Regulation of innate and adaptive immunity by the female sex hormones oestradiol and progesterone
*Corresponding author. Tel.: +61 (2) 49236157; Fax: +61 (2) 49236488, E-mail address: firstname.lastname@example.org
Women mount more vigorous antibody- and cell-mediated immune responses following either infection or vaccination than men. The incidence of most autoimmune diseases is also higher in women than in men; however, during pregnancy many autoimmune diseases go into remission, only to flare again in the early post-partum period. Successful pregnancy requires that the female immune system tolerate the presence of a semi-allogeneic graft for 9 months. Oral contraceptive use can increase susceptibility to certain genital tract infections and sexually transmitted diseases in women. Moreover, treatment of mice and rats with female sex hormones is required to establish animal models of genital tract Chlamydia, Neisseria and Mycoplasma infection. This review describes what is currently known about the effects of the female sex hormones oestradiol and progesterone on innate and adaptive immune responses in order to provide a framework for understanding these sex differences. Data from both human and animal studies will be reviewed.
The immune system associated with the female reproductive tract (FRT) is particularly important because it is the first site of immunological contact with pathogens such as chlamydia, gonorrhoea, syphilis, human immunodeficiency virus (HIV) and human papillomavirus (HPV). The extent of this problem is emphasised by the fact that there are an estimated 333 million new cases of non-HIV sexually transmitted diseases (STDs) each year (WHO). There are, however, constraints on immune responses at this site. To be compatible with reproductive functions, immune responses against allogeneic spermatozoa or the developing foetus must be prevented. The vagina, cervix and uterus of human females and many animal species contain the full complement of immune cells that are responsible for both innate and specific immunity. However, the numbers and activity of most of these cell types vary significantly throughout the phases of the reproductive cycle and this is believed to be controlled by changes in the levels of the female sex hormones oestradiol and progesterone [1,2]. For example, recruitment of neutrophils into the FRT, antibody levels in cervicovaginal fluids, T cell responses and susceptibility to STDs are all affected by the stage of the oestrous cycle. In this review, we will describe the effects of the female sex hormones oestradiol and progesterone on both innate and adaptive immunity and review how these hormones affect susceptibility to STDs, vaccine-induced immune responses and the development of autoimmune disease.
2Hormonal effects on innate immunity
Innate immune responses are activated by the binding of microbial pathogen-associated molecular patterns (PAMPs) to Toll-like receptors (TLR) found on phagocytic cells and epithelial cells [3,4]. This PAMP-TLR binding results in the secretion of antimicrobial peptides (defensins), secretion of chemokines that attract phagocytic cells to the affected tissue and secretion of type 1 interferons that attract and activate natural killer (NK) cells. Epithelial cells in the FRT express TLR-4 and CD14 (G. Yeaman, personal communication) as well as TLR1, 2, 3, 5 and 6 . However, it is not known if changes in sex hormones affect TLR expression.
Defensins, or cationic peptides, represent an important component of innate immune protection at mucosal surfaces, including the FRT [6,7]. Several of these broad-spectrum natural antimicrobial peptides are expressed in urogenital tissues and their expression may be regulated by cycle-associated changes in sex hormones. For example, human intestinal defensin-5 (HD-5) mRNA is expressed in the vagina, ectocervix and inflamed fallopian tube and variably in the endocervix, endometrium and fallopian tube . Endometrial expression of HD-5 mRNA was highest during the early secretory phase of the cycle. Concentrations of secreted HD-5 peptide in cervicovaginal lavage were also highest during the secretory phase of the menstrual cycle. Human β-defensin 1 (HBD-1) mRNA has also been detected in human endocervix, ectocervix and vagina and the HBD-1 protein was detected in vaginal lavage fluid . This study did not determine if oestrous cycle stage influenced HBD-1 expression, but this is likely since levels of other antimicrobials such as lactoferrin and lysozyme are affected by oestradiol concentrations and stage of the oestrous cycle [10,11]. Apical rinses obtained from cultures of polarised uterine epithelial cells recovered from women at the proliferative and secretory phases of the menstrual cycle, but not from epithelial cells obtained from postmenopausal women, were able to kill both Staphylococcus aureus and Escherichia coli. This cytotoxic activity was shown to be due partly to the secretion of secretory leukocyte protease inhibitor (SLPI). The endometrial epithelial cell line HEC-1B was also shown to secrete SLPI . SLPI is produced by normal human endometrium, first trimester decidua and trophoblast . Endometrial expression of SLPI was found to be menstrual cycle-dependent with highest secretion occurring in the secretory phase. Collectively, these data indicate that menstrual status may regulate this important ‘first line’ of defense against microbial infection in the FRT and may explain why susceptibility to infection with some sexually transmitted microbes is greater at certain stages of the menstrual cycle (see below).
Recruitment of neutrophils into the vaginal epithelium of mice is controlled by hormone-regulated increases in the mouse interleukin (IL) 8 homologue MIP-2, that peaks at metoestrus-2 coincident with the peak of neutrophil infiltration . Similarly, the increase in macrophages in the FRT observed during luteolysis is associated with a hormone-dependent increase in levels of the chemokine MCP-1 in several species [15–17]. In the mouse ovary, macrophage numbers are significantly higher at proestrus and metoestrus compared to dioestrus and oestrus . Macrophages are also found throughout the endometrium and myometrium of the mouse uterus . Macrophages were evenly distributed throughout the uterine tissues during dioestrus but were concentrated in the sub-epithelial stroma during proestrus and oestrus. Ovariectomy resulted in a decrease in the number of uterine macrophages and this could be reversed by hormone administration . Changes in the distribution of macrophages at different stages of the cycle were also seen in the rat uterus and vagina . In this study MHC class II-positive cells were detected using the OX-6 monoclonal antibody and macrophages, granulocytes, and dendritic cells (DC) were detected by OX-41 monoclonal antibody. MHC class II-positive cells, macrophages, granulocytes and DC were present in large numbers in the stroma of the endometrium and around the glandular epithelium in the uterus at oestrus, the stage of the reproductive cycle when oestradiol levels are known to be high, relative to those seen at dioestrus, when oestrogen levels are low and progesterone is the predominant hormone. Immune cells were more numerous in the vagina, relative to the uterus. OX-6- and OX-41-positive cells were present in greater numbers in the sub-epithelial layers of the vagina at dioestrus, in contrast to oestrus. In the bovine ovary, eosinophils appear to be selectively recruited during the periovulatory period whilst mast cell numbers were not affected by the stage of cycle .
NK cell activity in the FRT is also influenced by hormone changes and in women decreases significantly following menopause. Interestingly, hormone replacement therapy was recently shown to restore NK cytotoxicity in postmenopausal women to premenopausal control levels . The endometrium of both humans and mice contains a unique subset of uterine-specific NK cells (uNK). The number of these cells also appears to be regulated by sex hormones. The number of uNK cells increases in the mid–late secretory phase of the menstrual cycle and they are also abundant early in pregnancy but progressively disappear from mid-gestation onwards and are absent at term. uNK cells express the oestrogen receptor β1 and glucocorticoid receptor but not the oestrogen receptor α and progesterone receptors , suggesting that their function(s) are controlled mainly by oestradiol. Local expansion of uNK cells in the uterine decidua is driven by IL-15 , and the secretion of interferon γ (IFNγ) and IL-18 [25–27] by NK cells appears to be essential for successful implantation and pregnancy. How, or if, these functions are regulated by pregnancy-associated changes in oestradiol and progesterone remains to be determined, although IL-15 mRNA in human endometrium is significantly increased during the secretory phase of the cycle, compared to the proliferative phase, and IL-15 production by endometrial stromal cells is enhanced by progesterone . Further discussion of the role of uNK cells in implantation, foetal survival and parturition is outside the scope of this review and the reader is referred to recent review articles [29,30].
Oestradiol may also have generalised anti-inflammatory activity. This has been demonstrated in animal models of carrageenan-induced lung injury in the rat , burn injury-induced organ inflammation in rats  and liver damage induced by warm ischaemia/reperfusion injury in mice . Oestradiol has also been shown to prevent lipopolysaccharide-induced inflammatory responses in primary cultures of rat microglia . In these systems, the anti-inflammatory effects of oestradiol may be due to a combination of decreased tumour necrosis factor α (TNFα) production, reduced neutrophil recruitment into affected tissues or inhibition of inducible nitric oxide synthase activity.
There is a growing body of evidence showing that the innate immune response determines the type of adaptive immune response elicited by an infection [35,36]. Thus, hormonal effects on the innate immune response to infection in the FRT may affect subsequent adaptive immunity to STDs such as chlamydia and HIV .
3Hormonal effects on antigen presentation
Regulation of antigen presentation represents a key control point for induction or suppression of adaptive immune responses. In the FRT, sex hormones influence antigen presentation by both professional antigen-presenting cells (APCs: DC and macrophages) and non-conventional APCs found within uterine and vaginal tissues (epithelial cells and stromal cells) [38,39]. The best evidence for a direct effect of sex hormones on DC function comes from studies of the suppressive effects of oestradiol on the development of experimental autoimmune encephalomyelitis (EAE), a mouse model of multiple sclerosis (MS) [40–42]. EAE is due to the induction of myelin basic protein (MBP)-specific T cells with a T helper 1 (Th1) cytokine profile that mediate spinal cord injury. Pretreatment of mice with oestradiol, prior to the induction of EAE, significantly decreased DC numbers in the brains of treated mice. Furthermore, the frequency of CD11c+/CD8α+ DC producing TNFα and IFNγ was reduced in oestradiol-treated mice compared to control mice with EAE . Oestradiol treatment of splenic DC, matured in granulocyte-macrophage colony-stimulating factor and IL-4, decreased TNFα, IFNγ and IL-12 production by DC and suppressed their ability to present antigen to MBP-specific T cells. Furthermore, when oestradiol-treated DC were cultured with MBP-specific T cells and antigen, a shift towards the production of the T helper 2 (Th2) cytokines IL-4 and IL-10 and a concomitant decrease in production of the Th1 cytokines TNF and IFNγ was noted . We have also shown that oestradiol treatment of bone marrow-derived DC inhibits their ability to present antigen to naive CD4 transgenic T cells (personal observation). Thus, oestradiol can have multiple effects on DC function that may include inhibition of DC homing or migration, suppression of antigen presentation and shifting of helper T cell cytokine production towards a Th2 profile.
Vaginal and uterine tissues from both humans and rats contain cells that can present antigen to memory T cells. Mixed uterine cell suspensions, prepared from hysterectomy tissues isolated from pre- and postmenopausal women, were assayed for their ability to present antigen to autologous T cells. Uterine and vaginal cells from women at the proliferative and secretory stages of the cycle and following menopause were able to present tetanus toxoid antigen to sensitised T cells. This was the first demonstration that human reproductive tract cells present antigen and that antigen presentation varies with both the site in the reproductive tract analysed and the stage of the cycle. Subsequently, it was shown that freshly isolated epithelial cells from the uterine endometrium constitutively express MHC class II and could process and present tetanus toxoid antigen to autologous T cells [44,45].
Antigen presentation by purified rat uterine and vaginal cells was demonstrated by studies in which reproductive tract cells were incubated with ovalbumin (OVA)-specific T cell lines and OVA [38,39,46,47]. Analysis of cells from rats throughout the reproductive cycle indicated that antigen presentation by uterine epithelial cells was high at dioestrus (low oestradiol) and low at oestrus (high oestradiol), antigen presentation by uterine stromal cells was low at dioestrus and high at oestrus and vaginal cell antigen presentation was high when oestradiol levels in blood were low (dioestrus). These studies indicate that the FRT of the rat contains non-professional APC that can activate memory T cells and that this function may be either enhanced or suppressed depending on tissue site and endocrine balance.
The suppressive effects of oestradiol on antigen presentation by both DC and non-conventional vaginal APC may be mediated through oestrous cycle stage-dependent production of transforming growth factor β (TGFβ). Migration of Langerhans cells from the skin  and the ability of DC to activate T cells [49–51] are inhibited by TGFβ, due to inhibition of DC maturation . It has recently been demonstrated that oestradiol-induced inhibition of antigen presentation by rat vaginal cells was partially reversed by incubation of vaginal APC with anti-TGFβ antibody . Furthermore, secretion of TGFβ by isolated vaginal cells was significantly higher in oestradiol-supplemented cultures compared to controls . Thus, primary activation of naive CD4 T cells by DC, which have migrated into the FRT, or activation of memory T cells by vaginal (and possibly uterine) APC may both be inhibited at the oestradiol-dominant phase of the oestrous cycle as a result of local production of TGFβ. The function of APC in the FRT following sexual intercourse may also be inhibited by seminal plasma, which has been shown to contain high levels of TGFβ[54,55]. In seminal plasma, TGFβ is found mainly in the latent form; however, the pH of the human vagina would likely cause conversion to the active form of TGFβ that could then inhibit APC function. Interestingly, the suppression of immune responses induced by feeding of protein antigens, the phenomenon of oral tolerance, also involves TGFβ effects on DC and the induction of TGFβ-secreting Th3 cells [56,57].
4Hormonal effects on T cell immunity
Studies of human uterine endometrium have demonstrated the existence of large lymphoid aggregates (LA) consisting predominantly of a central core of B lymphocytes surrounded by large numbers of CD8+ T lymphocytes . The size of these LA was greatest at the secretory phase of the cycle when levels of both oestradiol and progesterone increase. At this stage, LA contained 2000–3000 cells and were distributed between the endometrial glands in the stratum bursalis. Such LA were significantly smaller during the proliferative stage of the cycle and were absent in tissues from postmenopausal women. Analysis of cell division and distribution of T cell receptor variable region β expression within LA in the endometrium suggest that LA develop largely by trafficking of cells into the nucleation sites within the endometrium, rather than by division of a limited number of resident precursor cells . This could involve sex hormone-mediated regulation of chemokine secretion throughout the menstrual cycle, although this remains to be proven. LA also exist in the mouse uterus . Mouse uterine LA contain equal numbers of CD4 and CD8 T cells, more macrophages but fewer B lymphocytes than are found in human LA. In rat uterine tissues, CD8 T cell numbers were also influenced by the stage of the cycle, being highest at oestrus .
Cytotoxic T lymphocyte (CTL) activity in the human FRT is also cycle stage-dependent . CTL activity in the uterine endometrium is present during the proliferative phase of the menstrual cycle and absent during the subsequent secretory (postovulatory) phase. In contrast, in postmenopausal women the entire reproductive tract, including the uterus, retains the capacity for strong CD3+ T cell cytolytic activity. These findings suggest that the high levels of oestradiol and progesterone present during days 14–28 of the menstrual cycle down-regulate CTL activity in the uterus. Cytotoxic activity of decidual lymphocytes may also be controlled by local sex hormone concentrations as high levels of progesterone block perforin expression in these cells . T cell receptor γδ intraepithelial lymphocyte numbers in the vaginal epithelium also vary throughout the oestrous cycle, being highest at proestrus . Numbers of CD8 T lymphocytes also fluctuate according to the stage of the cycle in the bovine corpus luteum with numbers increasing significantly just before luteolysis .
Women consistently mount higher Th2 responses than men do when peripheral blood mononuclear cells (PBMC) are activated by polyclonal activators such as phytohaemagglutinin , consistent with higher immunoglobulin production and increased prevalence of some autoimmune diseases (see below). Recent studies in mice, however, do not support a role for oestradiol in the preferential induction of Th2 responses . In this study, female mice were ovariectomised then transplanted with slow-release oestradiol pellets, which maintain a continuous serum oestradiol concentration of 60–100 pg ml−1, equivalent to levels found during late pregnancy in the mouse. Control mice were ovariectomised only. Following immunisation with OVA in incomplete Freund's adjuvant, draining lymph node CD4 cells were stimulated in vitro with various doses of OVA. Even in Th2-prone BALB/c mice, OVA-induced proliferation and IFNγ secretion were significantly higher in lymph node cultures from oestradiol-supplemented mice compared to non-supplemented controls. IL-4 secretion was significantly reduced in lymph node cultures from oestradiol-supplemented mice. These authors went on to show that the oestradiol-induced increase in Th1 cell priming was independent of any hormone effect on APC and required a functional oestrogen receptor α. Despite the strong enhancement of Th1 T cell responses caused by continuous oestradiol supplementation a significant decrease in the absolute numbers of both CD4 and CD8 T cell numbers was observed in hormone-treated animals, consistent with other studies demonstrating oestrogen-mediated reductions in thymocyte numbers . This could be explained by the observation that oestrogen suppresses both IL-2 production and IL-2 receptor expression by human peripheral blood T lymphocytes and CD4+ T cell lines . This would prevent IL-2-mediated T cell expansion following antigen- or mitogen-induced activation. The mechanism of this suppression was due in part to inhibition of the positive regulatory IL-2 transcription factors NF-κB and AP-1 by oestradiol. Studies showing that oestradiol markedly increases the activity of the IFNγ promoter in human T cell lines and increases IFNγ mRNA expression in mitogen-stimulated murine spleen cells  also suggests that oestradiol may favour Th1 responses. The differences between the Th2-inducing effects of oestradiol demonstrated in studies with human T cells and the Th1-inducing effects of oestradiol demonstrated in studies with mouse CD4 cells could be explained if oestradiol had biphasic effects on T cell development, depending on the concentration used. This may in fact be the case as recent studies, using proteolipid protein-specific human T cell clones, suggest that oestradiol may enhance the secretion of IFNγ at low concentrations but enhance IL-10 secretion by the same cells at higher concentrations (>5000 pg ml−1) [70,71].
While further studies will be required to determine the effects of sex hormones on T cell function, receptors for oestradiol have been demonstrated on both CD8 and CD4 T cell populations and thymocytes [40,66,72–74]. The presence of progesterone receptors on T lymphocytes remains controversial [75,76]. Future studies will need to consider hormone concentrations representative of those found during normal cycling and during pregnancy as well as time of exposure to hormones, which will differ significantly between human and murine lymphocytes. Furthermore, cycle-associated changes in hormone levels may affect T cell function directly, or via modulation of APC function (see Section 3).
5Hormonal regulation of immunoglobulin secretion
Levels of IgG and IgA antibodies in reproductive tract fluids are also affected by the stage of the oestrous cycle. Immunoglobulins (IgG and IgA) in human cervical mucous increase over the days leading up to ovulation, then drop to their lowest levels one day prior to ovulation [77,78]. IgA in mouse and rat vaginal lavage is highest at oestrus and lowest at dioestrus. IgG levels show a reverse pattern, being highest at dioestrus and lowest at oestrus [1,79,80]. Treatment of ovariectomised rats with oestradiol increases IgA levels in rat uterine secretions but suppresses IgA levels in vaginal secretions [1,81] suggesting that changes in hormone levels may have different effects on immune responses in different regions of the FRT. Levels of the polymeric immunoglobulin receptor (pIgR), responsible for the transport of IgA into the FRT, are also under hormonal control. PIgR levels in rat uterus are increased by administration of oestradiol whilst levels in vaginal tissues are depressed [82,83]. In human females, the number of cervical epithelial cells containing secretory component is increased in the late secretory phase of the cycle, suggesting that expression is up-regulated by progesterone . Oestrous cycle stage also affects antibody levels in primate cervicovaginal secretions . The highest levels of total immunoglobulins (IgG and IgA) occurred during menses, and the lowest levels occurred around the time of ovulation. Following intramuscular and oral immunisation antigen-specific IgG and IgA levels in cervicovaginal fluids also peaked during menstruation . Further studies from this group  showed that total and antigen-specific IgG- and IgA-secreting cells (ASC) in the cervix of female rhesus macaques were significantly higher during the periovulatory stage of the menstrual cycle (days 11–15) than at other stages of the cycle. Somewhat surprisingly, this cycle-associated difference in numbers of ASC was also seen in non-reproductive tract immune tissues. IgG ASC numbers in spleen, axillary lymph node, mesenteric lymph node and PBMC were also significantly higher during the periovulatory stage of the cycle compared to other stages. IgA ASC numbers were highest at the periovulatory stage of the cycle in cervix and vagina but also in spleen, inguinal lymph nodes, axillary lymph nodes and PBMC. These changes could not be accounted for by changes in the relative frequency of B cell numbers in the different tissues. Using PBMC from male macaques these authors showed that addition of progesterone to in vitro PBMC cultures decreased both IgG and IgA ASC numbers while addition of oestradiol increased the numbers of cells secreting both isotypes . Interestingly, the effect of both progesterone and oestradiol required the presence of CD8 T cells because removal of CD8 T cells from PBMC cultures ablated the antibody-enhancing or -inhibitory effects of the hormones. The authors suggest that, in monkeys, the actions of steroid hormones on CD8 T cells can regulate B cell antibody secretion not only in the FRT but also in some systemic lymphoid tissues. There is some discrepancy between the results this group has generated regarding immunoglobulin levels in cervicovaginal secretions, which were highest around menses , and the numbers of ASC in FRT tissues, which peaked during the periovulatory stage of the cycle and were lowest during the luteal phase and menstruation . The authors suggest that the increase in local immunoglobulin production cannot compensate for the dilution effect caused by increased production of cervical mucous at mid-cycle, or that the stratified squamous epithelium, which is thickest at ovulation, prevents IgG and monomeric IgA diffusion into cervicovaginal fluids.
Oestrogen has been shown to increase total IgG and IgM production by human PBMC from normal individuals [87,88]. In the former study, the immunoglobulin-enhancing effects of oestradiol were due to inhibition of CD8 suppressor cells , whilst in the latter study the effect of estradiol was neutralised by anti-IL-10 antibody  and the authors suggested that the increased IgG production was mediated by secretion of IL-10 by monocytes in the PBMC cultures. It is tempting to speculate that the effects of oestradiol and progesterone on immunoglobulin levels and secretion are mediated indirectly by hormonal effects on CD8 (and possibly CD4) T cells, or accessory cells such as macrophages. In this regard it is interesting that the LA found in human uterus have a central core of B lymphocytes surrounded by large numbers of CD8+ T lymphocytes  which, in response to hormonal changes, would be in exactly the right anatomical location to regulate local antibody secretion.
6Sex hormones and genital tract infection
The stage of the cycle or the presence of exogenous oestrogen or progesterone affects susceptibility to a number of sexually transmitted infections in both humans and animal models (reviewed in ). Oral contraceptives have previously been found to increase the incidence of vaginal candidiasis . Furthermore, mice have been found to be most susceptible to infection by Neisseria gonorrhoeae when endogenous oestrogen levels are highest at proestrus . Conversely, progesterone increased the mortality rate of mice intravaginally infected with herpes simplex virus 2 (HSV-2) . Studies of HSV-2 infection showed mice are mostly susceptible to infection during the early to late periods of dioestrus. In contrast, virus replication was not found in mice infected at oestrus . Steroid hormones were also implicated as likely cofactors in HPV-16 infection and subsequent pathogenesis of cervical carcinogenesis whereby the levels of apoptosis induced by HPV-E2 and -E7 proteins were increased by both oestrogen and progesterone . Susceptibility to genital tract infection with Mycoplasma hominis and Ureaplasma urealyticum, in a mouse model, was increased by oestradiol treatment [94,95]; however, progesterone increased susceptibility to infection with Mycoplasma pulmonis in the same model . In a widely used mouse model of genital tract infection with Chlamydia muridarum, vaginal infection is enhanced by pretreatment of mice with progesterone .
Using a rat model to study Chlamydia trachomatis infection, endocrine balances were found to influence local immune responses thereby enhancing or compromising immune protection. Specifically, a persistent infection was established in rats pretreated with progesterone prior to exposure to Chlamydia whereas untreated rats exposed to C. trachomatis at oestrus and dioestrus displayed a self-limiting infection . Furthermore, rats treated with progesterone were found to be more vulnerable to chlamydial intrauterine infection whereas oestradiol treatment reduced the susceptibility to infection .
7Oestrous cycle influence on vaccine-induced immunity
In order to induce protective immunity to pathogens such as N. gonorrhoeae, C. trachomatis, HIV, HSV and HPV, vaccination must induce and maintain immune memory in the FRT throughout the menstrual cycle. To date, most experimental vaccines have been unsuccessful and therefore understanding how endogenous factors such as changes in hormonal status influence immune induction in the FRT may assist in developing protective vaccines against STDs.
Studies in animal models have shown that treatment of mice with progesterone before mucosal immunisation (peroral, intraperitoneal, vaginal and intranasal) significantly increased numbers of antigen-specific antibody-secreting cells in vaginal tissues, particularly after vaginal immunisation . The importance of gender and hormone levels was further highlighted in a study examining protective vaccination against Plasmodium chabaudi malaria in mice. Specifically, 55% of male mice and 90% of female mice were protected by immunisation, but only 34% of female mice pretreated with testosterone were protected . Most recently, studies from this laboratory have shown that the stage of the oestrous cycle at which mice are immunised (perorally, transcutaneously or intranasally) affects the magnitude of the antigen-specific humoral and cell-mediated response in the mouse FRT and lymph nodes draining the FRT. However, results indicated that oestrous cycle did not affect antigen-specific immune responses uniformly following each route of immunisation [102,120]. These studies suggest that the development of successful vaccines to prevent the spread of STDs will not only require the development of immunisation strategies that target immunity to the FRT, but may also require the use of female sex hormones as adjuvants to insure that immunity is effectively induced and maintained throughout the reproductive cycle.
8Gender differences in immunity and autoimmunity
Evidence obtained mainly from murine models indicates that basic immune responses differ between males and females. In response to immunisation female mice mount a more vigorous T cell response and produce more antibody [103,104]. Women have higher CD4 T cell numbers than men  and also consistently mount higher Th2 responses than men, when PBMC were activated with polyclonal activators such as phytohaemagglutinin . There is also substantial epidemiological evidence showing that autoimmune disease is more common in women than in men (reviewed in ). Sex differences are most notable in Sjogren's syndrome and systemic lupus erythematosus (SLE) where >80% of patients are women. A hormonal influence on some autoimmune diseases, such as MS and rheumatoid arthritis (RA), is further suggested both by the increased frequency of occurrence of these diseases in women versus men as well as by the finding that disease severity decreases during pregnancy, particularly during the third trimester when progesterone and oestradiol levels are highest. There is also often a flare of disease post partum when hormone levels fall [107,108]. This is not the case for all autoimmune conditions, however, as patients with SLE may either worsen or remain unchanged throughout pregnancy . Some of the findings associated with human autoimmune disease have been duplicated in animal models. For example, in animal models of EAE and collagen-induced arthritis, disease severity decreases during pregnancy . Oophorectomy of female non-obese diabetic mice decreases the severity of insulin-dependent diabetes mellitus  and injection of the male hormone testosterone suppresses murine EAE . This effect of pregnancy on autoimmune disease has been attributed to the induction of a Th2-dominant immune system induced by pregnancy-associated hormonal changes [112,113]. A generalised shift towards Th2 responses would explain the decrease in severity of RA and MS during pregnancy as Th1 responses directed against self-tissues are thought to mediate tissue damage in these conditions. Furthermore, it would be expected that the severity of SLE, which is mediated mainly by antibodies, would be exacerbated under the Th2-dominant conditions associated with pregnancy. However, it should be noted that not everybody in the field agrees that pregnancy is associated with a generalised Th1 to Th2 shift [114,115]. It has recently been shown that IFNγ and IL-18 production by uNK cells is essential for successful implantation and pregnancy. Production of these cytokines would be expected to favour Th1-type immune responses. Furthermore, IL-12 production by monocytes isolated from normal pregnant donors is enhanced compared to monocytes from non-pregnant individuals  again suggesting that pregnancy may favour Th1 responses. Recent studies, using proteolipid protein-specific T cell clones isolated from MS patients and normal controls, suggest that oestradiol may enhance the secretion of both Th1 cytokines (IFNγ) and Th2 cytokines (IL-10) [70,71]. In these studies, higher concentrations of oestradiol (>5000 pg ml−1) stimulated IL-10 secretion by T cell clones, whereas lower doses of oestradiol stimulated IFNγ secretion. These studies also showed a biphasic effect of oestradiol on secretion of TNF, with enhancement of secretion occurring at low doses of oestradiol while inhibition was observed at higher hormone doses. Oestradiol had no effect on secretion of IL-4 or TGFβ by T cell clones, but progesterone enhanced IL-4 secretion in a dose-dependent manner. These studies suggest that in order to really understand how sex hormones affect both normal immune responses and autoimmunity it will be necessary to study the effects of oestradiol and progesterone on immune cells at concentrations that reflect those found both during the normal oestrous cycle and during pregnancy.
Studies of B lymphocytes isolated from human SLE patients and B lymphocytes from a mouse model of lupus suggest that oestradiol may also directly affect B cells in ways that exacerbate disease. Addition of oestradiol to PBMC cultures of patients with lupus, but not PBMC cultures from normal controls, enhanced secretion of IgG anti-double-stranded DNA (dsDNA) antibodies , which are believed to have a major pathogenic role in SLE. In an animal model of SLE, oestradiol has been shown not only to block tolerance induction in autoreactive B cells  but also to mediate expansion of autoreactive marginal zone B cell populations , thereby enhancing the production of pathogenic anti-dsDNA antibodies.
Data from both human and animal studies clearly demonstrate that both oestradiol and progesterone influence most components of both innate and adaptive immunity. Evidence of these effects is found in the differences in immune responses between females and males. For example, women mount more vigorous T and B cell responses and have higher circulating CD4 T cell numbers than men. The incidence of most autoimmune diseases is higher in women and in many women the hormone changes associated with pregnancy lessen the severity of disease.
Hormonal regulation of innate immunity in the FRT involves changes in production of defensins, changes in the recruitment of phagocytic cells including neutrophils, macrophages and DC and changes in both NK cell function and numbers throughout the oestrous cycle. Changes in oestradiol and progesterone levels affect both conventional and uterine NK cell subsets. With regard to inflammation, oestradiol can be considered to be anti-inflammatory. This may be due to a combination of decreased inflammatory cytokine production, inhibition of inducible nitric oxide synthase activity and decreased inflammatory cell recruitment.
Oestradiol and progesterone also influence the APC functions of DC, with oestradiol generally suppressing APC function. This may be due to decreased recruitment of DC into specific tissues, as demonstrated in studies of EAE in mice, or as a result of hormone-induced TGFβ production that maintains DC in an immature state, thereby suppressing APC functions. Many studies suggest that oestradiol may instruct DC to preferentially stimulate Th2-dominant responses. However, not all data support the hypothesis that Th2 immunity is predominant in an oestrogen-rich environment with recent studies in mice suggesting the exact opposite, that oestradiol promotes Th1 responses.
The numbers of CD8 T cells in the human FRT changes dramatically throughout the oestrous cycle as large LA appear during the secretory phase of the cycle then virtually disappear during the proliferative phase of the cycle. Changes in LA composition and numbers are also seen in the rodent FRT during the oestrous cycle, however, the changes are not as dramatic as those in the human FRT. Cytotoxic T cell function in human uterine endometrium is also affected by the oestrous cycle, being present during the proliferative phase but absent during the secretory phase of the cycle. Oestradiol may suppress IL-2 and IL-2 receptor expression by human T cells and T cell lines. Furthermore, the effects of oestradiol may be concentration-dependent, with low levels stimulating IFNγ production (Th1 immunity) whilst high levels may stimulate IL-10 secretion promoting Th2 responses.
In both humans and animal models, changes in hormone levels also regulate the levels of antibody in vaginal and uterine secretions, the numbers of antibody-secreting cells in FRT tissues and pIgR-mediated transport of IgA into the FRT. Different regions of the FRT appear to be affected differently as levels of oestradiol and progesterone rise and fall. The effect on immunoglobulins may be indirect, due to hormonal effects on either CD8 T cells or macrophages. Studies in primates suggest that the influence of hormone changes on immunoglobulin levels and local production may extend to tissues beyond the FRT, affecting tissues such as peripheral lymph nodes and spleen. These reports will need to be confirmed in further studies.
Future studies designed to understand how hormonal changes regulate immune responses must consider oestradiol and progesterone in combination, at concentrations found during both the normal oestrous cycle and during pregnancy in order to more closely replicate the in vivo situation. Studies of hormonal regulation of chemokine and chemokine receptor expression may provide new tools for controlling inflammation in the FRT as many chemokine agonists and antagonists are now becoming available. Further understanding of how female sex hormones influence immunoglobulin responses may result in the use of hormones as adjuvants in order to target immunity against STDs to the reproductive tract. These studies may also provide important insights into how many environmental oestrogens could affect immunity in both man and animals.
The authors wish to thank Professor Charles Wira, Professor John Aitken and Professor Richard Barry for their critical review of the manuscript.