Insight into innate immunity of the uterine cervix as a host defense mechanism against infection and preterm birth

In 2001, the WHO established the external Child Health Epidemiology Reference Group (CHERG) to develop epidemiological estimates for the various etiologies of death in young children [1]. In 2003, building on the work of CHERG, it was established that 10% of the 10.6 million yearly deaths in children younger than 5 years old were attributable to preterm birth (PTB) [1].

Spontaneous PTB complicates approximately 12.8% of all deliveries, with a marked increase over the last decade [2]. Concerned by this increase in the rate of PTB, the March of Dimes Scientific Advisory Committee on Prematurity [3] and the most recent report issued by the Institute of Medicine in 2008, Preterm Birth: Causes, Consequences and Prevention, suggested that studies to identify biomarkers that may predict adverse outcomes for infants born preterm should become a priority [4].

Perinatal infection & preterm birth

The greatest etiological factor for PTB worldwide is infection, mainly due to malaria, HIV and parasites [1]. This is in contrast with most developed countries, where iatrogenic delivery is responsible for approximately half of the births between 28 and 35 weeks of gestation [5]. Adverse pregnancy outcomes related to infection are due to a direct microbial attack on the fetus or from prematurity resulting from early activation of the uterine contractile machinery without fetal infection [6,7]. The first evidence of intrauterine infection involvement in triggering PTB was provided by Larsen et al. at Yale more than 30 years ago [8]. Since then, data from the same and other institutions directly implicate intrauterine infection as an etiological factor for a quarter of pregnancies delivered before 34 weeks of gestation [9,10]. Evidence of the causal role of infection in PTB is also supported by a body of work demonstrating that microbial invasion of the amniotic cavity, as identified by positive amniotic fluid (AF) cultures, occurs in 10% of patients with preterm labor and intact membranes and in as many as 38% of patients with preterm premature rupture of membranes [9].

Inflammation represents a highly orchestrated process designed to combat infection [11]. Remarkably, identification of intra-amniotic inflammation in the absence of a positive microbial culture result is not an unusual finding [9]. Thus, a relevant question remains: what accounts for this discrepancy? The field of molecular diagnosis applied contemporary knowledge to further discoveries of novel pathogens [12]. It was recently shown that AF cultures provide only a glimpse of the microbial spectrum responsible for triggering intra-amniotic and fetal inflammation [13]. The spectrum of AF microbiota may be more diverse than was once believed [13]. Molecular biological techniques such as PCR, which are more sensitive than culture because they do not require a priori knowledge of bacterial growth conditions, can detect bacteria in the AF cavity of up to 60% of pregnancies complicated by PTB [14]. Moreover, 16S rDNA fingerprinting technology demonstrated that approximately 60% of the species detected by these culture-independent methods could not be identified by culture alone [13]. Both uncultivated and difficult-to-cultivate species such as Fusobacterium nucleatum, Leptotrichia/Sneathia, Bergeyella, Peptostreptococcus, Ureaplasma parvum, Bacteroides and Clostridiales spp., can be isolated from the AF of women with intra-amniotic inflammation even in the absence of a positive microbial culture [13,15]. This finding suggests that the actual prevalence of AF infection as a cause of PTB may be even higher then suspected clinically [13]. Interestingly, the majority of the identified bacterial species are normal constituents of the vaginal flora with relatively low virulence [16,17]. Recognizing how and why some bacteria manage to evade host antimicrobial defense systems remains a significant challenge.

Uterine cervix is not only a mechanical barrier against infection

Although anatomically part of the uterus, the cervix is perhaps best viewed as a separate, complex and heterogeneous organ [18]. Its biology undergoes major molecular, enzymatic and biomechanical transformations that differ from those of the myometrium [19–23]. Recent data derived from animals with different inborn genetic collagen remodeling phenotypes suggest that the uterine cervix is a metabolically active organ and thus its ability to shorten and remodel may be genetically determined [23,24]. For centuries it was held that the function of the closed cervix was limited to maintaining the fetus inside the uterus [25]. As a consequence, a great deal of attention was focused on cervical length, since significant shortening increases the risk of PTB in both nulliparous and parous women [26].

Intra-amniotic and fetal microbial pathogens are thought to originate primarily from the vagina and rarely from other parts of the body, such as oral cavity flora [6,15]. To gain access to the fetus these microorganisms have to pass the cervico–vaginal barrier, then colonize and spread along the chorio–amniotic structures [16]. This view has been maintained since studies almost a decade ago first demonstrated that a short cervix is associated with an increased incidence of intra-amniotic infection and histological chorioamnionitis [27]. It has been implied that the uterine cervix functions as a protective mechanical barrier against ascending infection. This concept is supported by studies that demonstrate increased risk of microbial invasion of the AF in women with a disrupted cervical barrier as seen with advanced cervical dilation [9,28,29].

Yet, an increasing body of research has recently accumulated in support of the view that the endocervical epithelium and the cervical mucus plug are not only passive mechanical barriers but also have significant innate and adaptive immune functions [30,31]. Thus, a critical question is whether activation of the cervical ‘inflammasome’ accompanying intra-amniotic infection is the result or consequence of defective innate immune function of the cervix. If true, the latter implies that transcervical passage of bacteria occurs first, followed by microbial invasion of the AF cavity and activation of uterine contractility [23,32,33].

The cervix should be envisioned as a protective organ with an ability to generate an innate immune response. Recent advances in science and technology (genomics and proteomics) have facilitated improved insight into the molecular mechanisms responsible for the host defense function of the uterine cervix that would not have been possible by hypothesis-driven approaches [34]. Taken together, these findings have heightened the need to understand the molecular components of cervical innate immunity and how their dysregulation may transform the upper genital tract in an environment that is ultimately hostile for the fetus. The final goal of this line of research should be to improve outcomes based on targeted interventions in a modern diagnostic–therapeutic framework as part of the new scientific field of ‘theranostics’ [34,35].

Innate immunity of the uterine cervix

Much progress has been made toward understanding the mechanisms involved in microbial pathogenesis and host microbe symbiosis [11]. From this perspective, the uterine cervix has a unique structural and functional organization that allows peaceful cohabitation of a myriad of vaginal bacterial species (vaginal microbiota) with the maternal innate immune system.

Innate immunity is an archaic defense mechanism, preserved phylogenetically to provide the first line of resistance against infections [36]. This response requires interaction between the ‘modules’ of the innate immune system, which respond rapidly, nonspecifically and without memory [11]. These modules include receptors, humoral and serum factors (complement, cytokines and natural immunoglobulins) [37]. The cellular module of innate defenses include natural killer cells, granulocytes, macrophages and dendritic cells [37].

A large body of evidence proves that the cervical immune system involves each ‘limb’ of the innate immune apparatus. However, many challenges remain in determining why, in some instances, this peaceful cohabitation ceases, leading to a failed antimicrobial barrier where virulent microorganisms are able to reach the upper genital tract where they elicit an inflammatory response [30,38–41].

Cervical mucus, neutrophil & epithelial antimicrobial peptides

The endocervical mucosa and the cervical mucus plug are positioned strategically between a rich microbial vaginal environment and a presumed sterile uterine cavity. While an antibacterial role of the cervical mucus was proposed almost 50 years ago [42], the first demonstration of its antimicrobial properties against a variety of Gram-positive and Gram-negative bacteria was only recently provided by Hein et al.[30]. Specifically, it was proposed that low-molecular-weight substances with antibacterial activity in the cervical mucus plug, such as defensins and calprotectin, may protect the fetus against ascending infections [30]. The mucosal surface of the cervix is a portal for sexual transmission of microbes and viruses, such as HIV, with defensins playing a key role in the pathogenesis of primary infection. Cationic antimicrobial polypeptides, including defensins and cathelicidins, are one of the primary mechanisms used by the epithelium and neutrophils in the early stages of immune defense [9,43]. As a general rule, defensins have broad antibacterial activity against Gram-positive and Gram-negative bacteria but also display antifungal and antiviral activity [44]. Defensins are the main group of natural antimicrobial peptides in humans and their antimicrobial activity is the result of unique structural features that enable them to disrupt the microbial membrane [45]. Traditionally, two classes of defensins have been described: α-defensins (HNP) and β-defensins. Recently, a novel β-defensin structure has been added to this large family of antimicrobial peptides [46]. In general, α-defensins are expressed in the azurophil granules of the neutrophils (HNP-1–4) [47], whereas β-defensins are expressed primarily by epithelial cells, including those of the cervix and amniochorion [48,49]. To exert their natural endogenous antibiotic function, defensin molecules require proteolytic excision of their anionic N-terminal inhibitory propeptide [44]. Defensins also act on host cells to stimulate cytokine production, cell migration, proliferation, maturation and extracellular matrix synthesis. Defensins have recently been shown to be important players in mediating the innate immune response in the reproductive tract, including cervical mucus, amniochorion membranes and amniotic fluid [9,30,48,50–53]. Interestingly, the production of defensins by cervical epithelium occurs constitutively but also greatly increases in the setting of infection and inflammation [30,45,53]. A specific description of the mechanisms in charge for this unique regulation is highly desirable.

The S100 proteins have gained attention recently for their ability to act as alarmins, although their specific roles vary [54]. They are expressed in mammals exclusively, display a cell-specific distribution and regulate a large variety of intracellular activities. Some S100 proteins can signal by engaging the newly described receptor for advanced glycation end products [55]. S100A8 (calgranulin C), S100A12 (calgranulin A) and S100A9 (calgranulin B) are part of a multigenic family of calcium-modulated proteins (calgranulins) involved in intracellular and extracellular regulatory activities with a connection to inflammation [54]. Our group was one of the first to demonstrate the presence of S100A8 and S100A12 in the cervical secretions of nonpregnant and pregnant women by using surface-enhanced laser desorbtion ionization time-of-flight (SELDI-TOF) mass spectrometry [50]. We concluded that the S100A8 and S100A12 biomarkers are native constituents of the cervico–vaginal secretion involved in the host defense system, most likely offering constant protection against ascending infection. In the absence of isoform-specific ELISAs, mass spectrometry remains the only way to discriminate between S100A8 and S100A12 as biomarkers [50].

A significant development was the discovery that the cervix also acts as a reservoir for proteins with direct microbicidal activity such as secretory leukoprotease inhibitor, lysozyme, lactoferrin, calgranulins and defensins [56]. Thus, it is clear that cervical mucus has specialized antimicrobial functions. By incorporating peptides that limit the viability and multiplication of the pathogens and establishing a symbiotic relationship with vaginal micobiota the cervical mucus acts as a dynamic barrier between the vagina and uterus. By employing gene microarray technology, Gankovskaia et al. studied the expression levels of human defensin-1 and Toll-like receptors (TLRs; TLR1, TLR2, TLR6) in cells of the cervical mucosa of pregnant women [57]. The molecular profile generated in this study demonstrated that infection was associated with not only a remarkable increase in TLR1 and TLR2 gene expression, but a significant decrease in expression of the human defensin-1 gene as well [57]. This suggests that different arms of cervical innate immunity can be defective or dysfunctional in their response and may explain why, in certain clinical circumstances, bacteria and viruses can circumvent cervical mechanisms set in place to protect the fetus.

The findings of a recent study by Xu et al. require a short commentary [58]. The authors discovered that the relationship among human neutrophil defensin peptide 1–3 levels in the cervico–vaginal fluid, bacterial vaginosis and PTB vary by race. In African–Americans, midpregnancy human neutrophil defensin peptide 1–3 levels were more predictive of PTB than the presence of bacterial vaginosis. This underscores a key role for host response. In addition, elevated human neutrophil defensin peptide 1–3 levels could be a marker for a particular high-risk vaginal milieu that is not distinguished by the current bacterial vaginosis Nugent scoring system. Last, in the setting of infection, these results, in comparison with a study that demonstrated a decrease in expression of the human defensin-1 gene, hint that the human genome may fail to reflect the enormity and complexity of the human proteome [59]. In an era where the ‘one gene – one protein’ concept has proven incorrect and interaction conditioning, rather than changes in expression, may explain biological complexity, it seems that the role of innate cervical immunity as a risk factor for PTB can only be elucidated by the combined knowledge of gene expression profiles (genomics), RNA expression (transcriptomics), protein expression (proteomics) and exploring protein-to-protein interactions (‘interactomics’) [59,60].

Genetic markers of defective cervical innate immunity

Large epidemiological studies established that genetic risk factors play an important role in the pathogenesis of PTB [61,62]. A future challenge, however, will be to ascertain whether a quantitative or a functional disarray of the antimicrobial transcriptome is responsible for the consequent disruption of the cervical immune barrier and onset of an intrauterine infection. Availability of the human genome provides the basis for studying the nature of sequence variation, particularly single nucleotide polymorphisms (SNPs) [61] and copy number variations [63].

A recent study by Simhan et al. investigated the contribution of the maternal genome to the concentration of a palette of cytokines in the cervical secretions during the first trimester in women with bacterial vaginosis [32]. The results of this research are provocative because they show that SNPs can impact mediators of infection. Out of the 72 SNPs in six genes, an IL-10 receptor β SNP (rs6517158) demonstrated significant association with IL-10 concentration in the African–American population. Among Caucasian women, there were four SNPs in the IL-10 receptor α with a highly significant association with IL-10 concentration, adjusted for bacterial vaginosis and smoking. A polymorphism in the promoter region of the TNFα allele (-308A) has been associated with an increased risk of PTB by over-secretion of TNFα [64]. By contrast, concurrent carriage of TNFα and IL-6 (-174C) alleles may decrease the risk of PTB in women with bacterial vaginosis [65].

While SNPs are single-letter changes in the genetic code, copy number variation involves the multiplication or deletion of entire areas of DNA and is a major type of genetic variation in humans and inbred strains of mice. Analysis of the growing number of copy number variations reveals that there is an enrichment for genes involved in general defense responses (immune response, xenobiotic stimuli and regulation of cell-surface integrity and cell-surface antigens) [63]. Further support for the importance of copy number variation as ‘disease-modifier’ comes from a recent study by Gonzalez et al.[66], which has determined that HIV/AIDS susceptibility was significantly conditioned by differences in the copy number of a segmental duplication encompassing the gene encoding CCL3L1, a potent HIV-suppressive chemokine and ligand for the HIV coreceptor CCR5. Possession of a low CCL3L1 copy number was a major determinant of enhanced HIV susceptibility among individuals in four different populations and both mother-to-child and adult-to-adult transmission. This study points out that the individual differences in the quantity of immune response genes may constitute a genetic basis for the variable responses to infectious agents in a given population.

These studies confirm that the differential maternal inflammatory responses may ultimately result from our genetic backgrounds [67]. Mothers have a genetic predisposition toward either a hyper- or hypo-reactive immune system [61]. Given that the activity of TLRs also seems to exhibit genetic control, deciphering the complexity of the human genome should provide at least partial explanation for the extent to which the genetic makeup impacts the inflammatory response of different individuals. The multiple layers of gene-level complexity, however, may explain why selection of women who could benefit from targeted interventions based on recognition of genetic markers alone would be difficult at this time and why proteomic biomarkers may be better suited for this purpose.

Proteomic biomarkers of defective innate immunity in cervicovaginal secretions

The major advantage of proteomics over genomics-based technologies is that proteomics more closely relates to phenotypes rather than the source code. Comprehension of the underlying pathophysiologies and successful identification of relevant cervicovaginal protein biomarkers that can revolutionize early diagnosis and treatment of PTB is theoretically possible through application of proteomics high-throughput technologies [68]. The success of any proteomics study depends largely on experimental design and computational approach [69]. The challenge, however, is to isolate the biomarkers with true biological relevance amongst the huge number of proteins resolved and identified [69].

Discovering relevant biomarkers that can be used noninvasively to diagnose inflammation and women at risk for PTB remains critical. Proteomics holds the promise of providing direct insight into the mechanisms of cervical innate immune apparatus. A first step was to provide a complete mapping of the cervicovaginal fluid proteome [70,71].

A comprehensive review of the proteomics studies aimed at identifying cervicovaginal biomarkers relevant for PTB was recently published in Expert Review of Obstetrics and Gynecology[34]. Since the results of the first studies [50,72,73] that applied proteomic analysis of the human cervico–vaginal fluid were published, several other reports confirmed that the cervico–vaginal proteome is rich in defense proteins in both pregnant [74] and nonpregnant women [75]. Several of the identified immunoregulatory proteins, such as lacto-transferin, neutrophil gelatinase, S100 A8 (calgranulin A), S100A9 (calgranulin B), haptoglobin, defensins, lactoferrin, azurocidin and transferrin, were previously linked to intra-amniotic infection [9,52,74,75]. The presence of these proteins in cervico–vaginal secretions does not appear to be the result of an innate inflammatory response triggered, for example, by infection of the AF cavity, but rather that these proteins are normal components of the cervical innate immune system.

Together, these studies suggest that identification of longitudinal changes in the antimicrobial biomarker profiles by proteomic tools (i.e., SELDI-TOF or multidimensional liquid chromatography–mass spectrometry/mass spectrometry ) may be possible. The goal should be to identify biomarkers that may play a key role in identifying patients at risk for ascending infection and PTB. The discovery phase in this area of research is still in its infancy. Nevertheless, the task remains formidable owing to the complexity of such biological samples and the data generated, thus necessitating the development of specific mathematical models for optimal integration and interpretation. Structured network knowledge-base approaches, such as the ones provided by MetaCore™ (GeneGo, Jerkintown, PA, USA), Ingenuity Pathways Analysis Protein™ (Ingenuity Systems, Redwood City, CA, USA) or the publicly available Pathway Analysis Through Evolutionary Relationships resource [76,77,101], enable analysis of proteome-wide responses in the context of known functional inter-relationships and inter-related signaling pathways among proteins [7]. Such tools may provide future insight into significant biological processes that cannot be uncovered through hypothesis-driven approaches.

The role of the cervical innate immune response in mediating cellular damage is the subject of powerful debate. Although pro-inflammatory cytokines released as a result of TLR engagement by microbes are essential for host defense against infection, their exaggerated production can have deleterious consequences [78]. Explaining the disparity in the frequency of bacterial vaginosis and risk of PTB or preterm premature rupture of membranes based on genetic make-up, race and ethnicity is another challenge ahead [58,79].

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.