Hyaluronan in cervical epithelia protects against infection-mediated preterm birth

Increased synthesis of cervical hyaluronan (HA) from early to late pregnancy has long been proposed to play an essential role in disorganization of the collagen-rich extracellular matrix to allow for maximal compliance and dilation of the cervix during the birth process. Here, we show that HA is not essential for increased cervical distensibility during late pregnancy. Rather, cervicovaginal HA plays an unanticipated important role in epithelial barrier protection of the lower reproductive tract. Specifically, HA depletion in the cervix and vagina resulted in inappropriate differentiation of epithelial cells, increased epithelial and mucosal permeability, and strikingly increased preterm birth rates in a mouse model of ascending vaginal infection. Collectively, these findings revealed that although HA is not obligatory for cervical compliance, it is crucial for maintaining an epithelial and mucosal barrier to limit pathogen infiltration of the lower reproductive tract during pregnancy and thereby is protective against infection-mediated preterm birth.

Preterm birth (PTB) affects 11.7% of all pregnancies in the United States. Worldwide, PTB is the primary cause of infant morbidity and mortality in the first postnatal weeks of life and a risk factor for lifelong health problems in those that survive (1). Of the known causes for PTB in women, infection accounts for the largest percentage (25%–40%) of spontaneous PTB. Ascending microbes from the vagina and cervix are thought to be the primary route for infection. Prophylactic antimicrobial treatment of women has not reduced PTB rates, which suggests that the factors predisposing women to infection-mediated PTB are not confined to the microbiome composition of the reproductive tract (2, 3). Parturition, the process of birth, requires timely orchestration of multiple processes that include uterine contractions and remodeling of the cervix from closed and rigid to soft and compliant. Throughout pregnancy, the cervix must protect the fetus from external insults, such as ascending infection from the vagina through the cervix; provide structural integrity for maintenance of pregnancy; and simultaneously undergo progressive remodeling to a distensible organ to allow a safe birth. Regardless of cause, cervical changes precede onset of term and preterm labor in both women and mice (4, 5); thus, defining the mechanisms of cervical remodeling in term birth and PTB will allow for improved detection and prevention.

Hyaluronan (HA) is a nonsulfated glycosaminoglycan (GAG) found ubiquitously in the extracellular matrix of all mammalian tissues. During physiologic pregnancy and preterm labor in women and mice, HA levels increase significantly in the cervix (6–10). HA is synthesized by cervical fibroblasts, epithelial cells, and immune cells and released into the extracellular matrix (ECM) and cervicovaginal mucus. Beginning in early pregnancy, increased synthesis and deposition of collagen with reduced cross-links, along with declined synthesis of specific matricellular proteins, correlates with a significant increase in distensibility by gestation day 12 (d12) in the mouse cervix (11). At the end of pregnancy, increased expression of the HA synthase 2 (Has2) gene, and consequent increases in epithelial and stromal HA, correlate with maximal cervical distensibility, collagen disorganization, and hydration of the cervix (10). In mice, HA at term pregnancy accounts for 1% of the dry cervical mass, a 5-fold increase compared with early pregnancy (9). Cervical HA is also increased in mouse models of infection-mediated and non-infection-mediated PTB, although expression is limited to the stromal matrix (9). In addition to its structural role in the ECM, HA has diverse functions — such as cell growth, migration, adhesion, and differentiation — that are dependent on HA size and expression of HA-interacting molecules such as CD44, versican, and inter-alpha-trypsin inhibitor (12). The increase in both high– and low–molecular weight HA and cell-specific expression of HA-binding partners in the term cervix suggest that HA may have additional functions in cervical remodeling during parturition that remain to be elucidated (6).

The objectives of our present study were to evaluate the functions of cervical HA during pregnancy and infection-associated PTB using a tissue conditional transgenic mouse model devoid of HA in reproductive tissues. We tested the hypothesis that the increase in cervical HA at the end of pregnancy leads to maximal cervical distensibility and collagen disorganization that is crucial for a successful parturition, and also evaluated the role of HA in cervical epithelial function in term birth and infection-mediated PTB.

Conditional HA-depleted mouse models. To evaluate the specific functions of HA in cervical remodeling, we crossed mouse lines with global and tissue-specific KO of the 3 genes that encode HA synthase (Has1, Has2, and Has3) to generate mice in which HA was depleted in the female reproductive tract (see Methods). Targeted mutations in Has2 result in embryonic lethality (13), whereas mice with global double-KO of Has1 and Has3 (referred to herein as Has1/3 KO) develop and grow normally (14, 15). To achieve HA depletion in the reproductive tract, we used a Has2^fl/fl and heterozygous progesterone receptor Cre knockin cross (Has2^fl/fl;PR-Cre^+/– mice; referred to herein as Has2 KO) (16). In the Has2 KO mice, progesterone receptor Cre was expressed throughout the reproductive tract in both epithelia and stroma of the uterus and cervix (Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI78765DS1). Crossing Has1/3 KO mice with Has2 KO mice resulted in mice with reproductive tract HA depletion (referred to herein as Has1/2/3 KO mice). Cervical HA content, assessed by fluorophore-assisted carbohydrate electrophoresis (FACE), was decreased on average 93% in Has2 KO mice and 95% in Has1/2/3 KO mice (Figure 1A). Loss of Has2 gene expression was not compensated with upregulation of Has1 or Has3 in cervices of Has2 KO mice (Figure 1B). Immunofluorescence staining of HA revealed that on d18, the Has1/2/3 KO cervical epithelia and mucus were completely devoid of HA, with low levels of HA in stroma (Figure 1, C and D). While immunofluorescence revealed cell localization of insoluble HA that was bound to other ECM structures or to cells, the majority of HA in the cervix was soluble (>93% in WT mice) and could be removed from the sections during the staining procedure (Supplemental Figure 2).

Figure 1

The majority of cervical HA is synthesized by Has2, with ≥93% depletion in Has2 KO and Has1/2/3 KO mice. (A) HA and CS/DS GAGs (sGAG) were quantified by FACE in d18 cervices from WT, Has2 KO, Has1/3 KO, and Has1/2/3 KO mice. n = 3 per genotype. (B) Decreased Has2 expression in Has2 KO and Has1/2/3 KO cervix. Has1 and Has3, while expressed at low levels in the cervix, did not increase to compensate for Has2 loss. n = 4 per genotype. (C and D) Visualization of HA (green) and cell nuclei (DAPI; blue) in d18 cervices. While HA was completely depleted from Has1/2/3 KO epithelia and mucus (C), low levels of HA were detectable in the stroma of Has2 KO (not shown) and Has1/2/3 KO mice (D), likely derived from progesterone receptor–negative cells. 4 cervices per genotype were evaluated for immunofluorescence. S, stroma; ME, mucosal epithelia (defined as terminally differentiated epithelia with visible mucus-laden vacuoles); NME, nonmucosal epithelia (defined as epithelia with no visible mucus-laden vacuoles); M, mucus; L, cervicovaginal lumen. *P < 0.05, 1-way ANOVA with pairwise multiple comparisons by Tukey test. Scale bars: 20 μm.

Parturition rates and cervical ripening. The increase of hydrophilic, high–molecular weight HA (500 to ≥2,000 kDa) during cervical ripening has been proposed to play an important role in promoting tissue hydration and collagen matrix disorganization in the cervical stroma, making HA essential for the final changes in tissue compliance required for birth (6, 17). Given HA’s established function as a structural molecule important in tissue hydration in other ECM-rich tissues, we anticipated that the cervix of HA-depleted mice would fail to undergo reorganization of the collagen matrix, resulting in a parturition defect. Surprisingly, lack of HA did not affect the time of delivery (Table 1), which suggests that cervical HA is not required for normal ripening. Moreover, HA depletion did not affect parity, the increase in hydration during cervical ripening, pubic symphysis elongation, or litter size (Supplemental Table 1).

Table 1

Parturition timing

To determine whether loss of HA affects cervical ripening, biomechanical properties of cervices from WT and Has1/2/3 KO pregnant mice were assessed late in d18 and compared with those of WT nonpregnant animals. Tissue stiffness decreased significantly in the d18 WT cervix compared with the nonpregnant WT group (Figure 2A). Interestingly, loss of cervical HA did not prevent a similar decline in tissue stiffness of Has1/2/3 KO mice in late gestation (Figure 2A and Supplemental Table 2). Although it did not affect cervical function, HA did play a role in space filling, as the spacing between collagen fibers appeared to be decreased in its absence (Figure 2B). Consistent with the biomechanical data, cervical collagen ultrastructure, as determined by measurement of fibril diameter, was similar in d18 WT and Has1/2/3 KO mice (P = 0.87; Figure 2C). These results collectively suggest that HA is not obligatory for increased tissue compliance, hydration, and matrix remodeling of the cervix during pregnancy and parturition.

Figure 2

HA is not obligatory for increased tissue compliance during cervical ripening and dilation. (A) Biomechanical assessment of cervical stiffness indicated similar declines in WT and Has1/2/3 KO cervices at term (d18) compared with the nonpregnant (NP) WT cervix. n = 5–8 per group. (B) Collagen fibers (blue) exhibited similar organization between genotypes, although they appeared more condensed in Has1/2/3 KO than in WT mice. 4 cervices per genotype were evaluated for Masson trichrome stain. (C) TEM of collagen fibrils within the cervical stroma, which were similar in diameter between WT and Has1/2/3 KO animals. At least 4 different areas in each of the 3 term cervices per genotype were evaluated for the micrographs. *P < 0.05, 1-way ANOVA with pairwise multiple comparisons by Tukey test. Scale bars: 20 μm (B); 200 nm (C).

Structure of cervical epithelia. Next, we sought to understand the potential roles of HA in the cervical epithelia, where HA is a component of the pericellular matrix and secreted mucus (Figure 1C). Mucosal epithelia are critical for defense against microorganisms: the mucus forms a physical barrier, and the cells form a junctional barrier and also express pathogen recognition receptors, chemokines, and cytokines that assist in immune surveillance. Staining for alcian blue, Masson trichrome, and the cell-cell adhesion molecule E-cadherin revealed striking disorganization of cervical epithelium in d18 Has1/2/3 KO mice (Figure 3). In particular, the ordered progression of basal epithelia near the basement membrane to polarized mucus-laden epithelial cells facing the lumen in the WT cervix was abnormal in the absence of HA. At the luminal interface, the volume of mucus-containing vacuoles appeared reduced, and the vacuoles were less uniform in shape; moreover, unlike the WT group, epithelial cells with small vacuoles were interspersed with epithelial cells lacking vacuoles, and E-cadherin expression was not uniform across the apical cells. In some regions in the Has1/2/3 KO mice, the epithelia lacking mucus-laden vacuoles appeared sloughed away from tissue; in other areas, it assumed an altered morphology of elongated squamous epithelia. This pattern was most pronounced in the ectocervix and contiguous vaginal epithelia. Despite the differences in morphology and organization, Has1/2/3 KO luminal mucosal epithelia expressed markers of terminal differentiation similar to the WT group, including normal expression of trefoil factor 1 (TFF1), an estrogen-regulated secreted peptide important in restitution and repair of gut epithelia (18). These morphological observations suggested defective proliferation or differentiation of cervical epithelia in the absence of HA. Ki67 and TUNEL staining were indistinguishable between d18 WT and Has1/2/3 KO tissues (Figure 4), which suggests that increased proliferation rate and reduced cell death, respectively, do not account for the aberrant epithelia in the HA-depleted lower reproductive tract.

Figure 3

HA depletion results in aberrant organization and differentiation of cervical epithelia. Ectocervical sections from term (late d18) WT and Has1/2/3 KO mice were evaluated. Alcian blue stained mucus vacuoles (blue) and nonmucosal cells (pink); arrows denote basal epithelial cells lining the basement membrane. Masson trichrome stained keratin and muscle fibers (red), cytoplasm (pale red), collagen (blue), and nuclei (dark brown/black). E-cadherin (green) was present on the cell surface of epithelial cells. TFF1 (green) immunofluorescence was localized to the mucus present in vacuoles of luminal epithelia and secreted mucus. In E-cadherin and TFF1 images, nuclei were stained with DAPI (blue). 4 cervices per genotype were evaluated for each stain. Scale bars: 20 μm.

Figure 4

Aberrant organization of epithelia in HA-depleted cervices does not result from increased proliferation or apoptosis. Proliferation of cervical basal epithelial cells, as assessed by Ki67 staining (green), indicated similar proliferation rates between WT and Has1/2/3 KO mice. Apoptosis was assessed by TUNEL (green) counterstained with nuclear marker (red). Arrows denote basal epithelial cells lining the basement membrane. 3 cervices per genotype were evaluated for each stain. Scale bars: 20 μm.

Compromised cervical epithelial cell differentiation. Epithelial structure was further assessed by immunofluorescent staining of proteins localized to the nonmucosal epithelia (occludin and CD44) in d18 WT and Has1/2/3 KO mice (Figure 5). Expression of CD44, a HA receptor, was strongest in the basal epithelial cells close to the basement membrane and adjacent nonmucosal epithelia, with a gradual decline in expression as cells differentiated into mucosal-laden epithelia. This pattern of expression was similar between WT and Has1/2/3 KO cervical epithelia. Occludin, a tight junction protein important in paracellular junctional barrier formation (19), was localized to the same layers of nonmucosal epithelia in Has1/2/3 KO and WT cervix; however, its expression was less apparent in the epithelial cell layers with the strongest CD44 expression (Figure 5A). Regional variations were also observed: occludin expression appeared greater in areas where the luminal-facing epithelia appeared as stratified squamous epithelia with no visible mucus-containing vacuoles (Figure 5B). The apparent disorganization and/or delay in differentiation of the epithelia layers and reduced occludin expression in the absence of HA are suggestive of compromised epithelial function. Expression analysis of genes encoding tight junctions, focal adhesions, gap junctions, adherens junctions, desmosomes, and hemidesmosomes did not reveal significant differences between d18 HA-depleted and WT cervices (Supplemental Figure 3 and Supplemental Table 3). Similarly, expression of genes encoding mucins (Muc2, Muc4, Muc5b, Muc13, and Muc15) were similar in d18 WT and HA-depleted cervices, with the exception of significantly lower Muc5b expression in the Has1/2/3 KO group (Supplemental Figure 3).

Figure 5

HA depletion diminishes occludin expression in CD44⁺ epithelia. (A) CD44 (red) and occludin (Occ; green) revealed decreased occludin expression in the CD44⁺ epithelial cells of Has1/2/3 KO versus WT mice at term (d18). (B) The luminal ectocervical epithelia and contiguous vaginal epithelia displayed reduced volume of mucus-laden vacuoles and increased occludin expression in d18 Has1/2/3 KO versus WT mice. Nuclei were stained with DAPI (blue). 4 cervices per genotype were evaluated for each stain. Scale bars: 20 μm.

PTB rates after infection. The aberrant organization of cervical mucosal epithelia in mice lacking HA in the lower reproductive tract led to the hypothesis that loss of HA may result in greater susceptibility to PTB induced by ascending infection of microbes. Live E. coli were inoculated into the vagina of WT and HA-deficient mice on d16. Delivery within 48 hours of inoculation was considered PTB, whereas mice delivering on d19 were considered term. In the WT group, 50% of mice exhibited PTB after vaginal inoculation of 10⁷ CFU E. coli, compared with 100% of the Has1/2/3 KO group (P = 0.02; Table 2). Similar exposure in Has2 KO mice resulted in a 92% PTB rate, although this difference was not statistically significant compared with WT (P = 0.11; Table 2). The marked difference in susceptibility to ascending infection was confirmed in both Has2 KO and Has1/2/3 KO mice at a lower E. coli dose of 10⁵ CFU: only 27% of WT mice exhibited PTB, compared with 92% of HA-depleted mice (Table 2). Thus, loss of epithelial HA increased susceptibility to infection-induced PTB. This increased susceptibility was dependent on loss of Has2, as PTB rates of Has1/3 KO and WT mice were similar (Table 2). The increased sensitivity to infection in HA-depleted mice was accompanied by a more robust inflammatory response, with marked increases in gene expression of both classical and alternatively activated macrophage markers in Has1/2/3 versus WT cervices (Figure 6). Neither oral administration nor vaginal application of HA (200 μg/d on d12–d17) rescued E. coli–induced PTB in the Has1/2/3 KO mice (Supplemental Figure 4A). The failure to rescue may be due to the fact that neither route of HA administration led to significant accumulation of HA in Has1/2/3 KO cervicovaginal epithelia (Supplemental Figure 4B). In contrast to ascending exposure to infection, PTB rates were similar between WT and Has1/2/3 KO mice given systemic LPS (Table 2). These data indicated that HA-depleted mice were not inherently hypersensitive to systemic infection–induced PTB, but demonstrated increased susceptibility to localized bacteria and ascending infection due to the breach in epithelial integrity. Thus, disrupted barrier function of the cervical epithelia in HA-depleted reproductive tract is the primary defect.

Figure 6

HA depletion compromises epithelial barrier function and increases susceptibility to infection-mediated PTB. Gene expression of classical (Ptgs2, Il1a, and Tnfa) and alternative (Ym1 and Arg) immune response after 24-hour live vaginal E. coli exposure (10⁵ CFU) in WT and Has1/2/3 KO (KO) mice. n = 4 per group. *P < 0.05 vs. uninfected, 1-way ANOVA with pairwise multiple comparisons by Tukey test.

Table 2

Infection-induced PTB rates

Epithelial and mucosal barrier function. In the next series of experiments, potential mechanisms by which HA depletion affects epithelial and mucosal barrier function were assessed using in vitro permeability assays in a human ectocervical cell line, ECT-1, and in vivo inoculation of microparticles into the mouse vagina. Epithelial permeability assays were carried out in human ECT-1 cells seeded on inserts with or without siRNA targeting Has3, the primary HA synthase expressed in these cells. In contrast to the pregnant mouse and human cervix, in which HA is synthesized predominantly by HAS2, ECT-1 cells express HAS3. Despite differences in HA synthase isoform expression in the cell line versus cervix, it is well established that HA synthesized from either isozyme is functionally the same (20). The movement of FITC-labeled dextran present in the medium on the inserts to the receiver plate provides a direct measure of cell permeability. siRNA-mediated Has3 knockdown decreased HA production 60%–70% and increased ECT-1 cell permeability (Figure 7A). However, adding exogenous HA to Has3 siRNA–treated ECT-1 cells did not restore epithelial permeability (Figure 7A). In addition to the epithelial tight junctions, the mucus contributes to the overall barrier function of the epithelia. The integrity of the mucosal barrier was assessed by placement of polyethylene glycol–coated (PEG-coated) polystyrene microparticles (PSPEGs) that were larger in diameter than the mucus pore spacing, but similar in size to E. coli (1 μm), into the mouse cervicovaginal canal on d17. Particles densely coated with PEG have mucoinert surfaces, such that coated particles smaller in diameter than the mucus pores were able to penetrate human and mouse cervicovaginal mucus (21, 22). However, if the mucus pore spacing is smaller than the particle diameter, even mucoinert particles will become sterically trapped in the mucus mesh (21). At 15 minutes after application of 1 μm PSPEGs into the vagina, similar fractions traversed into the uterus in both WT and Has1/2/3 KO mice; however, in both genotypes, the majority of the PSPEGs were retained in the cervicovaginal canal (Figure 7B). Higher-magnification images in the ectocervical areas did reveal differences in PSPEG movement: the particles moved deeper into the mucosal layer in Has1/2/3 KO versus WT mice, appearing closer to epithelia (Figure 7B). The substantial difference in penetration of mucus by the PSPEGs, which were similar in size to bacteria, suggests that HA within the cervicovaginal mucus plays a role in mucosal barrier protection from pathogens.

Figure 7

HA depletion compromises epithelial and mucosal barrier functions. (A) siRNA-mediated Has3 silencing increased permeability of confluent ECT-1 cells. However, treatment with exogenous HA did not restore siRNA-treated cells to control levels. FACE analysis confirmed 60%–70% HA knockdown in Has3 siRNA–treated groups. (B) Fluorescent PSPEGs (red) penetrated deeper into the mucus layer in Has1/2/3 KO mice than in WT controls. TFF1 staining (green) marks mucus and mucosal epithelia. Nuclei were stained with DAPI (blue). n = 4–5 per group. *P < 0.05, 1-way ANOVA with pairwise multiple comparisons by Tukey test. Scale bars: 50 μm.

Although a substantial number of PTBs in women are associated with ascending infection from the vagina through the cervix, no specific microbe or microbiome profile has been shown to be causative of PTB, and prophylactic antibiotic treatment during or prior to pregnancy does not alter PTB rates (23, 24). Our present findings offer a rationale for previous observations and provide new insights into the importance of the cervical epithelia and mucus to form a protective barrier against pathogen invasion, especially during extensive remodeling of the cervix during late gestation. HA secreted by the epithelia plays key roles in barrier function. These roles include regulation of epithelial cell differentiation and organization as well as modulation of cervical mucus permeability. The lower expression of the junctional protein occludin observed in CD44⁺ epithelial cells of HA-depleted mice (Figure 5, Supplemental Figure 3, and Supplemental Table 3) suggests that HA affects the protein expression and/or subcellular distribution of adhesion/junction molecules, rather than the transcription of these genes. Our observations that HA was necessary for appropriate differentiation of the cervical epithelia, and that its depletion within the cervicovaginal mucosal epithelia increased the vulnerability to microbia-induced PTB, endorse the importance of HA synthesis in the epithelia in maintenance of pregnancy. This function expands HA’s proposed role in protection of mucosal epithelia, as a recent study reported the importance of HA in breast milk to activate antimicrobial expression in gut mucosal epithelia of the offspring ingesting the milk (25).

The fact that vaginal HA application from d12 to d17 failed to rescue the sensitivity to PTB in Has1/2/3 KO mice, as well as the failed restoration of permeability in HA-treated ECT-1 cells with siRNA-mediated Has3 knockdown, suggest that HA function is more complex than simply improving the barrier properties of the mucus. It is possible that the program of cell differentiation of basal epithelia may require HA exposure specifically to the basal epithelia, which is likely not efficient with vaginal HA application, or that exposure may be required at earlier time points in pregnancy. Our present data indicated that the absence of HA leads to abnormal differentiation and epithelial layer organization and altered mucosal properties that lead to declined barrier function. Barrier function encompasses both the mucosal barrier properties and the paracellular regulation between epithelial cells regulated by junctional proteins. Differences in the biophysical characteristics of cervical mucus from women at high risk of PTB supports the importance of mucosal properties in barrier function (26). Given that transcript expression of mucins, antimicrobial peptides, tight junctions, and adherens junctions were similar between d18 WT and HA-depleted cervices, the function of HA and its binding partners in cervical epithelial function and cell signaling remains to be elucidated. The observed regional and differentiation-specific differences in expression of occludin and E-cadherin protein suggest that epithelial barrier function is compromised, although further studies are required to understand HA’s role. Although numerous HA-interacting molecules are expressed in the pregnant cervix to modulate HA function, we propose that the interaction of HA with its receptor, CD44, is likely critical for epithelial differentiation and barrier function. This notion is supported by the fact that squamous epithelial cells, such as keratinocytes, are also dependent on CD44 for normal cellular function. Cd44^–/– mice have skin permeability dysfunction, and HA-CD44 interactions are required for mediating intracellular signaling pathways necessary for keratinocyte migration, adhesion, proliferation, and differentiation (27, 28).

A second novel finding of the current study is that HA within the stromal ECM was not essential for collagen disorganization and increased hydration of the cervix during the accelerated phase of cervical ripening at the end of pregnancy. During cervical ripening, HA is produced in a range of sizes from low to high molecular weight. High–molecular weight HA has a well-established function as a structural molecule that increases tissue flexibility, elasticity, and/or resilience: it can form cross-links with the proteoglycan versican to generate structures that are space-filling and increase elasticity of a tissue while maintaining tissue strength. HA has been proposed to be directly linked with ECM disorganization in the term cervix of many species. While collagen bundles were more closely packed in HA-depleted mice, suggesting that HA is involved as a space-filling molecule, collagen fibril size and resulting tissue mechanical strength were not affected (Figure 2). However, our present studies only evaluated cervical stiffness, and it remains possible that HA loss in the cervix may affect other biomechanical parameters, such as recoverability of the tissue (elasticity) or its ability to resist fracturing and breaking (strength). Low–molecular weight HA has proposed immunomodulating functions that include recruitment and activation of leukocytes. HA depletion did not alter the induction of either classical or alternatively activated macrophages; rather, the gene expression response was similar between WT and Has1/2/3 KO mice (Figure 6).

Globally, PTB and stillbirths associated with PTB result in 3.3 million perinatal deaths per year (29). While the direct causes of PTB remain unknown, the primary risk factors for PTB and resulting perinatal death in developing countries are poor nutrition and infection. Despite the magnitude of this health problem, there are few effective strategies for diagnosis or prevention of PTB. The unanticipated role of HA as a protective factor against pathogen-mediated PTB warrants further studies to determine whether reduced cervical HA expression (or increased HA metabolism) may be a susceptibility factor predisposing certain women to infection-associated PTB and whether long-term exogenous HA may reduce risk of infection-mediated PTB. Such findings could readily be translated into cost-effective preventative therapies.

View Supplemental data

We thank F.J. DeMayo (Baylor College of Medicine, Houston, Texas, USA) for providing PR-Cre^+/– mice, J. McDonald (Mayo Clinic Scottsdale, Scottsdale, Arizona, USA) for generation of the Has3^–/– mouse, N. Itano and K. Kimata (Aichi Medical University, Aichi, Japan) for generation of the Has1^–/– mouse, and E.V. Maytin and V.C. Hascall (Cleveland Clinic, Cleveland, Ohio, USA) for providing Has1/3 KO mice. This work was supported by NIH grant 5-PO1-HD011149.

Address correspondence to: Mala Mahendroo, Department of Obstetrics and Gynecology and Green Center for Reproductive Biological Sciences, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, Texas 75235-9032, USA. Phone: 214.648.3091; E-mail: mala.mahendroo@utsouthwestern.edu.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2014;124(12):5481–5489. doi:10.1172/JCI78765.

1. Martin JA, Hamilton BE, Ventura SJ, Osterman MJ, Mathews TJ. Births: final data for 2011. Natl Vital Stat Rep. 2013;62(1):1–69.
   View this article via: PubMed Google Scholar
2. Romero R, Mazor M. Infection and preterm labor. Clin Obstet Gynecol. 1988;31(3):553–584.
   View this article via: PubMed Google Scholar
3. Romero R, Chaiworapongsa T, Espinoza J. Micronutrients and intrauterine infection, preterm birth and the fetal inflammatory response syndrome. J Nutr. 2003;133(5 suppl 2):1668S–1673S.
   View this article via: PubMed Google Scholar
4. Akgul Y, Mahendroo M. Cervical changes accompanying birth. In: Croy BA, Yamada AT, DeMayo FJ, Adamson SL, eds. The Guide to Investigation of Mouse Pregnancy. Boston, Massachusetts, USA: Academic Press; 2014:391–401.
5. Cunningham F, Leveno K, Bloom S, Spong CY, Dashe J. Williams Obstetrics 24/E. 24th ed. New York, New York, USA: McGraw Hill Professional; 2014.
6. Ruscheinsky M, De la Motte C, Mahendroo M. Hyaluronan and its binding proteins during cervical ripening and parturition: dynamic changes in size, distribution and temporal sequence. Matrix Biol. 2008;27(5):487–497.
   View this article via: PubMed CrossRef Google Scholar
7. Osmers R, Rath W, Pflanz MA, Kuhn W, Stuhlsatz HW, Szeverenyi M. Glycosaminoglycans in cervical connective tissue during pregnancy and parturition. Obstet Gynecol. 1993;81(1):88–92.
   View this article via: PubMed Google Scholar
8. Tateyama N, Asakura H, Takeshita T. Correlation between an absence of cervical gland area on transvaginal sonography and cervical mucus hyaluronic acid levels in women with threatened preterm delivery. J Perinat Med. 2013;41(2):151–157.
   View this article via: PubMed Google Scholar
9. Akgul Y, Holt R, Mummert M, Word A, Mahendroo M. Dynamic changes in cervical glycosaminoglycan composition during normal pregnancy and preterm birth. Endocrinology. 2012;153(7):3493–3503.
   View this article via: PubMed CrossRef Google Scholar
10. Straach KJ, Shelton JM, Richardson JA, Hascall VC, Mahendroo MS. Regulation of hyaluronan expression during cervical ripening. Glycobiology. 2005;15(1):55–65.
    View this article via: PubMed Google Scholar
11. Akins ML, Luby-Phelps K, Bank RA, Mahendroo M. Cervical softening during pregnancy: regulated changes in collagen cross-linking and composition of matricellular proteins in the mouse. Biol Reprod. 2011;84(5):1053–1062.
    View this article via: PubMed CrossRef Google Scholar
12. Dicker KT, Gurski LA, Pradhan-Bhatt S, Witt RL, Farach-Carson MC, Jia X. Hyaluronan: A simple polysaccharide with diverse biological functions. Acta Biomaterialia. 2014;10(4):1558–1570.
    View this article via: PubMed CrossRef Google Scholar
13. Camenisch TD, et al. Disruption of hyaluronan synthase-2 abrogates normal cardiac morphogenesis and hyaluronan-mediated transformation of epithelium to mesenchyme. J Clin Invest. 2000;106(3):349–360.
    View this article via: JCI PubMed CrossRef Google Scholar
14. Kobayashi N, et al. Hyaluronan deficiency in tumor stroma impairs macrophage trafficking and tumor neovascularization. Cancer Res. 2010;70(18):7073–7083.
    View this article via: PubMed CrossRef Google Scholar
15. Bai K-J, et al. The role of hyaluronan synthase 3 in ventilator-induced lung injury. Am J Respir Crit Care Med. 2005;172(1):92–98.
    View this article via: PubMed CrossRef Google Scholar
16. Soyal SM, et al. Cre-mediated recombination in cell lineages that express the progesterone receptor. genesis. 2005;41(2):58–66.
    View this article via: PubMed CrossRef Google Scholar
17. Timmons B, Akins M, Mahendroo M. Cervical remodeling during pregnancy and parturition. Trends Endocrinol Metab. 2010;21(6):353–361.
    View this article via: PubMed CrossRef Google Scholar
18. Taupin D, Podolsky DK. Trefoil factors: initiators of mucosal healing. Nat Rev Mol Cell Biol. 2003;4(9):721–732.
    View this article via: PubMed CrossRef Google Scholar
19. Blaskewicz CD, Pudney J, Anderson DJ. Structure and function of intercellular junctions in human cervical and vaginal mucosal epithelia. Biol Reprod. 2011;85(1):97–104.
    View this article via: PubMed CrossRef Google Scholar
20. Spicer AP, Tien JY. Hyaluronan and morphogenesis. Birth Defects Res C Embryo Today. 2004;72(1):89–108.
    View this article via: PubMed CrossRef Google Scholar
21. Lai SK, et al. Rapid transport of large polymeric nanoparticles in fresh undiluted human mucus. Proc Natl Acad Sci U S A. 2007;104(5):1482–1487.
    View this article via: PubMed CrossRef Google Scholar
22. Ensign LM, et al. Mucus-penetrating nanoparticles for vaginal drug delivery protect against herpes simplex virus. Sci Transl Med. 2012;4(138):138ra79.
    View this article via: PubMed Google Scholar
23. Muglia LJ, Katz M. The enigma of spontaneous preterm birth. N Engl J Med. 2010;362(6):529–535.
    View this article via: PubMed CrossRef Google Scholar
24. Iams JD, Romero R, Culhane JF, Goldenberg RL. Primary, secondary, and tertiary interventions to reduce the morbidity and mortality of preterm birth. Lancet. 2008;371(9607):164–175.
    View this article via: PubMed CrossRef Google Scholar
25. Hill DR, et al. Human milk hyaluronan enhances innate defense of the intestinal epithelium. J Biol Chem. 2013;288(40):29090–29104.
    View this article via: PubMed CrossRef Google Scholar
26. Critchfield AS, et al. Cervical mucus properties stratify risk for preterm birth. PLoS One. 2013;8(8):e69528.
    View this article via: PubMed CrossRef Google Scholar
27. Kirschner N, et al. CD44 regulates tight-junction assembly and barrier function. J Invest Dermatol. 2010;131(4):932–943.
    View this article via: PubMed Google Scholar
28. Bourguignon LY. Matrix hyaluronan-activated CD44 signaling promotes keratinocyte activities and improves abnormal epidermal functions. Am J Pathol. 2014;184(7):1912–1919.
    View this article via: PubMed CrossRef Google Scholar
29. Oestergaard MZ, et al. Neonatal mortality levels for 193 countries in 2009 with trends since 1990: a systematic analysis of progress, projections, and priorities. PLoS Med. 2011;8(8):e1001080.
    View this article via: PubMed CrossRef Google Scholar
30. Matsumoto K, et al. Conditional inactivation of Has2 reveals a crucial role for hyaluronan in skeletal growth, patterning, chondrocyte maturation and joint formation in the developing limb. Development. 2009;136(16):2825–2835.
    View this article via: PubMed CrossRef Google Scholar
31. Akgul Y, Mahendroo M. Assessment of changes in the peripartum cervix. In: Croy BA, Yamada AT, DeMayo FJ, Adamson SL, eds. The Guide to Investigation of Mouse Pregnancy. Boston, Massachusetts, USA: Academic Press; 2014:723–731.
32. Calabro A, Benavides M, Tammi M, Hascall VC, Midura RJ. Microanalysis of enzyme digests of hyaluronan and chondroitin/dermatan sulfate by fluorophore-assisted carbohydrate electrophoresis (FACE). Glycobiology. 2000;10(3):273–281.
    View this article via: PubMed CrossRef Google Scholar
33. Carson FL, Hladik C. Histotechnology: A Self-Instructional Text. Chicago, Illinois, USA: ASCP Press; 2009.