

Factors target cells Leukocyte Inhibitory Factor Inhibits migration of neutrophils Tissue Factor Procoagulant activity Macrophage-Activating Activates macrophages against tumor

Factors (GM-CSF, cells and intracellular organisms

Interferon-γ,

Interleukin-3

Interleukin 4) Chemotactic Factors Selectively mobilize and attract onocytes, neutrophils, eosinophils, or basophils to inflammatory site

Cytokines involved in B cell regulation and antibody production

Interleukin-4 Activates resting B and T cells Inhibits antibody class switching from IgM

Interleukin-5 B cell growth factor Induces eosinophil differentiation Interleukin-6 Induces terminal maturation of

B cells Activates mitogen stimulated T cells Induces T cell proliferation

Interleukin-7 Induces proliferation, but not maturation, of early B cells Stimulates proliferation of early

T cells

Miscellaneous Cytokines

Colony-Stimulating Factors • Stimulates granulocyte and monocyte (G -CSF, Interleukin-3) differentiation

• Activates mature macrophages to kill tumor cells and certain microorganisms

Interleukin-1 • Aids in stimulation of T cell IL-2 production

• Endogenous pryogen, glucocorticoid synthesis, prostaglandin release, collagenase production, synthesis of acute phase reactants, chemotaxic activity

Tumor Necrosis • Induction of hemorrhagic necrosis Factor/Cachectin of certain tumors

• Causes cachexia, shock

• Enhances eosinophil ADCC

• Enhances neutrophil adhesion to endothelial cells

• Highly suppressive of B and T cell proliferation

• Potent chemoattractant for macrophages

• Render cells resistant to viral infection

• Promotes B cell differentiation

to the concentration of cytokine present in at least one standard. The standard can be a positive standard (which contains a concentration of cytokine indicative of a predisposition to immunologic reproductive failure) or a negative standard (which contains a concentration of cytokine indicative of the absence of a predisposition to immunologic reproductive failure) . For a cytokine that is a T„-l cytokine, a diagnosis of immunologic reproductive failure is made if the sample cytokine concentration is elevated compared to the negative cytokine standard or is substantially the same as the positive cytokine standard. For a cytokine that is a T„-2 cytokine, a diagnosis of immunologic reproductive failure is made if the sample cytokine concentration is reduced compared to the negative cytokine standard or is substantially the same as the positive cytokine standard. Thus, a diagnosis of immunologic reproductive failure can be made by observing an increase in the concentration of T..-1 cytokines, e.g., interleukin-2, interferon-gamma, tumor necrosis factor-alpha, tumor necrosis factor-beta or by observing a reduction in the level of certain cytokines, e.g., T„-2 cytokines such as interleukin-4 and interleukin-10, in stimulated leukocyte- or leukocyte secretion containing samples.

The invention also provides a method for preventing immunologic reproductive failure (see, e.g., Example 2 "Immunosuppressive Therapy in Pregnant Women for Prevention of Immunologic Spontaneous Abortion"). The method involves selecting a mammal diagnosed as having a predisposition to immunologic reproductive failure and administering to the mammal a therapeutically effective dose of an immunomodulating agent to prevent an immunologic reproductive failure. Preferably, the method for preventing immunologic reproductive failure further includes performing one or more of the above-described methods for diagnosing the condition for a leukocyte- or leukocyte secretion-containing sample as described above and administering to me mammal a subsequent

dose of the immunomodulating agent to cause a reduction in the amount of embryotoxic factor released in vitro or in vivo by the mammal's leukocytes.

In contrast to the time-consuming bioassays of the prior art, immunoassay quantitation of the embryotoxic factors (see e.g., Example 4) permits the rapid e ^v aluation of the efficacy of an immunomodulating treatment regimen. Accordingly, the instant invention provides a method for preventing immunologic reproductive failure, which method includes administering at least one subsequent dose of the immunomodulating agent to the patient. The instant invention is not limited to administration of a single immunomodulating agent, but rather embraces the administration of one or more agents, depending upon the particular immunologic response of the patient to immunomodulating agent drug therapy. This iterative treatment process, i.e., administration of an immunomodulating agent followed by determination cf the concentration of embryotoxic factor released by patient leukocytes in response to immunomodulating agent drug therapy, allows the tailoring of treatment to the cellular immune response for each patient.

As used herein, "immunomodulating agent" refers to an agent capable of modulating a cellular immune response and includes agents which directly or indirectly modulate the effective embryotoxic factor concentration in vivo. The immunomodulating agent can be a nonspecific immunomodulating agent (i.e., not targeted to modulating an immune response to a particular target antigen) that downregulates a T-,-1 immune response or that upregulates a T„-2 immune response. Exemplary nonspecific immunomodulating agents include glucocorticoids, cyclosporins, nifidipine, pentoxiphylline and progesterone. Alternatively, the immunomodulating agent can be a specific immunomudulating agent that modulates the cellular immune response to a specific reproductive antigen. According to a particularly preferred embodiment, the specific immunomodulating agent is

a vaccine including a reproductive antigen contained in an adjuvant. An adjuvant is selected that downregulates a T„-l type immune response or that upregulates a T„-2 type immune response to the reproductive antigen in vivo. Oral vaccines for modulating a cellular immune response to a specific antigen have been described (see, e.g., PNAS, USA 91:437-438 (1994); Immunology Today 12:383-385 (1991); Cellular Immunology 131:302 (1990); PNAS, USA 89:421-425 (1992); Science 259:1321-1324 (1993); and Science 261:1727-1730 (1993), the contents of which references are incorporated herein by references). Accordingly, a particularly preferred cellular vaccine of the invention is an oral vaccine prepared by placing the Jeg-3 antigen in adjuvants such as those described in the above-identified references. Adjuvants which downregulate a T-,-1 type response and/or upregulate a T„-2 type response can be selected by determining the T„-l and/or T„-2 cytokine profile following immunization of, for example, an animal with a vaccine containing a test antigen (e.g., BSA, reproductive antigen) contained in the test adjuvant. Exemplary adjuvants that upregulate a T _H -2 type response (and thereby downregulate a T-,-1 response) include alum and squalene in oil. Additional adjuvants which can be screened for their ability to downregulate a T„-l type response and/or to upregulate a T„-2 type response include the ISCOMS (Morein, B., et al. , Nature (Lond.). 308:457-459 (1984)), cholera toxin adjuvants (Quiding, M. , et al. , J. Clin. Invest. 88:143-148 (1991)) and complete Freunds adjuvant. The ISCOMS are prepared by removing detergent in a controlled fashion from a mixture of cholesterol, protein, phospholipid, detergent and Quil-A (e.g., by dialysis and cent ifugation on a Quil-A-containing sucrose gradient). ISCOM formation is confirmed by negative contrast electron microscopy and by their distinctive sedimentation constant (19S) in a sucrose gradient. Quil-A is a saponin extracted from the bark of the tree Quillaja saponaria. Purification

of these saponins and their use as adjuvants is described in

U.S. Patent No. 5,057,540, issued to Kensil et al. , the contents of which patent are incorporated herein by reference.

As used herein, "effective cytokine concentration" refers to the concentration of cytokine that is available for binding to cytokine receptors, i.e., the concentration of cytokine that is capable of triggering a cellular immune response. Thus, immunomodulating agents embrace agents which function by (1) reducing T„rl-l cytokine release from leukocytes; (2) reducing the concentration of receptors capable of binding to the T„-l cytokines; (3) binding directly to T„-l cytokines, thereby preventing cytokine binding to receptors; (4) competing with the T„-l cytokines for binding to cytokine receptors; as well as (5) agents which modulate the concentration of any of the above (e.g., T-,-2 cytokines) . Accordingly, immunomodulating agents include agents which are known in the art for their ability to suppress an immune response (e.g., progesterone), as well as TGF-beta and antibodies to the embryotoxic cytokines and/or antibodies to the embryotoxic cytokine receptors (e.g., antibodies to gamma-interferon, tumor necrosis factor-alpha, interleukin-2 and interleukin-6 or antibodies to cytokine producing cells such as CD-3 and CD56 cells or to their receptors). In a preferred embodiment, the immunomodulating agent is capable of modulating the cellular immune response in a localized area, i.e., the area in fluid or tissue communication with fetal cells, as distinguished from a humoral immune response.

The immunomodulating agents are administered in therapeutically effect amounts. A therapeutically effective amount is that amount which is sufficient to reduce the level of embryotoxic factor(s) to an amount(s) that is insufficient to cause a reproductive failure. The effective amount of agent will depend upon the clinical condition of the subject being treated. A therapeutically effective amount can be determined in a number of ways using medical tech.iiques

customary to one of ordinary skill in the art. For example, different amounts of immunomodulating agent are administered to mammals predisposed to reproductive failure, followed by assaying leukocyte- and/or leukocyte secretion-containing samples obtained from the subject. Exemplary assays for determining the level of embryotoxic factor in such samples are provided in the Examples. The therapeutically effective dose of immunomodulating agent is selected which reduces the level of extracellular embryotoxic factor to that level observed in the samples of mammals not having a predisposition to reproductive failure (i.e., a "normal" level). Such levels are considered non-toxic. Likewise, embryotoxic levels that are above normal levels, but which are insufficient to cause reproductive failure, also are considered non-toxic.

The selection of a therapeutically effective dose of the immunomodulating agent to prevent immunologic reproductive failure is made in accordance with standard procedures known to one of ordinary skill in the art, taking into consideration the patient's clinical condition. Such amounts will depend, of course, on the particular condition being treated, the severity of the condition, and individual patient parameters including age, physical condition, size, weight and concurrent treatment. These factors are well known to those of ordinary skill in the art and can be addressed with no more than routine experimentation. It is preferred generally that a maximum dose be used, that is, the highest safe does according to sound medical judgment. It will be understood by those of ordinary skill in the art, however, that a patient may insist upon a lower dose or tolerable dose for medical reasons, psychological reasons or for virtually any other reasons.

Optionally, a mouse model (see Example 3) is used to screen various immunomodulating agents (and/or the method of administration) for the ability to prevent immunologic reproductive failure. Such screening is performed, for

example, by beginning administration of the immunomodulating agent to the mouse prior to, or shortly after conception, and determining whether (1) the incidence of reproductive failure is reduced and/or (2) the level of embryotoxic factor present in mouse samples (i.e., leukocyte and/or leukocyte secretion containing) is reduced.

Administration of the immunomodulating agent is performed in accordance with methods known to one of ordinary skill in the art. Accordingly, a variety of administration routes are available. The particular mode selected will depend, of course, upon the particular drug selected, the particular condition being treated and the dosage required for therapeutic efficacy. The methods of this invention, generally speaking, may be practiced using any mode of administration that is medically acceptable, meaning any mode that produces therapeutic levels of the agents of the invention without causing clinically unacceptable adverse effects. Such modes of administration include oral, rectal, vaginal, topical, transdermal or parenteral (e.g. subcutaneous, intramuscular and intravenous) routes. Formulations for oral administration include discrete units such as capsules, tablets, suppositories, patches, lozenges and the like. For an example of a vaginal suppository containing an immunomodulating agent (progesterone), see U.S. Patent No. 5,084,277, the contents of which patent are incorporated herein by reference.

The compositions may conveniently be presented in unit dosage form, including oral vaccine form, and may be prepared by any of the methods well known in the art of pharmacy. Such methods include the step of bringing the active agents into association with a carrier which constitutes one or more accessory ingredients. In general, the compositions are prepared by uniformly and intimately bringing the agents into association with a liquid carrier, a finely divided solid carrier, or both, and then, if necessary, shaping the product

Compositions of the present invention suitable for oral administration may be presented as discrete units such as capsules, cachets, tablets or lozenges, each containing a predetermined amount of the agents. Compositions suitable for parenteral administration conveniently comprise a sterile aqueous preparation of the agents, which is preferably isotonic with the blood of the recipient. This aqueous preparation may be formulated according to known methods using those suitable dispersing or wetting agents and suspending agents. The sterile injectable preparation may also be a sterile injectable solution or suspension in a non-toxic parenterally-acceptable diluent or solvent, for example as a solution in polyethylene glycol and lactic acid. Among the acceptable vehicles and solvents that may be employed are water. Ringer's solution and isotonic sodium chloride solution. In addition, sterile, fixed oils are conventionally employed as a solvent or suspending medium. For this purpose any bland fixed oil may be employed including synthetic mono- or diglycerides. In addition, fatty acids such as oleic acid find use in the preparation of injectibles .

Other delivery systems can include sustained release delivery systems. Preferred sustained release delivery systems are those which can provide for release of the agents of the invention in sustained release pellets or capsules. Many types of sustained release delivery systems are available. These include, but are not limited to: (a) erosional systems in which the agents is contained in a form within a matrix, found in U.S. Patent Nos. 4,452,775 (Kent), 4,667,014 (Nestor et al.); 4,748,024 and 5,239,660 (Leonard) and (b) diffusional systems in which an active component permeates at a controlled rate through a polymer, found in U.S. Patent Nos. 3,832,252 (Higuchi et al . ) and 3,854,480 (Zaffaroni). In addition, a pump-based hardware delivery system can be used, some of which are adapted for implantation.

Generally, daily oral doses of active compound will be from about 0.01 milligrams/kg per day to 1000 milligrams/kg per day. Small doses (0.01 - 1 mg) may be administered initially, followed by increasing doses up to about 1000 mg/kg per day. In the event that the response in a subject is insufficient at such doses, even higher doses (or effective higher doses by a different, more localized delivery route) may be employed to the extent patient tolerance permits. Multiple doses per day are contemplated to achieve appropriate systemic levels of compounds.

An exemplary therapeutic dose for administration of progesterone in suppository form is provided in Example 2. The therapeutic dose of progesterone (a 50 mg progesterone suppository) was administered twice a day (beginning three days after ovulation in a subsequent conception cycle) to patients having a history of recurrent reproductive failure). In the preferred embodiments, the therapeutically effective dose of immunomodulating agent is initially administered in the luteal phase prior to implantation or within one week post-conception. Preferably, immunomodulating agent drug therapy is continued until the risk of immunologic reproductive failure is eliminated (about 20 weeks) .

Also within the scope of the invention are kits for determining the presence of an elevated concentration of an extracellular embryotoxic factor in a sample. In its simplest form, the kit contains a first vial including an antibody to the embrytoxic factor, a second vial containing a standard including a concentration of embryotoxic factor indicative of the absence of a predisposition to immunologic reproductive failure ("negative" standard) or including a concentration of embryotoxic factor indicative of a predisposition to immunologic reproductive failure ("positive" standard); and instructions for using the antibody to determine the concentration of the embryotoxic factor in the sample and to compare the sample embryotoxic

factor concentration to the standard embryotoxic factor concentration to determine whether the mammal has a predisposition to immunologic reproductive failure. The instructions delineate the assay protocol and the statistical analysis necessary for determining the relative amount or absolute concentration of embryotoxic factor present in the sample. Such instructions are tailored to the specific type of immunoassay being performed, e.g., a competitive assay, a sandwich assay, according to the standard procedures.

The antibody to the embryotoxic factor is specifically reactive with the embryotoxic factor. By "specifically reactive", it is meant that the antibody recognizes and binds to an epitope on the embryotoxic factor and does not bind to extraneous molecules present in the sample preparation, which molecules do not include the embryotoxic factor epitope. Tests for determining the specificity of an antibody for a particular protein are known to those of ordinary skill in the art. For example, to demonstrate antibody specificity in an immunoassay in which the signal measured in the immunoassay is generated by virtue of a label attached to the antigen (e.g., a radiolabel, fluorescent tag, or enzyme label), the immunoassay is performed in the presence of an increasing amount of unlabeled antigen. If the antibody specifically reacts with the antigen, then a reduction in signal is observed, the signal reduction being proportional to the amount of unlabeled antigen present in the assay. Other methods for evaluating the specificity and sensitivity of immunoassays are known to those of ordinary skill in the art without the need for undue experimentation.

According to one embodiment, the second vial includes an amount of embryotoxic factor indicative of the absence of a predisposition to immunologic reproductive failure (negative standard) . The amount of factor which satisfies this criterion is established by assaying (e.g. , bioassay and immunoassay) sample preparations (i.e., leukocyte - and/or leukocyte secretion- containing samples) obtained from

fertile women who have no propensity to reproductive failure (see, e.g., the Examples). Accordingly, the second vial serves as a standard to establish a reference value to which the amount of embryotoxic factor quantitated in the sample is compared. Preferably, the amount of embryotoxic factor present in the second vial is selected such that a ratio of sample embryotoxic factor concentration to standard embrytoxic factor greater than about 2:1 is indicative of a predisposition to immunologic reproductive failure.

In the preferred embodiments, the kits further contain a third vial including a positive control, i.e., a vial including an amount of embryotoxic factor indicative of a predisposition to immunologic reproductive failure. The amount of embryotoxic factor contained in the positive control vial is determined by assaying samples obtained from women diagnosed with immunologic reproductive failure as described above, (see also the Examples).

To determine the amount of embryotoxic factor in leukocyte-containing samples, the kits further contain a vial including a plurality of reproductive antigens for stimulating the leukocytes to release embryotoxic factors. Preferably, the reproductive antigens are sperm antigens, trophoblast antigens and/or antigens derived from a human choriocarcinoma cell line. In the most preferred embodiments, the reproductive antigens are purified from choriocarcinoma cell lines BeWo or Jeg-3. It will be understood that the above-described kits for determining an elevated concentration of the embryotoxic factor can easily be adapted for determining the concentration of a particular T„π-l or a Tπ _u -2 cytokine(s) in the sample (e.g., by substituting an antibody to the cytokine for the above-mentioned antibody to the embryotoxic factor) and including instructions which describe comparing the sample cytokine concentration to a positive or negative cytokine standard to determine whether the mammal has a predisposition to immunologic reproductive failure.

Also provided within the scope of the invention are vials containing various isolated reproductive antigens and/or embryotoxic factors. By isolated, it is meant that the antigen or factor has been removed from its natural matrix (e.g., blood, tissue) and exists in a substantially purified form (e.g., greater than about 80 % pure as assessed by gel electrophoresis) . The vial can include one or more isolated antigens. Preferably, the isolated antigen has a purity exceeding 90%. A higher degree of purity is needed for the in vivo than is required for in vitro applications.

According to one aspect of the invention, a vial is provided which includes an amount of isolated reproductive antigen capable of stimulating a plurality of leukocytes to release an embryotoxic factor. In the preferred embodiments, the vial containing the reproductive antigens for stimulating leukocytes in vivo further includes a pharmaceutically acceptable carrier. Such carriers are known to one of ordinary skill in the art and include, for example, an isotonic saline solution. To modulate a cellular immune response to a reproductive antigen in vivo, a vaccine is prepared which includes the reproductive antigen in the above-described adjuvants for downregulating a T-ri,-1 type response or for upregulating a T„-2 type response.

The following non-limiting examples illustrate representative utilities of the instant invention. Example 1 demonstrates that the preimplantation embryo is a vulnerable target of embryotoxic factors produced by activated leukocytes from many women with recurrent abortion. Example 1 further demonstrates that determination of an elevated concentration of embryotoxic factor is indicative of a predisposition to immunologic reproductive failure. Example 2 illustrates the use of immunomodulating agent drug therapy for the treatment of women previously diagnosed with im unologic reproductive failure. Example 3 illustrates the anti-fertility effects of antisperm cell-mediated immunity in mice. Example 4 demonstrates that lymphocytes from a

majority of women with unexplained recurrent abortion make Tπ„l-type cytokines when exposed to trophoblast extracts, whereas lymphocytes from parous women with normal reproductive histories and men make T„2-type cytokines.

Example 5 demonstrates that lymphocyte proliferation in response to trophoblast antigen stimulation was significantly higher in women with recurrent abortion of unknown etiology than in fertile controls and significantly correlated with embryotoxic factor activity in culture supernatants.

EXAMPLES Example 1: Immunomodulating Agent Drug Therapy in Pregnant

Women For Prevention of Spontaneous Abortion (l) Study Patients

The subjects in this study were 300 reproductive-aged, non-pregnant women undergoing evaluation for recurrent abortion at the Brigham and Women's Hospital Reproductive Endocrinology and Infertility Clinic between July 1, 1986 and June 31, 1989. The women were between 22 and 42 years old and had a history of at least three prior spontaneous abortions with or without a stillbirth, ectopic pregnancy, or live birth. These women were grouped by clinical history as follows: women with primary abortion (no live births), women with secondary abortion (abortions subsequent to live births or stillbirths), women with previous ectopic pregnancy, or women with other criteria (abortions interspersed with live births). After a complete medical, surgical, and social history was obtained from the couple and a thorough physical examination was performed on the woman, all couples had peripheral blood chromosome assessment by standard banding techniques. All women underwent hysterosalpingography followed by hysteroscopy-laparoscopy when indicated. Ovulation was assessed by a late luteal-phase endometrial biopsy with concomitant serum progesterone determination. When the endometrial biopsy specimen was out of phase according to the criteria of Noyes et al. (Fertil.

Steril.1:3-25 (1950), the suspected ovulation disorder was confirmed by repeating the biopsy in a subsequent cycle. In addition, all women had cervical cultures for mycoplasma and ureaplasma, and blood was obtained for thyroid-stimulating hormone, thyroxine, prolactin, anticardiolipin antibody, and lupus anticoagulant (Russell viper venom time) determinations. Cervical mucus and serum were also obtained for antisperm antibody assessment. After the preceding evaluation, study patients were categorized as having either a genetic, anatomic, endocrine, infectious, or humoral immune cause for their reproductive failure. Peripheral blood was obtained from all women on their initial visit for embryo- and trophoblast-toxic factor(s) testing.

Peripheral blood was also obtained from a comparison group of 30 non-pregnant, paid volunteer women who were randomly selected from the Obstetrical Nursing Service. Women in this control group were of reproductive age (23 to 41) with a history of at least three prior uncomplicated term deliveries and no history of either spontaneous abortion, ectopic pregnancy, or stillbirth. None of the women in the study were taking medication at the time of blood collection. (2) Assays

(a) Embryotoxic factor assay

(i) Reproductive antigen preparation. Trophoblast antigen extracts were prepared from the human gestational choriocarcinoma cell lines BeWo and Jeg-3 (American Tissue

Type Collection, Betheεda, Md.) These cells were grown in culture flasks by standard culture techniques until confluence. After trypsinization, the cells were washed twice in Hanks' balanced salt solution without calcium or magnesium (Gibco, Grand Island, N.Y.), and cells excluding trypan blue were counted on a hemocytometer (viability always

₇ > 90%). The cells were diluted to 10 cells per milliliter in tissue culture media comprising Roswell Park Memorial

Institute 1640 medium (RPMI;Gibco) , 10% fetal calf serum, 0.2 mmol/L glutamine, 100 IU/ml benzylpenicillin potassium, and

100 mg/ml streptomycin sulfate (Sigma, St. Louis). Cells were disrupted in a dounce homogenizer, and after centrifugation at 400g for 10 minutes, the membrane-enriched trophoblast antigen supernatant was divided into aliquots and stored at -70°C until use.

Semen was obtained from fertile donors by masturbation into sterile containers, and motile sperm were isolated by percoll density gradient centrifugation. Aliquots of 10 7 sperm per milliliter in tissue culture media were dounce homogenized. The suspension was then centrifuged at 400g for

10 minutes and the membrane-enriched sperm antigen supernatant was divided into aliquots and stored at -70°C until use.

(ii) Preparation of Leukocyte-containing samples.

Mononuclear leukocytes were separated from blood obtained from the study patients by Ficoll-Hypaque (Pharmacia,

Uppsala) centrifugation. Isolated cells were washed twice and resuspended in RPMI medium. Viable mononuclear cells excluding trypan blue were counted on a hemocytometer

(viability always > 90%) and diluted in tissue culture medium to a concentration of 10 cells per milliliter. Leukocyte suspensions were then divided into three 50 ml tissue culture flasks (Becton Dickinson, Lincoln Park, N.J.) containing

₇ 10 leukocytes per 10 ml of medium in each loosely capped flask. One milliliter of either trophoblast antigen suspension, sperm antigen suspension, or tissue culture medium alone was added to each flask. After a 72 hour incubation at 37°C in 3% carbon dioxide and 95% air, the cell cultures were transferred to separate 15 X 95 mm polystyrene round-bottom tubes with caps (Becton Dickinson) and pelleted by centrifugation at 400g for 10 minutes. The cells were next washed once in RPMI medium and once in Whitten's medium and then resuspended in 1 ml of Whitten's medium supplemented with antibiotic-antimycotic solution (Gibco) that had been previously passed through a 0.45 mm filter (Corning Glass,

Corning, N.Y.). The pH of the medium was adjusted before use

to 7.3 with carbon dioxide gas and supplemented with 0.3% bovine serum albumin (Miles Scientific, Naperville, ILL). After an additional 12-hour incubation at 37°C in 5% carbon dioxide and 95% air, the three suspensions from each original sample were centrifuged at lOOg for 10 minutes. The three supernatants were each filtered through a 0.22 m Millex-GS filter unit (Millipore, Bedford, MA) and stored at -70°C until use.

(iii) Preparation of Leukocyte secretion- containing samples. Peripheral blood, peripheral serum, peritoneal fluid, endometrial tissue (extracted or homogenized), vaginal secretions and saliva can be used directly or are processed to remove cellular components (e.g., leukocytes). The preparation of peripheral blood mononuclear cell (PBMC) supernatants is described in Example 4. The leukocyte-containing samples are stimulated and assayed as described above.

(iv) Preparation of Trophoblasts. Trophoblasts were prepared from a choriocarcinoma cell line as described above (Example 1) .

(b) Embryo development assay

Embryo culture was performed as previously described (Hill, J.A., et al., J. Immunol. 139:2250-2254 (1987)). Briefly, two-cell embryos were harvested from CD-I female mice (Charles River, Kingston, Ontario) 44 hours after human chorionic gonadotropin administration. Embryos were cultured in 20 ml drops containing a 1:1 dilution of study supernatant in Whitten's medium -0.3% bovine serum albumin under carbon dioxide - equilibrated paraffin oil in Falcon tissue culture dishes(Becton Dickinson) . Media and supernatants were equilibrated overnight in 5% carbon dioxide and 95% air before the addition of embryos. Embryos cultured in Whitten's medium - 0.3% bovine serum albumin alone served as an additional control. At least 11 embryos were cultured in each 20 ml drop. Embryos were assessed for normal development after 4 days in culture (100 hours after human

chorionic gonadotropin administration) by the criteria described by Ducibella (Dev. Biol. 79:356-66 (1980)). The experimental end point was the percentage of normal blastocysts in each culture after 4 days of development in vitro. The sensitivity and specificity of the assay were maximal when the median percentage of embryos advancing to the blastocysts stage of development was 50% of control values. The intraassay and interassay coefficients of variation were 8% and 12%, respectively, for the embryo-toxic bioassay.

(c) Trophoblast proliferation assay

Human gestational choriocarcinoma cells (10 cells per 0.1 ml tissue culture medium) were plated in triplicate into 96-well, flat-bottom microtiter trays (Becton Dickinson). Two hours after cell plating, 0.1 ml of leukocyte supernatant or media alone was added (four wells per test solution) . After a 2-day incubation at 37°C in 5% carbon dioxide and 95% air, 0.5 mCi of tritiated thymidine z(13.1 Ci/mmol; New England Nuclear, Boston) was added to each well. Cells were harvested at 72 hours after trypsinization onto glass fiber filter paper with a MASH II automatic sample harvester (Microbiological Assoc. Los Angeles). Viability of control cultures was assessed before harvesting by means of trypan blue exclusion (viability always > 90%). Filter paper disks corresponding to each well were air dried, transferred to scintillation counting vials, and suspended in 2 ml of Betafluor scintillation cocktail (National Diagnostics, Summerville, N.J.). Radioactivity (counts per minute) was measured in a liquid scintillation counter (Beckman Instruments Inc., Fullerton, Calif) (efficiency of the system 45%), and the mean = SD of each set of triplicate cultures was calculated. The data are presented as percentages of culture medium control. Trophoblast toxicity was assumed when the median percentage of trophoblast proliferation was _ 50% of control values.

(d) Leukocyte Proliferation assay

The leukocyte proliferation assay is performed according to procedures known to one of ordinary skill in the art (see Example 5 and e.g., Immunol. Rev. 75:61-85 (1983) and J. Reprod. Immunol. 6:377-391 (1984), the contents of which references are incorporated herein by reference) . Briefly, the above-described preparation of reproductive antigen(s) is added to 2.5 x 10(5) leukocytes in culture, followed by assessment of tritiated thymidine incorporation to determine the extent of DNA proliferation. All assays are performed in triplicate. A diagnosis of a predisposition to immunologic reproductive failure is made if the lymphocyte secretion-containing sample has a Stimulation Index that is greater than about 3.0. Fertile control subjects have a stimulation index that is less than 3.0. As used herein, the term "Stimulation Index" refers to the ratio of counts per minute (cpm) for unεtimulated cultures to the cpm for antigen-stimulated cultures.

(e) Gamma-interferon ELISA

An exemplary IFN-gamma ELISA, as well as other cytokine immunoassays useful for the purposes of the invention are provided in Example 4. A commercially-available enzyme-linked immunoassay (e.g., Genzyme Corp., Cambridge, MA) has been used to determine the amount of an embryotoxic factor present in a reproductive antigen-stimulated leukocyte-containing samples (see Example 4). These assays also are useful for determining the amount of an embryotoxic factor in a leukocyte-secretion containing samples.

(f) Statistical analysis

We compared the mean age distribution between the 300 women with recurrent abortion and the 30 control subjects with a two-sample test for independent samples and assumed equal or unequal variances based on the F distribution. Differences in mean ages by different categories of abortion and reproductive failure causes were assessed through analysis of variance. The percentages of mouse blastocyst

development or trophoblast proliferation in supernatants failed the test of normality with the Shapiro-Wilk W statistic, therefore median percentages were compared between the 300 women with recurrent abortion and the 30 controls overall and within each of the categories of abortion and reproductive failure causes with the use of non-parametric tests of medians and the Wilcoxon rank sum test. All directional p values were two tailed.

(3) Embryotoxic Factor Characterization

Trophoblast antigen-stimulated leukocyte supernatants from 15 patients that adversely affected embryo development were subjected to three different methods, to characterize the factor(s) responsible for embryo toxicity; molecular weight dialysis, heat inactivation, and affinity column purification. The supernatants were dialyzed for 12 hours against three changes of Whitten's medium - 0.3% bovine serum albumin in Spectrapor membrane tubing (Spectrum Medical Industries, Los Angeles) with molecular weight limits of 3500, 10,000 or 30,000 and then tested in the mouse embryo development assay. In separate experiments aliquots of these supernatants were heat treated at 56°C for 1 hour or passed through an affinity column containing anti-interferon gamma-coated beads (Endogen, Boston) before testing in the mouse embryo development assay. Five leukocyte supernatants not exhibiting an adverse effect and medium alone were used as controls in each of these experiments.

(4) Results: The leukocytes of women having a predisposition to immunologic spontaneous abortion released Embryotoxic Factors in response to stimulation with reproductive antigens in vitro.

Women with recurrent abortion were 80% white, 10% Hispanic, 8% black, 1% Native American, and 1% Asian. The 30 fertile controls consisted of 25 white, 3 Hispanic, and 2 black women. The 300 women evaluated for recurrent abortion had experienced 1367 prior pregnancies, 1213 of which resulted in reproductive failure (median 4, range 3 to 25).

Fetal loss occurred during the first trimester in 88%, whereas 12% of the abortions occurred during the second trimester (before 17 weeks of gestation) .

The control group of women without a history of abortion were slightly older than women with a history of reproductive failure (37.3 + 3.1 vs. 35.9 + 3.4, Table II). Among the 300 women with recurrent abortion, 215 women (72%) had primary abortions, 50 women (17%) had secondary abortions, and the remaining 35 women (11%) could not be identified as having either primary or secondary abortions. Seven of these women had a total of 9 ectopic pregnancies (all tubal gestations), and the other 28 women experienced recurrent abortions interspersed with live births.

The cause of recurrent abortion was unexplained in 180 (60%) women. The remaining 120 women (40%) had recurrent abortion presumably as a result of endocrine abnormalities (17%), anatomic anomalies (10%), infections (5%), humoral immune disturbances (4%), or chromosomal aberrations (5%). There was no correlation between the average age of the women and etiologic categories, except in women with an infectious cause (positive cervical culture for mycoplasma or ureaplasma), who were significantly younger than the other women with a history of reproductive failure (32.9 + 4.1 vs. 35.7 + 1.0, p < 0.01) .

There was no difference between effects of supernatants derived from unsti ulated peripheral blood mononuclear cells from women in any category and control medium in either the embryo development or trophoblast proliferation assays (data not shown) . Embryo development and trophoblast proliferation in response to trophoblast antigen-activated leukocyte supernatants and sperm antigen-activated leukocyte supernatants have previously been reported (see Hill, J. et al., Am. J. Obεtet. Gynecol . 166:1044-52 (1992)). Neither trophoblast antigen-activated nor sperm antigen-activated leukocyte supernatants from the 30 fertile women adversely affected embryo development or trophoblast proliferation.

Trophoblaεt and sperm antigen extracts alone had no effect on either embryo development or trophoblast proliferation. Neither race nor ethnicity was asεociated with either the production of toxic factorε or the variouε cauεeε of recurrent abortion. Both embryo development and trophoblast proliferation in the presence of trophoblast antigen-activated and sperm antigen-activated leukocyte supernatants were decreased in women with recurrent abortion as compared with fertile controls. In the 300 patientε with recurrent abortion, embryo development waε more often and more εeverely affected than trophoblaεt proliferation in the presence of both sperm antigen-activated and trophoblast antigen-activated leukocyte εupernatants (effects of trophoblast and εperm antigen-activated leukocyte supernatants on embryo development versuε trophoblast proliferation p < 0.01). Embryo development was more adversely affected by trophoblaεt antigen-activated leukocyte εupernatantε than by sperm antigen-activated leukocyte supernatants, except in those women with a prior ectopic pregnancy.

Embryo development in the preεence of trophoblaεt antigen-activated or εperm antigen-activated leukocyte εupernatantε from women with an unknown cause of abortion was significantly less than embryo development in comparable supernatantε from women with other cauεeε (p > 0.01). Embryo development waε also observed to be lower in trophoblaεt antigen-activated and εperm antigen-activated leukocyte εupernatantε from women with recurrent abortion cauεed by an endocrine or humoral immune abnormality (p < 0.001) but not in thoεe from the two groups of women with an anatomic or genetic etiology of recurrent abortions. In the two women with antisperm antibodies, embryo development was unaffected by sperm antigen-activated leukocyte supernatants but was adversely affected by trophoblaεt antigen-activated leukocyte εupernatants. However, within the anatomic group embryo-toxic factorε were produced in women with a diagnosis

of endometriosiε aε their only abnormal finding in reεponse to stimulation by both trophoblast and sperm antigens (p < 0.05). In the same group trophoblast proliferation was also significantly inhibited in response to trophoblast antigen-activated leukocyte supernatantε from women with endometriosis (p < 0.05). A woman with a pericentric inversion of chromosome 9 had low embryo development in response to both trophoblast antigen-activated and sperm antigen-activated leukocyte supernatants and low trophoblaεt proliferation in reεponεe to trophoblaεt antigen-activated leukocyte supernatants. Otherwise, cytogenetic abnormalities were not asεociated with either embryo or trophoblast toxicity.

The εenεitivity and εpecificity of the bioassays were greatest when embryo development and trophoblast proliferation was <_ 50% of control values (Table II).

Only one woman in the fertile control population had a value < 50%. This woman had 45% blastocyεt development in the preεence of εupernatant from her εperm antigen-activated leukocyte culture. Uεing 50% of control valueε aε our cutoff, we detected women with recurrent abortion aε follows: for embryo development assays, 63% using trophoblast antigen-activated and 51% uεing εperm antigen-activated leukocyte εupernatantε; for trophoblaεt proliferation, 32% using trophoblast antigen-activated and only 13% using sperm antigen-activated leukocyte supernatantε. Only one (3%) of the fertile controls would have been falsely clasεified as having recurrent abortion with sperm antigen-activated leukocyte supernatant in the embryo development asεay, and none would have been falεely claεεified with the other assay-εupernatant combinations.

The embryo development assay of reproductive antigen-stimulated leukocyte preparations was more sensitive than the trophoblast proliferation assay in predicting the propensity for abortion. Only four of 300 women had > 50%

TABLE I I

Sensitivity and specificity for recurrent abortion by aεεay and reproductive antigenic factor

Women Women with Without abortions abortions Sensitivity Specificity

(No.) (No.) (%) (%)

63 100

51 97

32 100

13 100

TABLE I I I

Sensitivity and specificity of the embryo development assay for predicting reproductive failure in women with known and unknown causes of their abortions

Women Women with Without abortions abortions Sensitivity Specificity

(No.) (No.) (%) (%)

All women ED-T or ED-S <_ 50% (pos. ) 236 1

₇₉ ! 97-

ED-T and

ED-S > 50% (neg. ) 64 29 Cause Unknown ED-T or ED-S <_ 50% (pos. ) 161 1

89i 97%

ED-T and

ED-S > 30% (neg. ) 19 29 Cause known ED-T or ED-S 50% (pos. ) 75 1

65% 97-

ED-T and

ED-S > 50% (neg. ) 45 29

embryo development when trophoblast proliferation was <_ 50%. If only trophoblast proliferation had been used to asseεs toxic factor production, 127 women would have been falsely screened as negative. Only 113 women produced toxic factors in response to reproductive antigens in the trophoblast proliferation assay compared with 236 women who tested positive with the embryo development bioaεsay (see Table III Hill, J. , et al. , Am. J. Obstet . Gynecol. 166:1044-52 (1992 for data) .

In Table III we considered a positive teεt for recurrent reproductive failure to be <_ 50% of embryo development in reεponse to either trophoblast or sperm antigen-activated leukocyte εupernatants. Thiε allowed us to detect 79% of women with recurrent abortion while still maintaining an extremely low number of false-poεitive results (3%) . This propensity to generate Embryotoxic Factors in reεponεe to either trophoblaεt or εperm antigen εtimulation was more pronounced in women with an unknown cause aε compared with a known cauεe for their abortion. The preεence of embryo- and trophoblaεt-toxic factors was not influenced by a change in partners because six women in whom toxic factors were produced in responεe to εperm-antigen εtimulation (n = 4) and trophoblaεt antigen εtimulation (n = 2) experienced recurrent abortionε with both their prior and current huεbandε.

The effects of embryotoxic factor-containing samples on embryo development after molecular weight dialysiε indicated that the factor(ε) responsible for embryo toxicity (in all 15 supernatants tested) was between about 10 and 30 kd. The embryo-toxic effect of the εupernatantε waε abrogated follow!-.g heat treatment at 56°C for 1 hour. In 12 of 15 embryo-toxic supernatants (from reproductive antigen-stimulated leukocyte preparations) the factor(s) was absorbed on affinity columns containing anti-interferon gamma beads. Passage of control media through the anti-interferon gamma column had no effect on embryo development. Further evidence identifying IFN-gamma as an embryotoxic factor is provided in Example 4.

Example 2: Immunomodulating Agent Drug Therapy For

Preventing Immunologic Spontaneous Abortion (1) Study Patients .

We reviewed the records of all women evaluated for recurrent abortion at the Fertility and Endocrine Unit of Brigham and Women's Hospital between July 1987 and June 1991. For the purposes of this study, recurrent abortion was

defined as two or more pregnancy losεeε before 20 week's gestation. All women were evaluated according to a standard protocol that included the following: (1) a thorough history and physical examination; (2) chromosome asεeεεment of the women and her partner using standard banding techniques; (3) hysterosalpingogram followed by laparoscopy/hyεteroεcopy when indicated; (4) cervical cultureε for mycoplasma and ureaplasma, with treatment if positive; (5) serum asεay for TSH, thyroxine, prolactin, anticardiolipin antibody, and lupus anticoagulant (Rusεell viper venom time); (6) evaluation of anti-sperm antibodieε in the woman' s εerum and cervical mucus; and (7) evaluation for luteal-phase insufficiency by an endometrial biopsy during the late luteal phaεe. When an endometrial biopsy was out of phase according to established criteria, the defect was confirmed on a subεequent biopεy and treated with ovulation induction until a repeat biopsy demonstrated an in-phase endometrium. At the conclusion of the evaluation, the patients were categorized as having either a genetic, anatomical, endocrine, infectious, or humoral immune etiology for their reproductive failure. Women in whom no presumed etiology could be detected were diagnosed with unexplained recurrent abortion. (2) Sample Preparation. Leukocyte and Leukocyte-secretion containing samples were prepared as described above from whole blood and εerum. To prepare leukocyte-containing samples, lymphocytes and macrophages from the patients' blood were isolated and maintained in tissue culture in the presence of antigens obtained from a human trophoblast cell line and human sperm (described above) . Supernatantε from theεe cultureε were collected and added to two-cell mouse embryos in culture, with assessment of subsequent blastocyst development. A toxic effect was asεigned when blastocyst development was leεs than 50% of control values as previously described. All samples were coded and the assays performed by the same technician, who was not aware of the sample source or patient status.

(3) Immunomodulating Agent Drug Therapy, - Preliminary Resultε.

(a) Methodε

Women found to be poεitive for Embryotoxic Factor(ε) upon initial evaluation were given progeεterone (50-mg εuppoεitory twice a day) beginning 3 dayε after ovulation in a εubsequent conception cycle. Progesterone was chosen for itε potential immunoεuppressive effects. The women were closely followed by the εame investigator and asked to return weekly for hCG assessment and then at 5 weeks' gestation for pelvic ultrasound and reasεeεεment of embryotoxic factors. Fetal viability was aεεessed by pelvic ultraεound every 2 weekε until 12 weekε' geεtation. Women with a viable intrauterine pregnancy documented by ultraεound at 12 weekε' gestation were referred to an obstetrician for continued care.

Patients and their referring physicians were asked to obtain karyotypes of any εubεequent abortions. In some cases, obstetric care waε continued at our hoεpital and outcomeε were eaεily obtained from the medical record. In all cases, we mailed a follow-up questionnaire to the patientε that asked for the outcomes of all pregnancies after the initial evaluation at our center.

Data were εtatiεtically analyzed uεing x 2 with Yateε correction.

(b) Reεults

We evaluated 450 non-pregnant women in our unit for a history of recurrent abortion (mean of four previouε loεεeε, range 2 to 25). Of theεe, 346 had a normal chromosomal analysiε, normal hyεteroεalpingogram, normal or corrected luteal-phaεe endometrial biopεy, negative or treated cervical cultureε, normal thyroid function, and negative anticardiolipin antibody and lupus anticoagulant. Embryotoxic factors were identified in 328 (95%) of theεe women.

After this evaluation, there were a total of 208 known pregnancieε in 166 women. Repeat embryotoxic factor and outcome data were available in 141 pregnancieε in 117 women. Of the 67 pregnancieε (61 women) for which outcomes were not known, 31 were positive and 36 were negative for embryotoxic factorε when reassayed in the firεt trimeεter.

The embryotoxic factor assay was predictive of pregnancy outcome (p < .01). Of 56 women still positive for embryotoxic factorε early in pregnancy, 40 (71%) miεcarried (37 in the firεt 10 weeks of pregnancy and three between 12-16 weeks) and 16 (29%) delivered a viable infant. Of the 85 women who were found to be negative for embryotoxic factors on repeat testing, only 11 (13%) miscarried (all in the first trimeεter), whereas 74 (87%) delivered a viable infant. The positive and negative predictive values of the embryotoxic factor asεay were 0.76 and 0.86, reεpectively.

Chromoεomal analyεeε of the aborted geεtationε were requested, but not performed in every case. Abnormalities were found in at leaεt four of 11 karyotyped miεcarriages in women negative for embryotoxic factors early in pregnancy, including one complete mole, two triploidic partial moleε, and one autoεomal triso y 18. Chromosomal analysiε waε available in nine εubsequent abortions of women who were εtill poεitive for embryotoxic factorε early in pregnancy. Abnormal karyotypeε (two triploidic partial moleε and one autosomal trisomy 21) were found in three of these nine cases.

Among the 21 women with more than one subεequent pregnancy after their initial evaluation (19 with two pregnancieε, one with three, and one with four), there were 14 patients in whom first-trimeεter embryotoxic factor values were discordant when compared in subεequent gestations. Ten of theεe 14 women miscarried during the pregnancy when their embryotoxic factors were positive and carried another pregnancy to term when their embryotoxic factors were negative. Two of the 14 spontaneously aborted all subsequent pregnancies (three pregnancies in one patient and four in the

other). The embryotoxic factor assay predicted all subsequent reproductive failures in these two women except for two caseε (one in each) in which the embryotoxic factors were negative yet the pregnancies still spontaneouεly aborted. The remaining two of theεe 14 women each carried two pregnancies to viability even though the embryotoxic factor assay predicted that one pregnancy in each would be aborted. All of these pregnancy losεeε occurred during the firεt trimester after documentation of fetal cardiac activity at 5 weeks' gestation. (4) Immunomodulating Agent Drug Therapy: Selection of

Immunotherapeutic Agent Doseε For Preventing Immunologic

Reproductive Failure

Leukocyte-containing and/or leukocyte εecretion- containing εampleε are prepared as described above and are teεted for the presence of an elevated concentration of Embryotoxic Factor(s) using the above-described asεayε. Preferably, εampleε are teεted uεing an assay such as an immunoaεεay for the Embryotoxic Factor to allow for rapid aεεeεεment of efficacy of the immunomodulating agent therapy. Accordingly, patient εampleε (containing leukocytes and/or leukocyte secretion products) are aεsayed for the presence of an elevated concentration of Embryotoxic Factor's) prior to and during the immunomodulating agent therapy regimen.

An immunomodulating agent, such as progesterone, is administered to a patient previously diagnosed (as above) as having recurrent reproductive failure. Leukocyte and/or leukocyte secretion-containing samples are prepared and asεayed as described above to determine the amount of Embryotoxic factor present in the sample. A subεequent dose of the immunomodulating agent is administered to the patient, the subεequent doεe being selected to cause a reduction in the amount of embryotoxic factor released in vivo by the leukocytes. In this manner, the immunomodulating agent therapy is precisely tailored to the cellular immune response

of each patient. Thiε iterative process is continued until the concentration of extracellular embryotoxic factor in vivo iε sufficiently low (as determined by assay and comparison with a standard concentration indicative of the absence of a prediεpoεition to reproductive failure) to preclude abortion of the fetuε.

Example 3: Antifertility Effects Of Antisperm Cell-Mediated Immunity In Mice

(1) Materials and Methods

(a) Preparation of Cells to be Used as Antigens

(i) Sperm. DBA/2 retired breeder males (Jackson Laboratories, Bar Harbor, ME) were killed by cervical dislocation and the cauda epididymides were removed and minced with fine sciεεors. Sperm were suspended in Krebs-Ringer solution, pasεed through a fine εcreen and collected from the cell pellet following centrifugation at 400 X g for 5 min. Motile εperm were further isolated from immotile sperm and εomatic cellε by centrifugation through a discontinuous gradient of Percoll (Pharmacia, Piscataway, NJ) (47%/90%) for 30 min. at 500 X g. The cell pellet containing motile εperm was washed 3 timeε and reεuεpended at a final concentration of 1 X 10 7/ml m saline. The cell suspenεion waε mixed 1:1 with CFA or IFA to form a homogenouε emulεion, aliquoted and εtored at -70°C until use.

(ii) Spleen Mononuclear Leukocytes. Splenic mononuclear leukocytes from DBA/2 retired breeders were isolated by centrifugation through Lympholyte-M

(Accurate Chemical & Scientific Corp. , Weεtbury, NY) for

30 min. at 500 X g. Leukocytes removed from the interface above the Lymphocyte-M were washed 3 times, resuεpended at a concentration of 1 X 10 7 in saline, mixed 1:1 with CFA or IFA to form a homogenous emulsion and aliquoted and stored at -70°C until use.

(iii) Red Blood Cells. Human red blood cells were isolated from heparinized peripheral blood by centrifugation through Ficoll-Hypaque (Pharmacia LKB

Biotechnology Inc., Piscataway, NJ) for 30 min. at a

500 X g. The buffy coat on the red blood cell pellet containing granulocytes was discarded. The bottom layer containing < 99% red blood cells (confirmed by microscopic evaluation) waε collected, washed, resuspended in saline at a concentration of 1 X

10 /ml, mixed 1:1 with adjuvant, aliquoted and stored at -70°C.

(b) Active Immunization Protocols

C57BL/6 female mice (Jackεon Laboratorieε, Bar

Harbor, ME) maintained on a 12-h light/dark cycle were immunized with 1 X 10 allogeneic (DBA/2) εperm, in

Complete Freund'ε Adjuvant (CFA; Cappel, Organon Teknika

Corp., Weεt Cheεter, PA). Control female mice of the εame age and εtrain were injected with: (1) εaline (0.9% NaCl in diεtilled H_0) aε a control for stresε, (2) εaline-pluε-CFA as a control for the effects of the CFA alone, and (3) human red blood cells and (4) paternal (DBA/2) mouse lymphocyteε in

CFA to aεεeεs the sperm-εpecificity of the antifertility

₇ effect. Booεter mjectionε of 1 X 10 εperm or control cellε, or saline alone in Incomplete Freund's Adjuvant (IFA;

Cappel, Organon Teknika Corp.) were given on day 14 and 21.

Injections (0.1 ml) were adminiεtered: (1) subcutaneously into the neck (s.c), (2) intraperitoneally (i.p.), or (3) intrauterine (i.u.) through the cervix which waε dilated by injection 12 h earlier with 1 IU of Pregnant Mare'ε Serum

(PMS; Sigma Chemical Co., St. Louiε, MO). For intrauterine injectionε, mice were aneεthetized with 0.2 ml of Advertine εolution (2.5 ml tert-amyl alcohol) Fisher Labs, Medford,

MA), 1.25 g 2,2,2, tribromoethanol 99% (Aldrich Chemical

Company, Milwaukee, WI) in 100 ml distilled water) and the cervix was visualized through a glass speculum. A catheter fitted with a 1-ml syringe and a 27 1/2-gauge needle (Becton

Dickinson, Rutherford, NJ) was threaded through the cervical opening and 0.1 ml of cells or εaline emulsified with adjuvant was introduced into the uterine lumen.

Twelve hourε before the final booεter injection, mice in all immunization groups were primed with 1 IU of PMS to cycle the animals but not to induce hyperstimulation. Thirty-εix hours after the booster mice were injected with 1 IU human chorionic gonadotropin (hCG; Sigma) and placed with breeder maleε from our εtud colony; one female per male in a εingle cage. Half of the mice were aεsesεed for number of ova and 2-cell embryos 40 h later and the other half were killed at day 15 of pregnancy and the numbers of viable fetuses, non-viable fetuseε and reεorption εiteε were aεεeεεed. Blood waε drawn by retroorbital puncture for aεεeεεment of antiεperm antibodieε and uteri were snap frozen and stored at -70°C for subsequent immunohistologic examination. Each experimental group consiεted of 10 mice and each experiment waε repeated at leaεt three times.

(c) Pasεive Immunization Studieε

Spleenε and lymph nodes were harvested from s.c./s.c/i.u. immunized C57BL/6 female mice one week after final immunization and from age and εtrain-matched non-immunized mice. The lymphoid organs were minced with fine scissors and cells were passed through a fine nylon screen. Red blood cells were lysed according to Maruyama et al . (1985), J. Androl . 6, 127-135 and mononuclear leukocytes were separated by gradient centrifugation on Lympholyte-M at 500 X g for 30 min. and waεhed three timeε in RPMI-1640 medium (GIBCO, Grand Island, NY).

T lymphocytes were obtained by nylon wool separation following a standard procedure (Trizio and

Cudgowicz (1974), J. Immunol. 113:1093-1097). Non-adherent cells (T cell enriched) were resuεpended in RPMI 1640 at 10 X

₇ 10 /ml and p.l ml waε injected intravenously into the tail vein of C57BL/6 female mice. PMS (1 IU) waε injected i.p. at the same time. Forty-eight hours later, mice were given 1 IU

hCG i.p. and mated. On day 15 of pregnancy, mice were killed by cervical diεlocation and the numbers of viable fetuses, non-viable fetuseε and resorption sites were recorded. Uteri were frozen in OCT gel (Baxter McGaw, Park, IL) on dry ice and stored at -70°C for subεequent immunohistological aεεeεεment. Blood waε collected by retroorbital puncture prior to termination and serum was εtored at -70°C for antiεperm antibody aεεay.

(d) Immunohiεtologic Staining of Mononuclear Cells

T lymphocytes used for the passive study were

₇ suspended in PBS at a concentration of 1 X 10 ; 5 μl of the cell suεpenεion waε applied to individual εpots of teflon-coated, eight-spot slides, air dried, fixed in acetone for 10 min. and stored at -70°C until use.

Uteri from actively immunized, passively immunized and non-immunized mice were embedded in groups of 5 in OCT gel on dry ice and εtored at -70°C until use. Before sectioning the uteri were equilibrated in the cryostat for 1 h to reach -20°C. Tiεεue εliceε (4-6 μm) were placed on

3-spot teflon slideε (Roboz Surgical Inεtru ent Co.,

Waεhington, DC) and were air dried overnight and fixed in 2% paraformaldehyde. Macrophages and lymphocyte subpopulations were detected in cell smearε and tiεεue sections by immunoperoxidase technique. A prewaεh in 0.02 N sodium azide/phosphate-buffered εaline (PBS) εolution was performed to destroy endogenous peroxidaεe before the primary rat anti-mouse monoclonal antibodies were applied (Helomy et al . ,

(1988), J. Immunol. Methods 111:101-106). All antibodies were diluted to optimal working concentration in PBS/1% BSA.

For the monoclonal antibodies Thy 1 (1:20), L3T4 (1:20), B

(1:20) and Ml/70, 15 (1:50) (Sera Lab, Accurate Chemical

Scientific Corp., Weεtbury, NY), the εecond immunoperoxidaεe-labelled antibody waε a peroxidase-labeled

F(ab') _? fragment mouse anti-rat IgG (Jackson Immunoresearch

Laboratorieε, Inc., West Grove, PA) diluted 1:100 in PBS containing 1% BSA. For the slideε stained with biotinylated

Lyt 2 antibody (1:20) (Becton Dickinson, Mountain View, CA) , peroxidase labeled εtreptavidin (Kirkegaard & Perry Laboratories Inc., Gaithesburg, MD) waε uεed diluted 1:200 in PBS/1% BSA. The substrate was 0.2 g of 3-amino-9-ethyl carbaεol (AEC) diεεolved in 50 ml dimethyl formamyl and diluted in acetate buffer (pH 4.9). The εlides were counterεtained with hematoxylin and mounted with an aqueouε mounting solution (GVA, Zymed Laboratories, South San Francisco, CA) . Spleen was uεed aε a poεitive tiεεue control and PBS inεtead of the firεt antibody as a negative control.

(e) Evaluation of Immunoperoxidase-Stained Uterine Sections

For each group of mice (5 uteri), three cross-sections from different areas of the block were evaluated. Localization of macrophages and lymphocyte subpopulations was noted deεcriptively. For quantitation, poεitive cells were counted in five random fields from each εection uεing a light icroεcope (X 500 magnification) . The mean numberε and εtandard deviationε of leukocyte

2 εubpopulationε per mm waε calculated. For statistical evaluation, Student's t-test waε uεed.

(f) Detection of Antisperm Antibodies by Indirect Immunofluorescence

The method used has been described in detail elsewhere (Madrigal et al . , 1986, Immunol. 9, 175-186). Epididymal sperm from DBA/2 retired breeders was washed and resuspended in PBS to 1 X 10 cells/ml and 5 μl of thiε suspension was applied to each spot of teflon-coated, eight-spot microscope slides (diameter 8 mm, Roboz Surgical). After drying, the slides were fixed in acetone for 15 min. and stored frozen at -70°C until uεe. On the day of aεεay, εlides were thawed, rinsed in PBS, then sera from immunized and non-immunized mice were added in four-fold serial dilutions and incubated for 1 h. Following a waεh step, fluorescein conjugated F(ab) ₂ fragment rabbit anti mouse IgG (Cappel) diluted 1:40 was applied for 30 min.

Following a final serieε of washes, coverslipε were mounted on the εlideε with a εolution of pheneylene diamine/glycerol and fluoreεcence patternε were analyzed on a Zeiss fluorescence microscope. (2) Results

Figure Summary: Fig. 1. shows the results of a single representative s.c. s.c./i.u. active immunization experiment with various control groups indicating reduced fertility and increased fetal reεorptionε in εperm-immunized mice. Immunization groupε: a, non-immunized; b, εaline + adjuvant; c, DBA/2 lymphocyteε + adjuvant; d, human RBC + adjuvant; e, DBA/2 εperm + adjuvant. Fig. 2. εhowε the effectε of paεsive transfer of T lymphocytes from s.c./s.c./i.u. immunized mice (from experiment depicted in Fig. l), on fertility of recipient mice. T lymphocyteε were from the groupε (a,b,c,d,e) identified in the deεcription of Fig. 1. Fig. 3. εhowε the mean numberε + S.D. of T lymphocyte subpopulations and macrophages in uterine sections of mice after s.c./s.c./i.u. immunizations (from experiment depicted in Fig. 1). Immunization groups were as described above (Fig. 1). Fig. 4. shows the mean numbers + S.D. of T lymphocyte subpopulationε and macrophageε in uterine sections of mice after pasεive tranεfer of T lymphocytes from s.c./s.c./i.u. immunized mice (same experiment as depicted in Figs. 1-3). Fig. 5 εhowε prei plantation embryo recovery in mice following allogeneic trophoblast immunization. Immunization groupε: f, εaline; g, adjuvant; adjuvant plus trophoblast. Fig. 6 εhowε the number of viable offεpring and fetal reεorption in mice following allogeneic trophoblast immunization. The groups are as described in Fig. 5.

(a) Comparison of Active Immunization Protocols

Different routes of immunization were compared to determine the moεt effective way to induce immunologic infertility with εperm antigenε. No εignificant difference was found between saline and sperm-immunized mice for any fertility parameter when either ε.c. or i.u. immunization

approaches were used excluεively (Table IV) . Intraperitoneal immunizations were discontinued after the first experiment because Freund's adjuvant alone produced inflammation and adheεionε that affected fertility. A combination of εyεtemic (s.c.) and local (i.u.) immunizations with sperm in adjuvant did not cause a significant decrease in the percentage of mice that become pregnant aε compared with the adjuvant control group (Table IV). However, a number of individual fertility para eterε were reduced in s.c./s.c./i.u. sperm-immunized animals as compared to adjuvant controls: on day 3 after mating the number of unfertilized oocyteε per pregnant mouse was significantly increased (P < 0.01) and the number of 2-cell embryos per mouse waε εignificantly decreased (P < 0.001); on day 15 of pregnancy the number of viable fetuses per pregnant mouse was εignificantly reduced (P < 0.005) and the number of fetal resorption siteε was significantly increased (P < 0.01) (Table IV, Fig. 1). (b) Pasεive Immunization Study

In the paεεive immunization εtudy, mice were adminiεtered T cell-enriched lymphoid cells from sperm-immunized, non-immunized and control-immunized syngeneic mice via the tail vein. The transferred cells were primarily CD4+ lymphocytes; less than 15% of the cells were macrophageε or CD8+ T cellε. Numbers of viable fetuses and resorption sites were stored on day 15 of pregnancy. There was no εignificant difference in any of the fertility parameters between mice receiving T cells from non-immunized mice and control immunized mice (Fig. 2). However, the fertility rate waε reduced in mice receiving lymphocytes from sperm-immunized animals (Table IV). Furthermore, the number of fetuses per pregnant mouse was significantly decreased in the group that received T cells from εperm-immunized mice (P < 0.001) and the number of reεorption εites was significantly increased (P < 0.005) (Table IV).

Table IV (i)

Effects of different sperm immunization protocols on fertility parameters. Active immunization*

Table IV (ii)

Effects of different sperm immunization protocols on fertility parameters.

Passive immunization T-cell transfter from s.c/s.c/i.u. Saline Sperm

2-Cell embryos NA NA

Fetuses per 6.2+4.0 2.6+2.0*** pregnant mouse (n=27) (n=28)

Resorption per 0.4+0.1 4.4+_2.2** pregnant mouse (n=27) (n=28) ^a Both saline and sperm immunizations were performed with Freunds's adjuvant

^• -'Fertility rate = (No. of pregnant mice/Total No. of mice) ⁰ ..

*?<0.01, **P<0.005, ***P<0.001

(c) Leukocyte Subpopulations in Uteri From Actively and Paεsively Immunized Mice

Uteri from pregnant and non-pregnant cycled animals (day 15 after hCG and mating) were studied. Sections of uteri from pregnant animals in both εperm-immunized and control groupε contained numerouε lymphocytes and macrophages and the patterns were indiεtinguiεhable from each other. In non-pregnant s.c./s.c./i .u. εperm-immunized mice, the number of T lymphocyteε/mm 2 of uteri•ne secti•on was εigni.fi.cantly higher than that obεerved in any control group (P < 0.001). Stage of eεtruε was not correlated with numbers of T cells in uterine horns of control mice or of sperm-immunized mice. Both CD4+ and CD8+ subpopulationε were obεerved in uterine tiεsue from sperm-immunized mice, but CD8+ cellε predominated. Moεt of the CD8+ cellε were located in the pe iglandular εpace, while CD4+ cellε were evenly εpread throughout the epithelium and periglandular region. In ut-ri from εperm-immunized mice, macrophageε were increaεed four-fold (P < 0.001) and were evenly diεtributed between epithelial, periglandular and interεtitial areaε. Data from a representative experiment are presented in Figε. 3 and 4.

When uteri from infertile paεεively immunized mice were studied, a significant increase in total T cells was found in the group which received T lymphocyteε from εperm-immunized animalε compared with controlε (P < 0.05). A εignificant increase in CD4+ cells was observed (P < 0.005); there waε not a εignificant increaεe in the number of CD8+ cellε. Macrophageε also were more prevalent in uterine tissues from the sperm passive-immunization group (P < 0.01). Both CD4+ lymphocytes and macrophages were evenly diεtributed throughout the uterine tiεεue. Very few B cellε were found in any of the uterine sections.

(d) Antisperm Antibody Evaluation

Titers of antisperm antibodies in pooled sera from s.c./s .c./s.c/ or i .u/i . ./i .u. εperm-immunized mice were 1/256 and 1/1014, respectively. Both groups showed a

predominantly acroεomal εtaining pattern. Pooled sera from s.c. s.c./i.u. εperm-immunized mice had a titer of 1/1014 and reactivity directed againεt both the sperm acroso e and tail regionε. Sera from mice which were paεεively immunized with lymphocyteε from actively sperm-immunized (ε.c./ε,c . /_..u. ) mice were negative for antiεperm antibodieε. Normal mouεe sera (negative control), sera from mice immunized with εaline and from mice injected with lymphocyteε from non-immunized mice were also negative.

(e) Trophoblast antigen Reεultε

The reεults of the trophoblast antigen immunizations are shown in figures 5 and 6. Immunizations were conducted using the above-described protocols for the immunization of mice with sperm. Specifically, C57BL/6 female mice were immunized with allogeneic trophoblast in complete Freund's adjuvant given subcutaneously, intraperitoneally and transcervically (intrauterine). These experiments demonstrate that trophoblast immunization resultε in a decreased number of recovered preimplantation embryos (Fig. 5), a εignificant reduction in the number of viable offspring and an increased number of fetal resorptionε in the trophoblast plus adjuvant immunized group as compared to the adjuvant alone and saline control groups (Fig. 6) . We have alεo obεerved that complete Freund's adjuvant injected between the implantation sites of day 11 viable C57BL/6 x DBA/2 fetuses will not only disrupt adjacent pregnancies, but will also cause recurrent abortion after two subsequent matings. These data provide further evidence of immunologic recurrent abortion involving trophoblast antigen activation. (3) Summary

C57BL/6 female mice were immunized with allogeneic (DBA/2) sperm in Freund'ε adjuvant either subcutaneously (s.c), transcervically into the uterine lumen (i.u.), or with a combination of s.c. and i.u. immunization approaches. Control mice received DBA/2 lymphocytes, human erythrocytes or saline in adjuvant using the same immunization protocols.

Immunization with εperm or control cellε in adjuvant excluεively by ε.c. or i.u. approacheε did not affect subsequent fertility, although sperm-injected mice from both protocols had high titerε of circulating antiεperm antibodies. In contrast, mice that were immunized with εperm in adjuvant by a combination of ε.c. and i.u. injectionε demonεtrated εignificant reductionε in fertilization rate and number of viable fetuεeε and an increaεed rate of fetal reεorption when compared with non-immunized and control-immunized mice. Mice receiving sperm by the ε.c./i.u. protocol had high titerε of antiεperm antibodies and a marked infiltration of T lymphocytes and macrophages into the uterine endometrium. To determine whether cellular immune mechanisms contributed to the infertility effect, T lymphocytes from spleens and pelvic lymph nodes of s.c./i.u. sperm-immunized mice and non-immunized mice were passively transferred to naive syngeneic female recipients which were subsequently mated. The total number of fetuses on day 15 of pregnancy was significantly reduced in mice receiving T-lymphocytes from sperm-immunized mice and a significant increaεe in fetal reεorption εiteε waε alεo obεerved. These mice did not have detectable titers of circulating antisperm antibodies, but had a significant infiltration of CD4+T lymphocytes and macrophages in the uterine epithelium and endometrium. Theεe data indicate that intrauterine antiεperm cell-mediated immunity can be induced in mice by a combination of εystemic and intrauterine immunizations and provide evidence for the existence of reproductive tract mucosal antisperm cellular immune responses that adversely affect fertility and pregnancy.

Example 4: T Helper 1-Type Cellular Immunity to Trophoblaεt Antigens in Women with Recurrent Spontaneous Abortion

(1) Patients

Peripheral blood was obtained from 244 non-pregnant women with a history of unexplained recurrent reproductive failure who were referred for evaluation to the Recurrent Miεcarriage Center, Brigham & Women's Hospital. The women were between 26 and 40 years of age and had a history of at least three prior first trimester spontaneous abortions with or without a prior ectopic gestation or live birth. The etiology of prior pregnancy losses was unexplained by conventional criteria (normal parental chromoεomeε, hysterosalpingography, luteal phaεe endometrial biopεy, hormonal analyεis, cervical cultures and antiphospholipid antibodieε) . The time of blood collection relative to the last pregnancy loss was not conεtant, but all women had experienced at leaεt one εpontaneous abortion within one year of testing. Peripheral blood was also obtained from a control group of 13 paid volunteer non-pregnant women between 27 and 41 years of age with a history of at least two uncomplicated term deliveries with their last birth occurring within one year of blood collection and no history of spontaneouε abort.on, ectopic pregnancy or εtill-birth. For an additional control, blood waε obtained from 10 men between 26 and 47 yearε of age. All of the individualε in this study were in excellent health, had no history of allergy or atopy, and were taking no medication at the time of blood collection.

(2) Aεεays

(a) Trophoblast Antigen Preparation A trophoblast antigen extract was prepared from the human choriocarcinoma cell line Jeg-3 (American Tissue Type Collection, Bethesda, MD) as described above (see alεo Hill, J. et al., 1992, Am. J. Obεtet. Gynecoi. 166:1044). Briefly, mycoplaεma-free Jeg-3 cellε were cultured in flasks until 80% confluence in Minimum Esεential Medium (MEM; Gibco, Grand

Iεland, NY) εupplemented with 10% fetal bovine εerum. The cellε were harveεted with a rubber cell scraper (Coaster, Cambridge, MA) and waεhed three times in Hank's Balanced Salt Solution (GIBCO) . The cells were then disrupted by Dounce homogenization (100 strokes) and the supernatant εaved after centrifugation at 400g for 10 minuteε. Protein concentration waε determined by BCA reagent kit (Pierce, Rockford, IL) and antigen extracts were adjusted to 333ug/mL. Fetal calf serum (GTBCO) was added to a final concentration of 10%, and 1 mL aliquots containing approximately 300ug of trophoblaεt extract each, were stored at -70°C until use.

(b) Peripheral Blood Mononuclear Cell (PBMC) Supernatants

Peripheral venouε blood waε collected into tubeε containing pyrogen-free sodium heparin (Riches, P. et al., 1992, J. Immunol. Method 153:125), and mononuclear cells were isolated by Ficoll-Hypaque centrifugation as deεcribed above (see also Hill, J. et al., 1992, Am. J. Obstet. Gynecol . 166:1044). Briefly, waεhed cellε were reεuspended to a concentration of 10 ⁸ cells/mL in Roswell Park Memorial Inεtitute RPMI medium Gibco εupplemented with 0.3 mmol/L glutamine, 100 IU/mL benzylpenicillin potaεεium and 100 ug/mL streptomycin sulfate (Sigma Chemical Co., St. Louis, MO) and 10% fetal bovine serum. Ten milliliterε of cell suspenεion and 1 mL of antigen extract and 1 mL of additional media for control were added separately to 50 mL tiεεue culture flaεks (Becton Dickinson, Lincoln Park, NJ) and cultured for 72 hours at 37°C in 5% CO 2 , 95% air. The cells were then washed and resuspended in 1 mL of Whitten's medium (Whitte., W.K., et al., J. Reprod. Fertil. 17:399-402 (1968)) supplemented with 0.3% bovine serum albumin (Mileε Scientific, Naperville, IL) . After an additional 24 hour incubation, cell εuspensionε were centrifuged and the εupernatantε were filtered through a 0.22um Millex GS filter unit (Miliipore, Bedford, MA) and stored at -70°C until use.

(c) Embryotoxic Factor Assay

This asεay waε performed aε described in Example 1.

(d) Cytokine Determination

Cytokines were measured in trophoblast-activated PBMC supernatantε by ELISA kitε according to the manufacturer' ε instructions. TNF-α, TNF-β, and IFN-gamma kits were obtained from Endogen (Boston, MA; lower limit of senεitivity, 10 pg/mL for TNF-α and IFN-gamma and 31pg/mL for TNF-β), IL-2 kits from R&D syεtemε (Minneapoliε, MN; loweεt limit of sensitivity, 31 pg/mL) , IL-4 kits, from Amerεham (International Place, UK; lowest limit of senεitivity, 31 pg/mL), and IL-10 kits from The Biosource CytoTM (Biosource International, Camarillo, CA; lowest limit of senεitivity, 15 pg/mL) .

All ELISA aεsays were εolid phaεe εandwich ELISAs in which specific monoclonal antibodieε were attached to wells in 96 well plates, and a secondary enzyme-conjugated specific antibody was added for cytokine detection. Following incubation with substrate, the colored product was measured in each well on a Dupont microplate reader (Dupont, Dover, DE) using a wavelength appropriate for the εubεtrate used. Cytokine concentrations were calculated from a standard curve generated with specific cytokine εtandardε provided with each kit.

All 244 culture εupernatantε from women with recurrent abortion and all 13 εupernatantε from parous controls were asεayed for IFN-gamma. Analyεiε of the other cytokines was performed on supernatants from a subgroup of 20 women chosen at random from the group with recurrent abortion and embryotoxic factor activity, and on all supernatants from the 13 women in the control group.

(e) Statistical Analysis

Nonparametric methods using Fisher'ε exact taεt were uεed to analyze data.

(3) Results: Lymphocyteε from a majority of women with unexplained recurrent abortion make T..l-type cytokines when exposed to trophoblast extracts, whereaε lymphocytes from parouε women with normal reproductive historieε and men take T _π 2-type cytokineε. Embryotoxic factor activity waε detected in trophoblaεt-stimulated mononuclear cell supernatantε from 160 (65.5%) of 244 women with unexplained recurrent abortion. None of the εupernatantε from parouε women with normal reproductive hiεtories or from men had embryotoxic activity. The T„l cytokine IFN-gamma was found in 125 (51.2%) of 244 supernatants from women with unexplained recurrent reproductive failure and significantly correlated with embryotoxic factor activity (75.6%) of 160 supernatants with embryotoxic activity vs 4 (4.8%) of 84 εupernatants without embryotoxic activity contained IFN-gamma (p<0.0001). The range of IFN-gamma levels in these εupernatantε waε 10.3 to 2295 pg/mL with a mean of 209 + 34 pg/mL.

All εupernatantε from the εubgroup of 20 women with recurrent reproductive failure and embryotoxic factor activity choεen for further study contained T _H l-associated cytokines (Table V). IFN-gamma (range 10.95-2,295 pg/mL; mean 517+141 pg/mL) and TNF-alpha (range 11.75-5646 pg/mL; mean 1199+388 pg/mL) were detected in all 20 caεeε. TNF-β waε detected in 17 of the 20 supernatants (range 47.6-578.1 pg/mL; mean 287 + 46.1 pg/mL) . IL-2 was initially detected (within the first 24 hourε) but waε not detected not detected after four days in any of the 20 supernatantε. Three supernatantε from the 20 women in thiε group contained detectable levels of T„π2-type cytokines: two contained low levels of IL-10 (17.19 and 20.34 pg/mL) , and one contained IL-4 (59.14 pg/mL) . Supernatants from unstimulated cultures of women with recurrent abortion contained neither T„ nor T„2 type cytokines. In contrast, none of the trophoblaεt-activated PBMC culture supernatants from the 13 parous women with normal reproductive histories or from the

Table V. T _H 1 and T _H 2 Cytokines in Supernatants of cultured Trophoblast-Antigen-Activated Peripheral Blood Mononuclear Cells from Nonpregnant Women with a History of Recurrent Spontaneous Abortion*

V V

*Data expressed as pg/mL

Table VI. T _H 1 and T 2 Cytokines in Supernatants of Cultured

Trophoblast Antigen Activated Mononuclear cells from Nonpregnant Women with

Normal Reproductive Histories*

THI TH2

IFN-gamma

1. 0

2. 0

*Data expressed as pg/mL

Table VII. Tyl and T _H 2 cytokines in supernatants of cultured trophoblast antigen mononuclear cells from men*

THI TH2

*Data expressed as pg/mL

10 men had detectable levels of T _H l-associated cytokineε

(INF-gamma, IL-2, TNF-β, or TNF-α); however, T„π2 cytokines were detected in all 13 supernatants from parous women (Table VI) and 9 of 10 εupernatants from men (Table VII). IL-10 was detected in all 13 supernatantε from parouε controlε (range 18.2-283 pg/mL; mean 65+23 pg/mL) , and IL-4 waε detected in three εupernatants (31.10, 40.58 and 81.65 pg/mL) . In men, IL-10 was identified in 9 of 10 trophoblast activated cell culture supernatants (range 22.4-10.90 pg/mL; mean 46 + 8.7 pg/mL), IL-4 was not demonstrable, while unstimulated cell culture supernatants from all parous women contained IL-10 (range 29.1-279.2 pg/mL; mean 76.8 + 19.2 pg/mL) . IL-10 was detected in 5 of 10 unstimulated cell culture supernatants from men (range 20.3-478.1 pg/mL; mean 260.2 + 75.4 pg/mL) . (4) Diεcuεsion

This study provides evidence that T lymphocytes from gravid women respond to trophoblast antigenε with a dichotomouε T helper immune response. Lymphocytes from many women (51.2%) with a hiεtory of unexplained recurrent reproductive failure produced T _H l-type cytokines following exposure to trophoblast cell extracts in vitro, whereaε lymphocyteε from women with a hiεtory of normal pregnancies and men when placed under identical conditions produced T„2-type cytokines. These data provide further evidence that trophoblaεt induces a T„2-type cytokine responεe that benefitε normal pregnancy, and of an abnormal T„l response to trophoblast antigens in women with recurrent pregnancy losε that playε a role in their reproductive dysfunction.

In this study peripheral blood mononuclear cells were obtained from 244 non-pregnant women with a history of unexplained recurrent reproductive failure by conventional testing criteria, 13 non-pregnant women with normal reproductive historieε, and 10 men. IFN-gamma was detected in 51.2% of the trophoblast-activated mononuclear cell εupernatants from 244 women with otherwise unexplained

recurrent reproductive failure and in none of the parouε controlε or men. Thiε data supports our hypothesiε for a new etiology for recurrent abortion based on an aberrant cellular immune response to trophoblast involving the T„l pathway.

To further define the cytokines produced by lymphocyteε of women with recurrent abortion following expoεure to trophoblast, and to investigate the poεεibility of T„2 immunity in normal pregnancy, εupernatantε from a εubgroup of 20 recurrent abortion patientε with embryotoxic activity, and εupernatants from 13 parous women and 10 men, none of which had embryotoxic activity, were εtudied in detail. TNF-α and TNF-β waε detected in 20 and 17 εupernatantε reεpectively from the recurrent abortion group, and in none from the parouε control group or from men. Theεe data provide further evidence of a T„l reεponse to trophoblast extracts in women with recurrent abortion since TNF-β iε a T„l-type cytokine and becauεe TNF-α is produced in higher amounts in a T„l response than in T„2 immunity (Romagnani, S. et al. , 1992, Int. Arch. Allergy Immunol. 4:279; Del Prete, G. et al . , 1993, J. Immunol. 150:353). The cytotoxic effect exerted in T-,1 immunity involves IFN-gamma, TNF-β and TNF-α (Tite, J. et al. , 1984, J. Immunol. 135:25). Our finding of elevated TNF-α levelε in trophoblaεt-activated mononuclear cell εupernatants from women with recurrent abortion εuggeεtε that theεe cytokineε, in addition to IFN-gamma, are part of an adverεe T„il immune response to trophoblaεt, thereby providing a new mechaniεm for reproductive failure. Accordingly, theεe reεultε εupport our hypotheεiε that theεe cytokineε are indeed, embryotoxic factorε.

Our finding of IL-10 in every trophoblaεt-activated mononuclear cell εupernatant from previously succesεful pregnant women without hiεtory of reproductive dysfunction provides evidence that normal human pregnancy is asεociated with the induction of a T„2-type immune reεponεe, and that TL-10 or other T _H 2-type cytokineε are involved in the maintenance of normal pregnancy, perhaps by suppreεεing

IFN-gamma and TNF-α production by T„l cells (Mosmann, T. et al., 1991, Immunol. Today 12:49; Maggi, E. et al. , 1992, J. Immunol. 148:2142). Our finding of IL-10 in 9 of 10 trophoblaεt-activated cell supernatantε from men provides further evidence that mononuclear cells normally respond to trophoblast with T„2-type cytokine production. The fact that IL-10 was detected in all unstimulated supernatantε from individualε not experiencing reproductive failure εuggests that IL-10 iε a natural occurrence.

The findingε of this study, together with data indicating that T„l and T„2 cells utilize distinct T-cell receptor-asεociated εignal tranεduction pathwayε (Gajewεki, T. et al., 1990, J. Immunol. 144:4110), and that high concentrationε of anti-CD3 monoclonal antibody inhibit T„l but not T„2 responεeε ( illiamε, M. et al. , 1990, J. Immunol. 144:1208) enableε the development of new immunotherapeutic regimenε for preventing human reproductive failure due to T„l mediated eventε. Additional therapieε involving the adminiεtration of an immunomodulating agent (to modulate the concentration the above-identified cytokineε that play a role in immunologic reproductive failure) , antibodieε to T„l cytokineε that abrogate the biological activity of theεe cytokineε, or adminiεtration of intravenouε immunoglobulin through anti-idiotype binding to T-cell receptor idiotypeε, εhould also prove beneficial in preventing reproductive failure due to T„l immunity.

Example 5: Diagnosis of Immunologic Reproductive Failure using a Lymphocyte Proliferation Assay (1) Subjects

Blood waε obtained from 57 non-pregnant women with a history of recurrent abortion at the Recurrent Miscarriage Center, Brigham and Women'ε Hoεpital between February and June, 1993 in accordance with Human Subjectε' Committee approval. The women, were between 29 and 41 years of age and had a history of at least three prior firεt trimeεter

εpontaneouε abortions, with or without a prior ectopic gestation or live birth. All women were evaluated according to a standard protocol by a single investigator (JAH) as previously described. Briefly, after obtaining a thorough history, all woman had a phyεical examination, hysterosalpingography, peripheral blood chromoεome aεεeεsment of both partnerε, luteal phase endometrial biopsy, hormonal analysis, and antiphospholipid antibody determinationε. Women who were normal by these criteria were designated as having recurrent abortion of unknown etiology. Peripheral blood was also obtained from a control group of 10 non-pregnant, paid volunteer women who were randomly εelected from the Obεtetrical Nurεing Service. Women in thiε control group were between 27 and 43 yearε of age with a hiεtory of at leaεt two uncomplicated term deliverieε and no history of either spontaneous abortion, ectopic pregnancy, or still-birth. None of the women in this study were taking medication at the time of blood collection. (2) Assays

(a) Antigen Preparation

Trophoblaεt antigenε were prepared from the human choriocarcinoma cell lines Jeg-3 and JAR (American Tisεue Type Collection, Betheεda, MD) . Theεe cellε were cultured in flaεkε until 80% confluence in Minimum Eεsential Medium (MEM; GIBCO, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS) . The cells were harveεted without trypεinization with a rubber cell εcraper (Coaεter, Cambridge, MA) and waεhed three times in RPMI 1640 media (GIBCO) . The cells were then disrupted by dounce homogenization (greater than 100 εtrokeε) and the supernatant was saved after centrifugation at 400g for 10 minuteε.

For fractionated trophoblaεt antigenε, Jeg-3 cellε were harveεted without trypεinization aε above and waεhed three times in ice-cold phosphate buffered saline (PBS) . The cell pellet was reεuεpended in ice-cold hypotonic buffer (lOmM Tris-HCl pH 7.6, 0.5 mM MgCl ₂ ) for 10 minutes and the cell

εuεpenεion waε added to a chilled Dounce homogenizer and delivered 35 εtrokes. Tonicity reεtoration buffer (lOmM Triε-HCl, pH 7.6, 0.5mM MgCl ₂ , 0.15M NaCl), waε then added to the homogenized cellε, and the mixture was centrifuged at 500g at 4°C for 5 minutes to obtain a crude nuclear pellet (designated aε Nuclear Fraction) . The εupernatant was then centrifuged at 20,000g, 4°C for 30 minutes to iεolate mitochondria, Golgi, microsomes and cell debris (designated the Organelle Fraction) . A membrane-enriched fraction (Membrane Fraction) waε εeparated from the Cytoεol Fraction after an additional ultra-centrifugation at 150,000g, 4°C, for 45 minutes. Nuclear, Organelle and Membrane antigen pelletε were εolubilized by incubation at 60°C for 5 minuteε in 20uL of 0.2% sodium dodecyl sulphate in 10 mM Triε-HCl, pH 7.6 and protein concentrationε were determined by BCA reagent kit (Pierce, Rockford, IL) . All trophoblaεt antigen sources were sterilized by gamma-irradiation (150 Gy) and stored at -70°C until use. A variety of concentrations of Jeg-3 extracts and Cytosol Fraction and Membrane Fraction antigenε were teεted to establish dose-dependent stimulation curves. Percent stimulation index (% SI) was calculated according to the following formula: % SI - cpm at the individual antigen concentration used/the higheεt cpm at any antigen concentration x 100. Additionally, the Jeg-3 antigen extract waε trypsinized with insoluble trypsin (GIBCO) at 37°C for 90 min or heat-inactivated at 70°C for 90 min, and teεted for antigenicity to determine whether the antigenε were proteinε. Red blood cellε (RBC) were εeparated from donorε ' white blood cells after Ficoll-hypaque gradient centrifugation and removal of buffy coat containing neutrophils. RBC were incubated in hypotonic solution and waεhed three timeε at 500g, 4°C for 10 minutes to remove hemoglobin. RBC membrane control antigen was then prepared after dounce homogenization and centrifugation as described above.

(b) Lymphocyte proliferation aεεay

Peripheral blood mononuclear cellε were iεolated from heparinized blood by Ficoll-hypague gradient centrifugation.

₅ Washed cells (2 x 10 ) were cultured in 96-well round-bottom microtiter plateε (Corning, Corning, NY) containing RPMI 1640 media, 10% human εerum (type AB, GIBCO) and antibioticε in a final volume of 200 uL. Twenty microtiterε of antigen extract or medium alone were added to triplicate wells. After a 6 day incubation at 37°C, 5%

CO,, 95% air, the cultureε were pulsed with 0.5 uCi of

[ 3H]-thymidine (New England Nuclear, Boston, MA) and harveεted 16 hourε later on a glaεε fiber filter. Liquid εcintillation in a beta counter (Beckman, Palo Alto, CA) waε used to determine lymphocyte proliferation expresεed aε countε per minute (cpm). The εtimulation index (SI) waε calculated according to the formula; SI - cpm in presence of antigen / cpm in absence of antigen. A SI greater than 3 was considered positive (significant lymphocyte proliferation in responεe to antigen) . All experiments were performed by the same investigator without knowledge of sample source or embryotoxic factor results.

(c) Embryotoxic Factor Asεay

This assay was performed as described above (Example 1). Briefly, peripheral blood mononuclear cellε were iεolated from heparinized blood by Ficoll-Hypaque gradient centrifugation aε deεcribed for the lymphocyte proliferation assay. Washed cells were reεuεpended in lOmL of RPMI medium εupplemented with 10% human εerum to a concentration of 10 cellε/mL. The cellε were cultured with or without Jeg-3 antigen (30ug/mL) for 96 hourε at 37°C in 5% C0 ₂ , 95% air. After waεhing, the cell pellet waε reεuεpended in 2mL of Whitten'ε media εupplemented with 0.3% bovine εerum albumin. After an additional 24 hour incubation, cell εuεpenεions were centrifuged and the supernatants were filter-sterilized and εtored at -70° until uεe. Two-cell embryos were harvested from εuperovulated CF-1 female mice that had been bred to

C57BL6 male mice. All media and supernatants were equilibrated overnight in 5% Co- before the addition of embryos. Embryos were cultured in 20uL drops containing a 1:1 dilution of study supernatant in Whitten's medium εupplemented with 0.3% bovine εerum albumin under C0 ₂ -equilibrated paraffin oil in tiεεue culture diεheε (Falcon, Oxnard, CA) . At leaεt 11 embryoε were cultured in each 20uL drop. The embryoε were aεεeεεed for normal development after 4 dayε in culture by the criteria described by Ducibella (Ducibella T., 1980, Dev. Biol., 79:356). Embryotoxic factors were considered present when the percentage of embryos advancing to the blastocyst stage of development was lesε than 50% of control valueε aε previouεly deεcribed (Hill J. et al. , 1992, Am. J. Obεtet. Gynecol. 166:1044; Ecker J. et al. , 1993, Obεtet. Gynecol. 84). The embryotoxic factor assay was performed by a single investigator without knowledge of sample source or results of the lymphocyte proliferation asεay.

(d) Statiεtical analyεiε

All data were analyzed by one factor analyεiε of variance. Fisher's PLSD (Protected Leaεt Significant Difference) teεts were used for poεt hoc pairwiεe compariεon. Regreεεion analysis was uεed to correlate lymphocyte proliferation with embryotoxic factor production. (3) Reεultε: Lymphocyte proliferation in reεponse to trophoblast antigen stimulation was εignificantly higher in women with recurrent abortion of unknown etiology than in fertile controlε, and εignificantly correlated with embryotoxic factor activity in culture εupernatantε.

The optimal concentration of trophoblaεt antigen derived from Jeg-3 extractε (30 ug/mL) uεed in the lymphocyte proliferation aεεay waε determined from doεe response curves from εix women reεponεive to Jeg-3 out of 12 women initially teεted (Fig. 7) .

Only one of the 57 women referred for evaluation of recurrent abortion was found to have a presumed etiology for reproductive loss. This individual had a εeptate uterus and underwent hysteroscopic resection. She did not respond to trophoblast antigen stimulation in either the lymphocyte proliferation asεay (SI - 1.4) or the embryotoxic factor assay (embryo development greater than 50%) .

Thirty out of 57 (52.6%) women with recurrent abortion had an SI greater than 3 to either Jeg-3 or J7AR; 14 (46.7%) responded to antigens derived from both Jeg-3 and JAR; while 13 of the woman (43.3%) responded to Jeg-3 only, and three (10.0%) to JAR only. Because more women responded to Jeg-3 than JAR, antigen fractionation εtudieε were performed with the Jeg-3 cell line.

Following stimulation with the optimal concentration of trophoblast antigen extract (30 ug/ml) the mean SI of the 57 women with a history of recurrent abortion was significantly higher than that of the control group (3.99+2.83 vs. 1.64+0.34, p<0.01). The mean SI in response to RBC membrane antigen was not εignificantly different between groupε (1.33+0.63 vε 1.10+0.29). When the lymphocyte proliferation reεponses to trophoblaεt antigen were divided into two groupε, poεitive (SI>3) and negative (SI<3), 52.6% of women (n=30) with recurrent abortion were poεitive (mean SI = 5.89+2.72) and 47.4% (n=27) were negative (mean SI = 1.89+ 0.58). Forty-four out of 57 women had an SI>2 in reεponεe to trophoblaεt antigen while all fertile controlε had a SI leεs than 2. The difference between women with recurrent abortion and fertile controls was εtatistically significant by Fisher exact tests (two tailed, p<0.002).

To determine which cellular fraction(s) were reεponεible for lymphocyte εtimulation in recurrent abortion patientε, lymphocytes from 27 women that had responded to the crude Jeg-3 antigen extract (30ug/mL) were teεted againεt antigen extracts made from Jeg-3 derived Nuclear, Organelle, Cytosol and Membrane fractions. The optimal concentration of each

antigen extract was lOug/mL, as determined by doεe reεponse studieε in the 6 positive women who had originally responded to Jeg-3 (data not shown). As shown in Fig. 8, trophoblast antigen derived from cytosol and membrane componentε were responsible for lymphocyte proliferation in the majority of individuals (55.6% and 37.0%, reεpectively). Antigenicity waε completely abrogated after trophoblaεt antigen treatment with either trypsin or heat (data not shown) .

Supernatants from two women with recurrent abortion had to be discarded due to bacterial contamination of the cultures. Therefore, embryotoxic factor data were only available from 55 of the original 57 study women. Embryotoxic factor activity was detected in lymphocyte εupernatantε from 36 of 55 women with recurrent abortion (66%) following εtimulation with crude trophoblaεt antigen extract. None of the supernatants from fertile control women contained embryotoxic factor activity, and supernatants from cells cultured without trophoblast antigen from both recurrent abortion and control groups did not have embryotoxic factor activity. Lymphocyte proliferation responses significantly correlated with the production of embryotoxic factor (r=0.61, Fig. 9). Embryotoxic factor production in response to trophoblast antigen stimulation as determined by impaired embryo development was found in 27 (90.0%) out of 30 women with recurrent abortion who had a SI>3 in their lymphocyte proliferation asεay in reεponse to trophoblast antigen εtimulation, while 9 (36-0%) out of 25 woman with recurrent abortion had evidence of embryotoxic factors despite a S.I.<3. The senεitivity and εpecificity for the lymphocyte proliferation aεεay compared to the embryotoxic factor aεεay waε 75% and 84%, reεpectively, with a poεitive and negative predictive value of 90% and 64%. (4) Diεcuεεion

In the preεent εtudy we provide further evidence that cellular immunity to trophoblaεt iε involved in the pathogeneεiε of immunologic reproductive failure, since 52.6%

of our patientε had evidence of activated lymphocyte reεponεeε to trophoblaεt antigen(ε) and 90.0% of theεe patientε produced embryotoxic factorε.

Many εtudies using maternal lymphocytes in proliferation aεεayε have reported depreεεed cellular immunity during pregnancy in reεponεe to the T-cell mitogen, phytohemagglutinin, although not all inveεtigatorε have corroborated theεe reεultε (reviewed in Goodfellow C, 1983, Immunol. Rev. 75:61). Few εtudieε have uεed lymphocyteε from women with recurrent abortion and none have teεted lymphocyte proliferation in reεponse to trophoblast antigen(s) free of human leukocyte antigen (HLA) expresεing cells. Other investigators have used placental cellε derived from chorionic membranes or normal human placenta as antigenic sourceε with conflicting reεultε reported (Goodfellow C. , 1983, Immunol. Rev. 75:61; Hunt J. et al. , 1984, J. Reprod. Immunol. 6:377). Stimulation may be due to contamination by nontrophoblaεt cellε εuch aε fetal or maternal white blood cells expressing HLA. In our εtudy, we uεed aε our trophoblaεt antigen-source choriocarcinoma cell lines to provide a consiεtent trophoblaεt antigen without contamination by other cellε expreεεing HLA moleculeε. Becauεe the reεponεe to trophoblaεt antigen was so εtrong in women with recurrent abortion, we uεed a more conεervative SI (greater than 3) in contraεt to the more conventional SI of greater than 2 to define εignificant lymphocyte proliferation in reεponεe to antigen. Jeg-3 and JAR cellε were found to be εuitable antigenic εourceε, with Jeg-3 cellε being more antigenic than JAR in εtimulating lymphocyte proliferation in women with unexplained recurrent abortion. The trophoblaεt antigen(ε) reεponsible for activated lymphocyte responεeε were abundant in both cytoεol and membrane fractions. Whether cell-mediated immunity was due to a single antigen which located in cytoεol and membraneε, or due to a variety of antigenε distributed in both fractions remains as yet unknown. However, our data provide evidence that responsible

antigenic epitopes contained a peptide/protein component since antigenicity was abrogated by trypεinization or heat treatment. Lymphocyte proliferation waε not obεerved in any of the samples following incubation with RBC membrane antigens. RBC membranes were chosen for the control antigen becauεe, like villouε trophoblaεt, RBC do not expreεs HLA.

The lymphocyte proliferation asεay iε an effective method to diεtinguiεh women with activated cellular immunity to trophoblaεt which may contribute to their recurrent abortion. This technique, in addition to the asεayε deεcribed in Example 4, is useful in the diagnostic evaluation of women with immunologic reproductive failure and has applicability regarding the testing of potential therapeutics such as the above-described immunomodulating agent drug therapies.

Equivalents

Each of the above-identified references iε incorporated herein by reference.

Thoεe εkilled in the art will be able to ascertain, using no more than routine experimentation, many equivalents of the specific embodiments of the invention described herein. These and all other equivalents are intended to be encompaεsed by the following claims.

What iε claimed iε: